Randomized Trial of Intravitreal Bevacizumab Alone or Combined with Triamcinolone versus Macular Photocoagulation in Diabetic Macular Edema

Randomized Trial of Intravitreal Bevacizumab Alone or Combined with Triamcinolone versus Macular Photocoagulation in Diabetic Macular Edema Masoud Soheilian, MD,1,2 Alireza Ramezani, MD,1,2 Arash Obudi, MD,1,2 Bijan Bijanzadeh, MD,1,2 Masoud Salehipour, MD,1,2 Mehdi Yaseri, PhD,3 Hamid Ahmadieh, MD,1,2 Mohammad H. Dehghan, MD,1,2 Mohsen Azarmina, MD,1,2 Siamak Moradian, MD,1,2 Gholam A. Peyman, MD4 Purpose: To compare the results of intravitreal bevacizumab (IVB) injection alone or in combination with intravitreal triamcinolone acetonide (IVT) versus macular laser photocoagulation (MPC) as a primary treatment of diabetic macular edema (DME). Design: Randomized 3-arm clinical trial. Participants: A total of 150 eyes of 129 patients with clinically significant DME and no previous treatment. Methods: The eyes were randomly assigned to 1 of the 3 study arms: the IVB group, patients who received 1.25 mg IVB (50 eyes); the IVB/IVT group, patients who received 1.25 mg of IVB and 2 mg of IVT (50 eyes); and the MPC group, patients who underwent focal or modified grid laser (50 eyes). Retreatment was performed at 12-week intervals whenever indicated. Main Outcome Measures: Change in best-corrected visual acuity (VA) at week 24. Results: VA changes among the groups were statistically significant at 6 (P⬍0.001) and 24 (P ⫽ 0.012) weeks. The significant treatment effect was demonstrated in the IVB group at all follow-up visits and in the IVB/IVT group at 6 and 12 weeks. VA changes ⫾ standard deviation at 36 weeks were ⫺0.28⫾0.25, ⫺0.04⫾0.33, and ⫹0.01⫾0.27 logarithm of minimum angle of resolution in the IVB, IVB/IVT, and MPC groups, respectively (P ⫽ 0.053). Significant central macular thickness (CMT) reduction was observed in all groups only up to 6 weeks; however, CMT changes were not significant among the groups in all visits. Overall, retreatment was required for 27 eyes up to 36 weeks (14 in the IVB group, 10 in the IVB/IVT group, and 3 in the MPC group). In the IVB group, in which a greater VA improvement was observed, only 1 injection was required in 72% of the cases. VA improvement ⬎2 Snellen lines at 36 weeks was detected in 37%, 25%, and 14.8% of patients in the IVB, IVB/IVT, and MPC groups, respectively. Conclusions: Intravitreal bevacizumab injection in patients with DME yielded a better visual outcome at 24 weeks compared with macular photocoagulation. A change in CMT beyond the 6-week time point that corresponded to the vision change was not detected. No adjunctive effect of IVT was demonstrated. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Ophthalmology 2009;116:1142–1150 © 2009 by the American Academy of Ophthalmology.

Diabetic macular edema (DME) is the main cause of visual impairment in diabetic patients.1,2 Conventional treatment is based mainly on laser photocoagulation with the probable mechanism of rejuvenation of retinal pigment epithelium cells or improvement of outer retinal oxygenation.3,4 The Early Treatment Diabetic Retinopathy Study (ETDRS) showed that macular laser photocoagulation (MPC) was beneficial for eyes with clinically significant macular edema.5 This beneficial effect was inferred only because MPC reduced the risk of moderate visual acuity (VA) loss by 50%. In their report, approximately 17% of the treated eyes had a 3-line improvement in VA. For diffuse DME, MPC has even more limited results. Lee and Olk6 demonstrated that with modified grid MPC, VA was stabilized in 60.9%, decreased in 24.6%, and increased in only 14.5% of eyes

1142

© 2009 by the American Academy of Ophthalmology Published by Elsevier Inc.

with diffuse DME. Therefore, alternative or adjunct treatments for DME, such as intravitreal triamcinolone acetonide (IVT)7–11 and anti-vascular endothelial growth factor (VEGF) therapy,12–15 have been the focus of the most recent attentions. Corticosteroids, by increasing tight-junction proteins and local vasoconstriction,3,16 as well as their angiostatic properties by VEGF inhibition,17 diminish vascular leakage and therefore may have a beneficial effect on DME.7–11 However, anti-VEGF drugs, by affecting endothelial tight junction proteins, decrease vascular permeability in ocular vascular diseases such as DME.18 VEGF-A levels are considerably higher in patients with DME showing extensive leakage in the macular region than in patients showing minimal leakage.19,20 ISSN 0161-6420/09/$–see front matter doi:10.1016/j.ophtha.2009.01.011

Soheilian et al 䡠 Bevacizumab with or without Triamcinolone vs. Laser for DME Human VEGF-A is found in at least 9 isoforms. Currently used anti-VEGF drugs are pegaptanib, ranibizumab, and bevacizumab. Pegaptanib is a ribonucleic acid aptamer that targets only the VEGF-165 isoform, and its beneficial effect on DME was recently reported.12 Ranibizumab and bevacizumab are recombinant antibodies with a pan-VEGFA– blocking activity and have shown promising effects in the treatment of DME.13,21 The use of anti-VEGF drugs is becoming increasingly more prevalent; however, some unresolved issues, such as ideal regimen, duration of treatment, potential of combination treatments, and safety concerns with long-term VEGF inhibition, deserve further investigations. We previously reported the 12-week results of a randomized 3-arm clinical trial evaluating the efficacy of intravitreal bevacizumab (IVB) injection alone or in combination with IVT in comparison with MPC as the first line of treatment in DME.22 We demonstrated that a single IVB injection yielded better visual outcome in comparison with laser photocoagulation, although it was not associated with a significant decrease in macular thickness. We did not observe any additive effect from IVT in such eyes. To complete the trial, we enrolled more cases (150 eyes) and repeated the treatments as required. In this second report, we present the results of the study up to 36 weeks.

Patients and Methods This clinical trial was approved by institutional review board of ophthalmic research center of Labbafinejad Medical Center (Tehran, Iran). The study protocol and its probable safety and efficacy of the interventions were explained to all participants before enrollment. Informed consent was obtained from all patients.

Participants Eligible cases were 150 eyes of 129 patients with clinically significant DME based on ETDRS criteria.5 Exclusions were previous panretinal or focal laser photocoagulation, prior intraocular surgery or injection, history of glaucoma or ocular hypertension, VA of 20/40 or better or worse than 20/300, presence of iris neovascularization, high-risk proliferative diabetic retinopathy, and significant media opacity. Monocularity, pregnancy, serum creatinine ⱖ3 mg/dl, and uncontrolled diabetes mellitus were also among the exclusion criteria.

Intervention A complete ophthalmic examination, including best-corrected VA, slit-lamp biomicroscopy, tonometry, and funduscopy, was performed at baseline. Ancillary tests consisted of fundus photography, fluorescein angiography, and optical coherence tomography. Best-corrected VA by the Snellen chart was recorded in logarithm of minimum angle of resolution (logMAR) scale. Lens opacity was graded from 0 to 4⫹ clinically. Optical coherence tomography mapping was performed using commercially available equipment (Zeiss, Dublin, CA). Retinal thickness was measured in a circle (3.5 mm diameter) centered on the fixation point. Mean thickness on the 1-mm circle centered on the fovea (central macular thickness [CMT]) was recorded and considered for statistical analysis. Eligible eyes were allocated to 1 of 3 study groups: IVB group, eyes receiving IVB injection alone; IVB/IVT group, eyes receiving

IVB injection plus IVT; and MPC group, eyes undergoing macular photocoagulation. Macular photocoagulation was carried out according to the modified ETDRS protocol. This treatment involved only areas of thickened retina, areas of retinal nonperfusion, and leaking microaneurysms. In bilateral cases, each eye was enrolled in the study individually; thus, both eyes of 1 patient could be allocated in 1 group. Indication for retreatment consisted of persistent clinically significant macular edema based on ETDRS criteria if VA was not better than 20/40. Retreatments were performed at 12-week intervals as required.

Surgical Technique Under restricted sterile condition, use of anesthetic eye drop, and insertion of a lid speculum, injections were performed. For the IVB group, 0.05 ml (1.25 mg) of bevacizumab (Avastin; Genentech, Inc., South San Francisco, CA [made for F. Hoffmann-La Roche, Ltd., Basel, Switzerland]) was injected intravitreally with a 27-gauge needle through the supratemporal quadrant. For the IVB/IVT group, in addition to IVB injection, 0.05 ml (2 mg) of IVT was injected with another 27-gauge needle through the inferotemporal quadrant. In bilateral cases, the injection for the second eye was performed after 2 days. In the MPC group, standard focal or modified grid laser was performed. Patients who received injections were examined at 1 and 7 days after injections for anterior chamber reaction and intraocular pressure measurement. Complete ocular examination and optical coherence tomography were performed again at 6, 12, 24, and 36 weeks. Fluorescein angiography was repeated as needed. Blood pressure measurement was performed initially and at each visit.

Outcome Measures Primary outcome measure was change in best-corrected VA (logMAR) at week 24. Secondary outcomes were VA changes at 6, 12, and 36 weeks, as well as CMT changes by optical coherence tomography and potential injection-related complications.

Sample Size To have a 90% power for detection of 0.2 logMAR difference (equal to 2 Snellen lines) in the mean VA among the groups as significant (at the 2-sided 5% level) with an assumed standard deviation of 0.33, 50 eyes for each group were required.

Randomization Randomization was performed using the random block permutation method according to a computer-generated randomization list. The block length varied randomly (6, 12). Random allocation sequence was performed by a biostatistician. The detail of series was unknown by the study investigators.

Masking A sham laser procedure (20 seconds) was performed by aiming the laser beam on the macula for the eyes in the IVB and IVB/IVT groups. In the MPC group, a sham injection was done by a needleless syringe pressed against the conjunctiva. To keep the masking process, patients were prevented from seeing the syringes. All procedures were run by staff members other than the study investigators to preserve investigator masking. Best-corrected VA measurement and OCT were performed by certified examiners masked both to the randomization and to the findings of previous measurements.

1143

Ophthalmology

Volume 116, Number 6, June 2009

Statistical Methods Statistical analysis was performed by the Statistical Package for the Social Sciences software (version 15; SPSS, Inc., Chicago, IL). For descriptive purposes, qualitative variables were stated using percentage, and quantitative data were reported by mean ⫾ standard deviation. To compare data in the baseline, we used chisquare test or Fisher exact test for qualitative data and analysis of variance for quantitative data. The following proportions were used to describe the ratio of VA and CMT improvement in each group at weeks 6, 12, 24, and 36: (VA at weeks 6, 12, 24, and 36 – baseline VA)/baseline VA and (CMT at weeks 6, 12, 24, and 36 – baseline CMT)/baseline CMT. For comparing VA and CMT with baseline values within each group, paired sample t test was used. The marginal regression model (based on generalized estimating equation methods) was used to compare VA and CMT in the treatment groups adjusted for the baseline values and to eliminate any possible correlation effects between the 2 eyes of patients in bilateral enrolled cases. P values less than 0.029 were considered statistically significant to control the study-wise type 1 error (primarily type 1 error was set at 0.05), based on ␣ spending function method presented by Pocock23 for an interim analysis.

Results In this trial, 150 eyes of 129 patients were enrolled and followed from September 2005 to May 2007. The mean age of patients ⫾ standard deviation was 61.2⫾6.1 years. Seventy-nine eyes (52.7%) were from male patients. A total of 141 (94%) eyes had nonproliferative diabetic retinopathy, and 9 eyes (6%) had early proliferative diabetic retinopathy. The general characteristics of each treatment group are summarized in Table 1. The eyes were randomly assigned to one of the treatment groups: (1) 50 eyes in the IVB group, (2) 50 eyes in the IVB/IVT group, and (3) 50 eyes in the MPC group. Retreatment was required for 27 eyes up to 36 weeks (14 eyes in the IVB group, 10 eyes in the IVB/IVT group, and 3 eyes in the MPC group). A third repeated treatment was required in 3 eyes in the IVB group, 3 eyes

Table 1. Distribution of Baseline Characteristics in Each Treatment Group Treatment Group

No. of eyes Mean age (yrs) ⫾ SD Female/male (No. of eyes) Mean diabetes duration (yrs) ⫾ SD Mean IOP (mmHg) ⫾ SD Retinopathy severity NPDR/early PDR

IVB

IVB/IVT

MPC

P

50 60.5⫾5.9 27/23 10.5⫾3.2

50 62.3⫾6.8 22/28 10.4⫾2.6

50 61.0⫾5.3 0.293* 22/28 0.379† 10.5⫾2.9 0.990*

16.7⫾2.4 46/4

14.4⫾2.6 48/2

15.9⫾2.2 0.136* 47/3 0.909‡

IOP ⫽ intraocular pressure; IVB ⫽ intravitreal bevacizumab; IVT ⫽ intravitreal triamcinolone; MPC ⫽ macular laser photocoagulation; NPDR ⫽ nonproliferative diabetic retinopathy; PDR ⫽ proliferative diabetic retinopathy; SD ⫽ standard deviation. *Based on analysis of variance. † Based on chi-square test. ‡ Based on Fisher exact test.

1144

in the IVB/IVT group, and 1 eye in the MPC group. The number of cases evaluated at each follow-up and finally analyzed, and the causes of incomplete data are presented in a flow chart (Fig 1). The means and proportions of improvement of the corrected VA and CMT in each group at every visit are presented in Tables 2 and 3. The comparisons were performed after adjustment of the parameters according to their baseline quantities to compensate for the influence of dissimilar baseline VA and CMT on the results (Table 3). Compared with the baseline, VA improvement was significant in the IVB groups at all follow-up visits up to 36 weeks (P⬍0.001). In the IVB/IVT group, VA improved significantly only at weeks 6 and 12 (P ⫽ 0.002 and 0.019, respectively). In the MPC group, however, VA changes in relation to the baseline did not change significantly (Table 3). Figure 2 demonstrates mean VA changes at different follow-ups in relation to the baseline in each group. In the marginal regression model, the generalized estimating equation analysis demonstrated that the differences in VA changes among the groups were statistically significant only at 6 and 24 weeks (P⬍0.001 and P ⫽ 0.012, respectively). Pairwise comparison between groups showed that the VA improvement at 6 weeks in both IVB and IVB/IVT groups was greater than in the MPC group (P⬍0.001); however, there was no significant difference between the IVB and IVB/IVT groups (P ⫽ 0.199). At 24 weeks, pairwise comparison disclosed that the difference of VA changes between the IVB and MPC groups was significant in favor of the IVB group (P ⫽ 0.003). This difference between the IVB/IVT and MPC groups was borderline (P ⫽ 0.033); however, no significant difference between the IVB and IVB/IVT groups was observed (P ⫽ 0.373). At 12 and 36 weeks, there was more VA improvement in the IVB group than in the other groups, although not to a meaningful level (Table 3). To evaluate the effects of different applied treatments on Snellen VA, we compared the percentages of the eyes with more than 2 lines VA improvement, the eyes with stable VA (within 2 lines changes), and the eyes with more than 2 lines VA decline among the groups. Overall, the percentage of the eyes with stable VA was relatively similar among the groups at all follow-ups. However, a greater percentage of cases gained more than 2 Snellen lines in the IVB and IVB/IVT groups than in the MPC groups. Furthermore, a greater percentage of eyes lost more than 2 Snellen lines in the MPC group than in the other groups. These differences were statistically significant among the groups at 6, 12, and 24 weeks (Table 4). Compared with the baseline, CMT decreased significantly in all groups only at 6 weeks. The reduction of CMT was more in the IVB group in relation to the other 2 treatment groups, although their differences did not reach a significant level at any follow-up (Table 3). In the marginal regression model, generalized estimating equation analysis showed no statistically significant difference in CMT changes among the 3 groups at all follow-up visits (Table 3). The eyes were categorized into 3 subgroups (⬍300 ␮m, 300 – 399 ␮m, and ⱖ400 ␮m) according to the initial CMT. There was no significant difference concerning the numbers of eyes in each category among the groups (Table 5). However, in the subgroup ⬍300 ␮m, the baseline CMT was significantly greater in the IVB/IVT group than in the other groups (P ⫽ 0.018). There was no meaningful difference in the other subgroups in terms of baseline CMT. After interventions, the reduction of CMT was significantly different among the treatment groups only in the subgroups with the initial CMT ⱖ400 ␮m at all follow-up visits except week 24. At week 6 in the subgroup ⱖ400 ␮m, CMT reduction (mean of percentages) was greater in the IVB group than in the other groups

Soheilian et al 䡠 Bevacizumab with or without Triamcinolone vs. Laser for DME

Figure 1. Flow chart showing progression of subjects through trial. HR-PDR ⫽ high-risk proliferative diabetic retinopathy.

(34.1⫾19.7%, 24.8⫾28.3%, and ⫺18.1⫾14.3% in the IVB, IVB/ IVT, and MPC groups, respectively; P ⫽ 0.026). At week 12 in the same subgroup, the difference in CMT reduction was significant among all 3 groups in favor of the IVB/IVT group (22.7⫾30.1%, 27.4⫾22.7%, and 10.9⫾18.5% in the IVB, IVB/IVT, and MPC groups, respectively; P⬍0.001). At week 36 in the subgroup of ⱖ400 ␮m, CMT reduction in relation to the baseline was observed in all groups; however, the difference was significant only between the IVB and MPC groups (⫺27.2⫾34.8%, ⫺8.8⫾35.9%, and ⫺15.1⫾14.6% in the IVB, IVB/IVT, and MPC groups, respectively; P⬍0.001). To reduce any potential bias regarding the cases

with missing data at any of the visits, we compared VA and CMT changes of these patients with those of the patients with complete data, and no significant difference was found. Transient anterior chamber reaction (trace to 1⫹ cell) was observed in 10 (20%) and 9 (18%) eyes in the IVB and IVB/IVT groups, respectively. This side effect resolved spontaneously in all after 1 week. Ocular hypertension (ⱖ23 mmHg) was detected in 8 eyes (16%) of the IVB/IVT group and was controlled in all by medical therapy except in 1 eye that progressed to neovascular glaucoma. Severe lens opacity developed in 5 eyes: 4 in the IVB/IVT group and 1 in the MPC group. Initially, retinal neovas-

Table 2. Mean Corrected Visual Acuity and Central Macular Thickness of Each Group before and 6, 12, 24, and 36 Weeks after Intervention Mean CMT (␮m) ⴞ SD

Mean VA (logMAR) ⴞ SD Groups

IVB

IVB/IVT

MPC

Before 6 wks 12 wks 24 wks 36 wks

0.71⫾0.28 0.54⫾0.26 0.49⫾0.28 0.50⫾0.28 0.50⫾0.28

0.73⫾0.28 0.60⫾0.29 0.60⫾0.35 0.58⫾0.33 0.62⫾0.40

0.55⫾0.26 0.60⫾0.26 0.55⫾0.36 0.60⫾0.41 0.52⫾0.37

P* ⬍0.001 0.556 0.324 0.825 0.272

IVB

IVB/IVT

MPC

P*

341⫾149 278⫾102 293⫾132 317⫾132 290⫾132

359⫾137 309⫾122 320⫾139 327⫾153 332⫾137

300⫾118 284⫾111 296⫾127 290⫾117 291⫾106

0.034 0.288 0.320 0.158 0.380

CMT ⫽ central macular thickness; logMAR ⫽ logarithm of minimum angle of resolution; IVB ⫽ intravitreal bevacizumab; IVT ⫽ intravitreal triamcinolone; MPC ⫽ macular laser photocoagulation; SD ⫽ standard deviation; VA ⫽ visual acuity. *P value based on generalized estimating equation analysis unadjusted for the baseline value.

1145

Ophthalmology

Volume 116, Number 6, June 2009

cularization was observed in 4, 2, and 3 eyes in the IVB, IVB/IVT, and MPC groups, respectively. These neovascularizations were resolved in all except 1 eye in the MPC group. Furthermore, 8 eyes developed early proliferative diabetic retinopathy during the study period (1 in the IVB group, 4 in the IVB/IVT group, and 3 in the MPC group). These eyes remained stable during the follow-up. Ten eyes progressed to high-risk proliferative diabetic retinopathy (4 in the IVB group, 3 in the IVB/IVT group, and 3 in the MPC group). These eyes were treated accordingly and excluded from the study. Significant blood pressure increase, thromboembolic events, and serious ocular complications such as vitreous hemorrhage, endophthalmitis, and retinal detachment were not detected in this study. Four patients (5 eyes) died during the course of the trial (2 in the IVB/IVT group and 2 in the MPC group).

Discussion This 3-arm randomized clinical trial demonstrated the superiority of IVB injection either alone or in combination Table 3. Mean Changes and Proportions of Improvement in Corrected Visual Acuity and Central Macular Thickness ⫾ Standard Deviation for Each Group at 6, 12, 24, and 36 Weeks after Intervention Treatment Groups

VA change (logMAR) 0–6 wks Improvement P within group 0–12 wks Improvement P within group 0–24 wks Improvement P within group 0–36 wks Improvement P within group CMT changes (␮m) 0–6 wks Improvement P within group 0–12 wks Improvement P within group 0–24 wks Improvement P within group 0–36 wks Improvement P within group

IVB

IVB/IVT

⫺0.18⫾0.19 ⫺24.8% ⬍0.001† ⫺0.21⫾0.19 ⫺31.2% ⬍0.001† ⫺0.23⫾0.22 ⫺31.0% ⬍0.001† ⫺0.28⫾0.25 ⫺35.1% ⬍0.001†

⫺0.11⫾0.21 ⫺15.4% 0.002† ⫺0.11⫾0.28 ⫺14.7% 0.019† ⫺0.07⫾0.28 ⫺7.3% 0.178 ⫺0.04⫾0.33 ⫺6.3% 0.579

⫺65⫾114 ⫺13.5% ⬍0.001† ⫺37⫾115 ⫺7.8% 0.038 ⫺24⫾103 ⫺3.1% 0.176 ⫺56⫾140 ⫺10.1% 0.044

MPC

P* among Groups

0.06⫾0.19 ⬍0.001 23.0% 0.047 0.02⫾0.31 0.616 13.3% 0.640 0.01⫾0.36 0.012† 8.6% 0.858 0.01⫾0.27 0.053 2.7% 0.865

⫺61⫾119 ⫺25⫾60 ⫺12.6% ⫺6.0% 0.004† 0.015† ⫺36⫾128 4⫾90 ⫺5.8% 3.1% 0.101 0.814 ⫺14⫾102 ⫺15⫾80 ⫺2.9% ⫺1.6% 0.420 0.269 ⫺5⫾113 ⫺8⫾67 1.0% 0.6% 0.845 0.541

†

with triamcinolone acetonide over MPC in VA improvement up to 24 weeks in primary treatment of DME. This improving effect persisted longer in the IVB group (up to 36 weeks) than in the IVB/IVT group (up to 12 weeks). In the MPC group, no improvement in VA was observed at all follow-up visits. In regard to CMT reduction, there was no meaningful superiority of the IVB and IVB/IVT groups Table 4. Level of Visual Acuity Changes According to the Snellen Lines at Each Follow-up in Relation to the Baseline Values

0.265

0.640

0.842

0.261

CMT ⫽ central macular thickness; IVB ⫽ intravitreal bevacizumab; IVT ⫽ intravitreal triamcinolone; logMAR ⫽ logarithm of the minimal angle of resolution; MPC ⫽ macular laser photocoagulation; VA ⫽ visual acuity. *P values based on generalized estimating equation analysis adjusted for the baseline value. † Significant based on method presented by Pocock23 for interim analysis (P⬍0.029).

1146

Figure 2. Mean VA changes in relation to the baseline values ⫾ 95% confidence interval in logMAR scale in the 3 groups at timely visits. IVB ⫽ intravitreal bevacizumab; IVT ⫽ intravitreal triamcinolone; MPC ⫽ macular laser photocoagulation.

P Values among the IVB* IVB/IVT† MPC‡ Groups Treatment Groups (%)

VA Changes Improvement ⬎2 lines Stable within 2 lines Worsening ⬎2 lines 12 wks Improvement ⬎2 lines Stable within 2 lines Worsening ⬎2 lines 24 wks Improvement ⬎2 lines Stable within 2 lines Worsening ⬎2 lines 36 wks Improvement ⬎2 lines Stable within 2 lines Worsening ⬎2 lines 6 wks

27.9 69.8 2.3 36.4 63.6 0 31.4 68.6 0 37.0 59.3 3.7

26.3 65.8 7.9 27.0 62.2 10.8 21.2 63.6 15.2 25.0 54.2 20.8

5.1 71.8 23.1 8.6 71.4 20.0 11.4 65.7 22.9 14.8 66.7 18.5

0.003*

0.002*

0.014*

0.164

IVB ⫽ intravitreal bevacizumab; IVT ⫽ intravitreal triamcinolone; MPC ⫽ macular laser photocoagulation; VA ⫽ visual acuity. *Significant based on method presented by Pocock23 for interim analysis (P⬍0.029).

Soheilian et al 䡠 Bevacizumab with or without Triamcinolone vs. Laser for DME Table 5. Number of Eyes and Mean Baseline Central Macular Thickness in Each Subgroup Categorized on the Basis of Initial Central Macular Thickness Mean Baseline CMT ⴞ SD

No. of Eyes (%) Initial CMT (␮m) ⬍300 300–399 ⱖ400

IVB

IVB⫹IVT

MPC

IVB

IVB⫹IVT

MPC

P Value

25 (50) 10 (20.8) 15 (29.2)

21 (42.6) 13 (25.5) 16 (31.9)

30 (59.2) 11 (22.4) 9 (18.4)

224⫾30 341⫾32 540⫾97

243⫾46 345⫾32 524⫾98

224⫾40 334⫾35 502⫾86

0.018 0.060 0.672

CMT ⫽ central macular thickness; IVB ⫽ intravitreal bevacizumab; IVT ⫽ intravitreal triamcinolone; MPC ⫽ macular laser photocoagulation; SD ⫽ standard deviation.

over the MPC group. Significant reduction of CMT in relation to the baseline measurement was observed only at 6 weeks in all groups. In addition, it was shown that in the primary treatment of DME, triamcinolone acetonide not only had no additive effect over bevacizumab but also caused less favorable visual outcome in the combined treatment group. In view of the key role of VEGF in the pathophysiology of diabetic retinopathy, VEGF blockade is an attractive therapeutic approach. Bevacizumab is a pan-VEGF–blocking agent and may impair normal physiologic VEGF-mediated functions.24,25 This might be considered as a disadvantage; however, recent popularity, more availability, and reasonable cost of this drug persuaded us to use this medication in this clinical trial. The beneficial effect of IVB on patients with DME has been demonstrated in recent published studies, including the preliminary results of the present clinical trial.22 In a phase II study, the Diabetic Retinopathy Clinical Research (DRCR) Network disclosed a median 1-line improvement at 3 weeks that was sustained through 12 weeks by 2 injections of either 1.25 or 2.5 mg IVB.26 However, our study demonstrated an almost 0.3 logMAR improvement in mean VA (⬇3 Snellen lines), persisting for 36 weeks. Furthermore, 72% of our cases required only 1 IVB injection during the study period. The DRCR Network26 also demonstrated an inferiority of MPC compared with IVB, similar to our study. This study also clarified that CMT reduction after IVB at 3 weeks appeared to plateau or decrease in most eyes between the 3- and 6-week visits. This finding was comparable to our results showing a diminishing effect on CMT after 6 weeks. In that study, it was suggested that 6 weeks might be too long for an optimal second-injection interval. However, our study underlines that the therapeutic effect of a single IVB injection may persist for up to 36 weeks in regard to VA improvement. This vision improvement without a significant decrease in CMT may be explained by increased macular perfusion rather than leakage reduction or fluid resorption. Moreover, it has been shown that VA changes are not always parallel to CMT changes in DME.27 The effective period of IVB in the treatment of DME was reported to be 6 and 12 weeks.14 The Pan-American Collaborative Retina Study Group reported that 20.5% of their cases needed a second injection and 7.7% needed a third injection within 6 months.28 In our study, similarly, 22% and 6% of eyes required a second and third injection, respectively. These repeated injections of VEGF inhibitors

may cause retinal atrophy by blocking neuroprotective cytokines.25,29,30 Therefore, routine prescheduled repeated injections may not be appropriate in the primary treatment of patients with DME and the decision for retreatment should be individualized in each patient. In a recent pilot study, the short-term beneficial effect of intravitreal ranibizumab therapy for improving VA and reducing CMT in patients with DME was reported.13 Comparison among efficacy end points of pegaptanib, ranibizumab, and bevacizumab in DME is limited by differences in study size, study design, inclusion/exclusion criteria, drug dosage, and drug administration schedule. Few prospective, randomized studies have evaluated the effect of intravitreal injection of triamcinolone on DME.9,11,31 From their 2-year study results, Gillies et al9 concluded that repeated IVT injections improved vision and reduced CMT in eyes with refractory DME. To enhance therapeutic effects, we combined IVB injection with IVT. However, we did not observe any additive effects of IVT in terms of VA improvement and CMT reduction. VA improvement in the IVB/IVT group had a shorter duration than in the IVB group (12 vs. 36 weeks). Theoretically, lens opacity progression, ocular hypertension, and adverse effects of triamcinolone preservatives might be the causes of this observation. Addition of triamcinolone also did not cause a greater reduction in CMT in our study. Nonetheless, in a recent comparative case series on 28 eyes of 14 patients by Shimura et al,32 IVT (4 mg) reduced CMT and improved VA better than IVB. Considering the results of this study and ours, one may conclude that injection of either IVB or IVT alone has a better outcome than combined treatment. In the study by Shimura et al, 16 of 28 eyes had undergone panretinal photocoagulation and 18 of 24 eyes had a history of cataract surgery before intervention. Both of these conditions would increase the inflammatory response and may explain the greater reported effect of triamcinolone, a reminder that both of these confounding factors were among our exclusion criteria. We injected 2 mg triamcinolone (instead of the usual dosage of 4 mg)33–35 to diminish the side effects of the drug and to avoid having a total volume of intravitreal injection of ⬎0.1 ml, which might necessitate anterior chamber paracentesis. The optimal dosage of IVT has been a subject of debate. Most researchers recommend 4 mg triamcinolone for intravitreal injection.33–35 Recently, however, Audren et al36 compared the efficacy and safety of intravitreal injection of 2 mg versus 4 mg of triamcinolone for treatment of

1147

Ophthalmology

Volume 116, Number 6, June 2009

DME and noticed no difference in dose. It has been shown that even with lower amounts of injected triamcinolone, corticosteroid receptors inside the eye become saturated.37 The ETDRS trial showed that focal MPC decreased the rate of moderate visual loss in eyes with clinically significant macular edema.5 Lee and Olk6 demonstrated that 24.6% of the eyes with diffuse DME lost 3 Snellen lines or more with modified grid MPC. With this treatment, only 14.5% of cases experienced vision improvement. In our study, mean VA did not change significantly up to 36 weeks after macular photocoagulation. Visual acuity improvement ⬎2 Snellen lines at 36 weeks was observed in 37.0%, 25.0%, and 14.8% in the IVB, IVB/IVT, and MPC groups, respectively. Therefore, considering the potential side effects of MPC, such as transient increased macular edema,38 paracentral scotoma, subretinal fibrosis, and inadvertent foveolar burns,39,40 we may suggest that in comparison with MPC, IVB is a better choice for primary treatment of DME. However, combining IVB with MPC may have an additive effect, although this has not been proven in the short term by the DRCR Network.26 In addition, the DRCR Network recently proved that MPC was more effective and had fewer side effects than IVT during a 2-year period for most patients with DME; although in the short term, VA was better in the IVT group.41,42 Therefore, one may assume that with a longer follow-up, a more beneficial effect of MPC might have emerged in our study. DME with various macular thicknesses may respond differently to treatments. In a recent published report, eyes with moderate macular thickening of 300 to 400 ␮m benefit most from MPC in relation to the eyes with greater or less macular thickening at baseline.43 To address this issue, a subgroup analysis based on initial CMT was performed in our study, which disclosed that in the eyes with a baseline CMT less than 400 ␮m, the CMT changes were almost comparable in all treatment groups. In the eyes with CMT greater than 400 ␮m at baseline, however, the response to the treatments varied among the groups. In this subgroup, a CMT reduction was observed in all treatment groups at all follow-up intervals. At the final visit in this study (week 36), this reduction was more prominent in the IVB group, followed by the MPC and IVB/IVT groups. One may conclude that decision making for the primary treatment of DME in choosing 1 of the 3 approaches used in this study is more important when the baseline CMT is greater than 400 ␮m in regard to anatomic outcome. Nevertheless, any definite conclusion in this regard should be made with caution, and this issue needs to be further investigated in larger studies. Retinal neovascularization regression was observed in 6% of the eyes in both intravitreally injected groups and in 4% of the eyes in the MPC group in our study. Progression to early proliferative diabetic retinopathy was observed in 5% and 6% and to high-risk proliferative diabetic retinopathy in 7% and 6% in the intravitreally injected groups and the MPC group, respectively. Retinal neovascularization regression in eyes with proliferative diabetic retinopathy in association with IVB treatment has been reported.21,44,45 The lack of a significant difference in the neovascularization regression and progression rates among the groups in the

1148

current study may be due to the low number of cases showing such events. Hypertension, thromboembolic events, and major intravitreal injection-related complications were not encountered in our study. Ocular hypertension (ⱖ23 mmHg) was observed in 8 eyes (16%) of the IVB/IVT group and was controlled in all by antiglaucoma medication, except in 1 eye that progressed to neovascular glaucoma. Mild anterior chamber reaction occurred in 20% and 18% of eyes in the IVB and IVB/IVT groups, respectively, 1 day after injection, which spontaneously resolved within 1 week. Although we did not encounter any significant complications in this study, further investigation to assess the safety of the drugs is needed. For patients presenting with DME for the first time, MPC is currently the standard treatment. However, for the reasons mentioned, MPC may not be the ideal treatment. In this randomized clinical trial, we compared the results of bevacizumab with or without triamcinolone versus laser macular photocoagulation in the treatment of DME. We conclude that, compared with MPC, IVB injection effectively increased VA for up to 24 weeks, although its effect on decreasing retinal edema was transient. In 72% of eyes, this beneficial effect persisted up to 36 weeks even with a single injection of bevacizumab. Moreover, we could not demonstrate any additive effect of triamcinolone. On the basis of the results of this study, IVB injection alone may be an alternative or even first-line treatment in such cases. In our study, baseline VA was better in the MPC group than in the other groups. This may affect the outcomes of our study in favor of the other treatment groups, because fewer changes in the group with a better initial VA are usually expected. To deal with this inadvertent substantial imbalance in baseline VA and CMT in the groups, statistical adjustment was performed to lessen this confounding effect. However, it should be remembered that the statistical methods could not always overcome the problem of unequal distribution of the factors among groups. Another source of bias was the cases with missing data, especially in the combined treatment group. In the present study, the combination of MPC with intravitreal drugs, such as anti-VEGFs or corticosteroids, was not assessed. Combination therapy may have a more beneficial effect on DME than either MPC or intravitreal drugs alone. Another study by the DRCR Network is evaluating a combination of IVT and MPC, as well as a combination of ranibizumab and MPC, in the treatment of DME.41 We did not address some issues concerning the initial characteristics of DME, such as mild versus moderate visual loss or focal versus diffuse macular edema, in this trial. Furthermore, in this study we report the results up to 36 weeks, which can be used as a basis for planning longer trials. A longer follow-up of the same patients for up to 2 years, however, is under way in our institution (Labbafinejad Medical Center). Larger studies with longer outcomes evaluating the therapeutic effects of bevacizumab focusing on different features of DME are recommended.

Soheilian et al 䡠 Bevacizumab with or without Triamcinolone vs. Laser for DME

References 20. 1. Klein R, Klein BE, Moss SE. Visual impairment in diabetes. Ophthalmology 1984;91:1–9. 2. Hardy RA, Crawford JB. Retina. In: Vaughn D, Asbury T, Riordan-Eva P, eds. General Ophthalmology. 15th ed. Stamford, CT: Appleton & Lange; 1999:178 –99. 3. Tranos PG, Wickremasinghe SS, Stangos NT, et al. Macular edema. Surv Ophthalmol 2004;49:470 –90. 4. Gottfredsdottir MS, Stefansson E, Jonasson F, Gislason I. Retinal vasoconstriction after laser treatment for diabetic macular edema. Am J Ophthalmol 1993;115:64 –7. 5. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study report number 1. Arch Ophthalmol 1985;103:1796 – 806. 6. Lee CM, Olk RJ. Modified grid laser photocoagulation for diffuse diabetic macular edema: long-term visual results. Ophthalmology 1991;98:1594 – 602. 7. Sutter FK, Simpson JM, Gillies MC. Intravitreal triamcinolone for diabetic macular edema that persists after laser treatment: three-month efficacy and safety results of a prospective, randomized, double-masked, placebo-controlled clinical trial. Ophthalmology 2004;111:2044 –9. 8. Massin P, Audren F, Haouchine B, et al. Intravitreal triamcinolone acetonide for diabetic diffuse macular edema: preliminary results of a prospective controlled trial. Ophthalmology 2004;111:218 –25. 9. Gillies MC, Sutter FK, Simpson JM, et al. Intravitreal triamcinolone for refractory diabetic macular edema: two-year results of a double-masked, placebo-controlled, randomized clinical trial. Ophthalmology 2006;113:1533– 8. 10. Jonas JB, Kamppeter BA, Harder B, et al. Intravitreal triamcinolone acetonide for diabetic macular edema: a prospective, randomized study. J Ocul Pharmacol Ther 2006;22:200 –7. 11. Audren F, Erginay A, Haouchine B, et al. Intravitreal triamcinolone acetonide for diffuse diabetic macular oedema: 6-month results of a prospective controlled trial. Acta Ophthalmol Scand 2006;84:624 –30. 12. Macugen Diabetic Retinopathy Study Group. A phase II randomized double-masked trial of pegaptanib, an anti-vascular endothelial growth factor aptamer, for diabetic macular edema. Ophthalmology 2005;112:1747–57. 13. Chun DW, Heier JS, Topping TM, et al. A pilot study of multiple intravitreal injections of ranibizumab in patients with center-involving clinically significant diabetic macular edema. Ophthalmology 2006;113:1706 –12. 14. Haritoglou C, Kook D, Neubauer A, et al. Intravitreal bevacizumab (Avastin) therapy for persistent diffuse diabetic macular edema. Retina 2006;26:999 –1005. 15. Starita C, Patel M, Katz B, Adamis AP. Vascular endothelial growth factor and the potential therapeutic use of pegaptanib (Macugen) in diabetic retinopathy. Dev Ophthalmol 2007;39: 122– 48. 16. Sears JE, Hoppe G. Triamcinolone acetonide destabilizes VEGF mRNA in Muller cells under continuous cobalt stimulation. Invest Ophthalmol Vis Sci 2005;46:4336 – 41. 17. Gardner TW, Antonetti DA, Barber AJ, et al., Penn State Retina Research Group. Diabetic retinopathy: more than meets the eye. Surv Ophthalmol 2002;47(suppl):S253– 62. 18. Antcliff RJ, Marshall J. The pathogenesis of edema in diabetic maculopathy. Semin Ophthalmol 1999;14:223–32. 19. Funatsu H, Yamashita H, Ikeda T, et al. Vitreous levels of interleukin-6 and vascular endothelial growth factor are re-

21. 22.

23. 24. 25. 26. 27.

28.

29. 30.

31.

32.

33. 34. 35. 36.

37.

lated to diabetic macular edema. Ophthalmology 2003;110: 1690 – 6. Funatsu H, Yamashita H, Sakata K, et al. Vitreous levels of vascular endothelial growth factor and intercellular adhesion molecule 1 are related to diabetic macular edema. Ophthalmology 2005;112:806 –16. Avery RL, Pearlman J, Pieramici DJ, et al. Intravitreal bevacizumab (Avastin) in the treatment of proliferative diabetic retinopathy. Ophthalmology 2006;113:1695–705. Soheilian M, Ramezani A, Bijanzadeh B, et al. Intravitreal bevacizumab (Avastin) injection alone or combined with triamcinolone versus macular photocoagulation as primary treatment of diabetic macular edema. Retina 2007;27:1187–95. Pocock SJ. Group sequential methods in the design and analysis of clinical trials. Biometrika 1997;64:191–9. Bhisitkul RB. Vascular endothelial growth factor biology: clinical implications for ocular treatments. Br J Ophthalmol 2006;90:1542–7. Gillies MC. What we don’t know about Avastin might hurt us. Arch Ophthalmol 2006;124:1478 –9. Diabetic Retinopathy Clinical Research Network. A phase II randomized clinical trial of intravitreal bevacizumab for diabetic macular edema. Ophthalmology 2007;114:1860 –7. Schmid KE, Neumaier-Ammerer B, Stolba U, Binder S. Effect of grid laser photocoagulation in diffuse diabetic macular edema in correlation to glycosylated haemoglobin (HbA1c). Graefes Arch Clin Exp Ophthalmol 2006;244;1446 –52. Arevalo JF, Fromow-Guerra J, Quiroz-Mercado H, et al., Pan-American Collaborative Retina Study Group Primary intravitreal bevacizumab (Avastin) for diabetic macular edema: results from the at 6-month follow-up. Ophthalmology 2007; 114:743–50. Jin K, Zhu Y, Sun Y, et al. Vascular endothelial growth factor (VEGF) stimulates neurogenesis in vitro and in vivo. Proc Natl Acad Sci U S A 2002;99:11946 –50. Oosthuyse B, Moons L, Storkebaum E, et al. Deletion of the hypoxia-response element in the vascular endothelial growth factor promoter causes motor neuron degeneration. Nat Genet 2001;28:131– 8. Ahmadieh H, Ramezani A, Shoeibi N, et al. Intravitreal bevacizumab with or without triamcinolone for refractory diabetic macular edema. Graefes Arch Clin Exp Ophthalmol 2008;246: 483–9. Shimura M, Nakazawa T, Yasuda K, et al. Comparative therapy evaluation of intravitreal bevacizumab and triamcinolone acetonide on persistent diffuse diabetic macular edema. Am J Ophthalmol 2008;145:854 – 61. Martidis A, Duker JS, Greenberg PB, et al. Intravitreal triamcinolone for refractory diabetic macular edema. Ophthalmology 2002;109:920 –7. Beer PM, Bakri SJ, Singh RJ, et al. Intraocular concentration and pharmacokinetics of triamcinolone acetonide after a single intravitreal injection. Ophthalmology 2003;110:681– 6. Nelson ML, Tennant MT, Sivalingam A, et al. Infectious and presumed noninfectious endophthalmitis after intravitreal triamcinolone acetonide injection. Retina 2003;23:686 –91. Audren F, Lecleire-Collet A, Erginay A, et al. Intravitreal triamcinolone acetonide for diffuse diabetic macular edema: phase 2 trial comparing 4 mg vs 2 mg. Am J Ophthalmol 2006;142:794 –9. Diabetic Retinopathy Clinical Research Network. A randomized trial comparing intravitreal triamcinolone acetonide and laser photocoagulation for diabetic macular edema. Version 4.0. 2005;1–13. Available at: http://www.t1diabetes.nih.gov/ investigator/resources/Protocol_B.pdf. Accessed December 21, 2008.

1149

Ophthalmology

Volume 116, Number 6, June 2009

38. Laursen ML, Moeller F, Sander B, Sjoelie AK. Subthreshold micropulse diode laser treatment in diabetic macular oedema. Br J Ophthalmol 2004;88:1173–9. 39. Thompson MJ, Ip MS. Diabetic macular edema: a review of past, present, and future therapies. Int Ophthalmol Clin 2004; 44:51– 67. 40. Christoforidis JB, D’Amico DJ. Surgical and other treatments of diabetic macular edema: an update. Int Ophthalmol Clin 2004;44:139 – 60. 41. Diabetic Retinopathy Clinical Research Network. A randomized trial comparing intravitreal triamcinolone acetonide and focal/grid photocoagulation for diabetic macular edema. Ophthalmology 2008;115:1447–59.

42. Schachat AP. A new look at an old treatment for diabetic macular edema. Ophthalmology 2008;115:1445– 6. 43. Estabrook EJ, Madhusudhana KC, Hannan SR, Newsom RS. Can optical coherence tomography predict the outcome of laser photocoagulation for diabetic macular edema? Ophthalmic Surg Lasers Imaging 2007;38:478 – 83. 44. Rosenfeld PJ, Fung AE, Puliafito CA. Optical coherence tomography findings after an intravitreal injection of bevacizumab (Avastin) for macular edema from central retinal vein occlusion. Ophthalmic Surg Lasers Imaging 2005;36:336 –9. 45. Mason JO III, Nixon PA, White MF. Intravitreal injection of bevacizumab (Avastin) as adjunctive treatment of proliferative diabetic retinopathy. Am J Ophthalmol 2006;142:685– 8.

Footnotes and Financial Disclosures Originally received: August 13, 2008. Final revision: January 12, 2009. Accepted: January 12, 2009. Available online: April 19, 2009.

Presented at: American Academy of Ophthalmology Annual Meeting, November 2008, Atlanta, Georgia. Manuscript no. 2008-978.

1

Ophthalmology Department and Ophthalmic Research Center, Labbafinejad Medical Center, Shaheed Beheshti Medical University, Tehran, Iran. 2

Negah Eye Hospital, Tehran, Iran.

3

School of Public Health and Institute of Public Health Research, Tehran University of Medical Sciences, Tehran, Iran. 4

Department of Ophthalmology, University of Arizona Health Science Center, Tucson, Arizona.

1150

Trial registration: clinical trials.gov identifier: NCT00370669. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Supported by the Ophthalmic Research Center of Shahid Beheshti University (MC) Tehran, Iran. Correspondence: Masoud Soheilian, MD, Ophthalmic Research Center, Labbafinejad Medical Center, Pasdaran Ave. Boostan 9 St. Tehran 16666, Iran. E-mail: [email protected].