To compare visual acuity and spectral-domain optical coherence tomography (SDOCT) outcomes associated with intravitreal (IV) bevacizumab vs IV ranibizumab for the management of diabetic macular edema (DME).
Prospective randomized trial.
Forty-eight patients (63 eyes) with center-involved DME were randomly assigned to receive 1.5 mg (0.06 cc) IV bevacizumab or 0.5 mg (0.05 cc) IV ranibizumab at baseline and monthly if central subfield thickness was greater than 275 μm.
Forty-five patients (60 eyes) completed 48 weeks of follow-up. At baseline, mean ± standard error best-corrected visual acuity (BCVA) (logMAR) was 0.60 (20/80) ± 0.05 in the IV bevacizumab group and 0.63 (20/85) ± 0.05 in the IV ranibizumab group. A significant improvement in mean BCVA was observed in both groups at all study visits ( P < .05); this improvement was significantly greater in the IV ranibizumab group compared with the IV bevacizumab group at weeks 8 ( P = .032) and 32 ( P = .042). A significant reduction in mean central subfield thickness was observed in both groups at all study visits compared with baseline ( P < .05), with no significant difference in the magnitude of macular thickness reduction between groups. The mean number of injections was significantly higher ( P = .005) in the IV bevacizumab group (9.84) than in the IV ranibizumab group (7.67).
IV bevacizumab and IV ranibizumab are associated with similar effects on central subfield thickness in patients with DME through 1 year of follow-up. IV ranibizumab is associated with greater improvement in BCVA at some study visits, and the mean number of injections is higher in the IV bevacizumab group.
Macular edema is the leading cause of decreased visual acuity in patients with diabetic retinopathy. Laser photocoagulation has been the standard-of-care treatment for diabetic macular edema (DME) for decades, based on the Early Treatment Diabetic Retinopathy Study (ETDRS) and other more recent clinical trials. However, because visual acuity improvement post laser is observed infrequently, and because of the frequent recurrence or persistence of DME after laser treatment, there is a need for better treatments for the management of DME (especially for diffuse DME involving the foveal center, since focal DME not involving the foveal center may have a good prognosis after focal laser treatment). Care must be taken, however, because the terms “diffuse” and “focal” DME have not been defined consistently in the literature; these terms have referred to a variety of diverse parameters (clinical and angiographic) itemized differently by various authors. In addition, the definition of center- and non-center-involved DME may vary; the Diabetic Retinopathy Clinical Research Network (DRCRnet) has defined non-center-involved DME as “a baseline central subfield thickness <250 microns and a baseline photograph assessment of retinal thickness at the center of the macula graded as none or questionable.” Moreover, the parameters of a “normal” central subfield threshold may vary depending on the optical coherence tomography (OCT) machine employed.
Among pharmacologic treatments currently available for DME, antiangiogenic agents such as bevacizumab and ranibizumab have been reported to be associated with visual acuity improvement and favorable remodeling of the macular architecture in patients with DME. Ranibizumab has been evaluated in phase III prospective randomized clinical trials and reported to be associated with better visual acuity outcomes compared to focal/grid laser in patients with DME. To our knowledge and based on a Medline search, there is no published study comparing intravitreal (IV) bevacizumab and IV ranibizumab for the treatment for DME. We conducted a randomized, prospective study to compare the visual acuity and spectral-domain optical coherence tomography (SDOCT) outcomes associated with IV bevacizumab vs IV ranibizumab for the management of DME.
The current study is a prospective randomized clinical trial registered at ClinicalTrials.gov ( NCT01487629 ). The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the local Institutional Review Board, Comitê de Ética em Pesquisa do Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto, and all participants gave written informed consent before entering into the study. All patients evaluated in the Retina Section of the Department of Ophthalmology, School of Medicine of Ribeirão Preto of the University of Sao Paulo with center-involved DME in at least 1 eye between July 1, 2010 and August 31, 2011 were invited to participate in the study.
Patient Eligibility and Baseline Evaluation
Inclusion criteria were as follows: (1) center-involved DME, defined as a central subfield thickness >300 μm on SDOCT, despite at least 1 session of macular laser photocoagulation performed at least 3 months previously; (2) best-corrected ETDRS visual acuity (BCVA) measurement between 0.3 logMAR (Snellen equivalent: 20/40) and 1.6 logMAR (Snellen equivalent: 20/800); (3) signed informed consent.
Exclusion criteria were: (1) vitreomacular traction on SDOCT; (2) proliferative diabetic retinopathy needing panretinal photocoagulation (PRP) or anticipated to need PRP in the next 12 months; (3) macular capillary dropout on fluorescein angiography; (4) history of glaucoma or ocular hypertension (defined as an intraocular pressure higher than 22 mm Hg); (5) an ocular condition (other than diabetes) that, in the opinion of the investigator, might affect macular edema or alter visual acuity during the course of the study (eg, retinal vein occlusion, uveitis or other ocular inflammatory disease, neovascular glaucoma, etc); (6) systemic corticosteroid therapy; (7) any condition that, in the opinion of the investigator, might preclude follow-up throughout the study period.
Each patient received a detailed ophthalmologic examination including measurement of BCVA according to the standardized ETDRS refraction protocol using a retroilluminated Lighthouse for the Blind distance visual acuity test chart (using modified ETDRS charts 1, 2, and R; Precision Vision, IL), as well as applanation tonometry, undilated and dilated slit-lamp biomicroscopic examination, indirect fundus examination, and fluorescein angiography using high-resolution angiography (HRA; Heidelberg Engineering, Heidelberg, Germany).
Fourier-domain OCT evaluation (Spectralis Eyetracker Tomographer, HRA-OCT; Heidelberg Engineering) was performed in all patients, and retinal thickness measurements were acquired using a standard 20 × 15-degree raster scan protocol consisting of 19 horizontal sections (each computed out of 25 frames) with a distance of 240 μm between each horizontal scan, covering a square of 20 × 15 degrees on the retina and centered on the foveal region. Follow-up mode was used to reduce test-retest variability. In order to optimize the accuracy of OCT data, automatic delineation of the inner and outer boundaries of the neurosensory retina generated by OCT built-in software was verified for each of the scans. Central subfield thickness values were calculated automatically as the average thickness of a central macular region 1000 μm in diameter centered on the patient’s foveola by built-in Heidelberg software using retinal map analysis.
If both eyes were eligible for treatment and the patient agreed to treat both eyes with anti-VEGF therapy, 1 eye received the randomized treatment according to a computer-generated sequence and the contralateral eye received the other anti-VEGF agent on the next day; thus, if an eye was randomized to the ranibizumab group, the contralateral eye was allocated to the bevacizumab group. All injections were performed using topical proparacaine drops under sterile conditions (eyelid speculum and povidone-iodine). Before the injection was performed, the eyelids were scrubbed with 10% povidone-iodine, and 5% povidone-iodine drops were applied to the conjunctiva. The time between application of 5% povidone-iodine solution to the conjunctiva and administration of the intravitreal injection was 2 minutes. Povidone-iodine was applied to the conjunctiva directly over the intended injection site. Care was taken in all cases to insure that the needle did not touch the lids or lashes. Bevacizumab (1.5 mg/0.06 cc; F. Hoffmann- La Roche Ltd., Basel, Switzerland) or ranibizumab (0.5 mg/0.05 cc; Novartis Pharma Stein AG, Stein, Switzerland) was injected into the vitreous cavity using a 29-gauge 0.5-inch needle inserted through the inferotemporal pars plana 3.0-3.5 mm posterior to the limbus. After the injection, central retinal artery perfusion was confirmed with indirect ophthalmoscopy. Patients were instructed to instill 1 drop of 0.3% ciprofloxacin into the injected eye 4 times daily for 1 week after the procedure.
Retreatment with the originally assigned treatment was performed monthly if central subfield thickness was greater than 275 μm.
If, after 3 consecutive injections, there was not a reduction in central subfield thickness of at least 10% or an increase in BCVA of at least 5 letters compared with baseline, the patient could, at the discretion of the treating ophthalmologist, receive focal/grid laser photocoagulation or continue to receive the same intravitreal medication for an additional 3 consecutive visits.
Follow-Up Examinations and Outcome Measures
Patients were scheduled for follow-up examinations at monthly intervals. At these visits, patients’ BCVA was determined after ETDRS refraction, and they underwent complete ophthalmic examination using the same procedures as at baseline, with the exception of fluorescein angiography, which was performed only at the final follow-up visit. Examiners (E.T., F.P.P.A., R.P.) were masked regarding which treatment drug was used for each patient. Throughout the study, a single masked, certified examiner performed BCVA measurements prior to any other study procedure. Patients, OCT technicians, and fundus photographers were also masked to treatment group. Outcome measures include changes in ETDRS BCVA, changes in central subfield thickness, and occurrence of complications.
BCVA and central subfield thickness measured at each follow-up visit were compared with baseline BCVA and central subfield thickness values for within- and between-group comparisons, which were performed using multiple analysis of variance (MANOVA) for repeated measurements. Proportions of eyes with central subfield thickness ≤275 μm were compared using the likelihood ratio χ 2 test. In addition, a multivariate analysis comparing BCVA and central subfield thickness outcomes in the IV bevacizumab group and IV ranibizumab group was performed, taking into account number of injections, baseline BCVA, and central subfield thickness as effects.
A statistically significant effect was defined if P < .05, and a trend towards significance was reported if P < .1. Statistical analyses were performed using JMP 10.0.0 (2010; SAS Institute Inc, Cary, North Carolina, USA) software.
Sample size and powering were based on a previous clinical trial on bevacizumab use for diabetic macular edema, where a mean change observed in central subfield thickness from baseline was −130 μm with a standard deviation of 122 μm. Therefore, to have 80% power to detect a difference of 50 μm between central subfield thickness change found in both groups, the sample size required in each group was 25 eyes. Thirty eyes per treatment group were required if one assumed a 10% dropout rate. With this sample size, there is a 20% chance for a failure to detect a true mean difference of at least 50 μm between the treatment groups (type I error), or for an incorrect conclusion that a difference of at least 50 μm exists between the treatment groups (type II error).
A total of 48 patients with center-involved DME in at least 1 eye were identified during the study period. Forty-five patients (60 eyes; IV ranibizumab: 28 eyes, IV bevacizumab: 32 eyes) were included in the outcomes analyses; all patients were included in the safety analyses. The 3 patients excluded from the outcomes analyses consisted of 1 patient in the IV ranibizumab group who developed Staphylococcus aureus endophthalmitis after the first injection (this patient chose to exit the study and he did not complete any further study visits); 1 patient in the IV bevacizumab group who developed advanced posterior subcapsular cataract, which precluded adequate SDOCT images, after the ninth follow-up visit; and 1 patient from the IV bevacizumab group who missed 3 consecutive follow-up visits.
Another patient in the IV ranibizumab group developed Streptococcus mitis endophthalmitis after the 44-week study visit, but he completed all study visits and his data were included in the analysis. One patient in the IV bevacizumab group developed transient inferior vitreous hemorrhage attributable to acute posterior vitreous detachment at week 36 and was also maintained in the analysis.
Fifteen patients with bilateral DME received IV ranibizumab in 1 eye and IV bevacizumab in the other eye, and 30 patients received unilateral treatment. Forty percent of eyes (24/60) had proliferative diabetic retinopathy treated with PRP at least 6 months before the initial evaluation. Mean duration of DME estimated by the patients’ reported duration of decreased vision was 37.3 months and 38.1 months in the IV bevacizumab and IV ranibizumab groups, respectively. The time interval between the last anti-VEGF or steroid treatment and study enrollment was at least 6 months. In the bevacizumab group, the number of eyes that had received IV triamcinolone, bevacizumab, or ranibizumab prior to entering the current study was 1, 3, and 2 eyes, respectively; in the ranibizumab group, the number of eyes that had received IV triamcinolone, bevacizumab, or ranibizumab prior to entering the current study was 2, 3, and 2 eyes, respectively. Baseline characteristics are summarized in Table 1 .
|Baseline Characteristics||Bevacizumab Group||Ranibizumab Group|
|Age (years) (mean + SE)||63.8 ± 8.8||63.7 ± 9.0|
|Duration of diabetes (years) (mean + SD)||16.2 ± 8.0||15.9 ± 8.0|
|Treatment regimen: no insulin||13||13|
|Treatment regimen: insulin||19||15|
|HbA1c (mean ± SD)||8.6 ± 1.3||8.7 ± 2.0|
|Systolic blood pressure (mm Hg) (mean ± SD)||139.3 ± 16.5||143.1 ± 20.1|
|Diastolic blood pressure (mm Hg) (mean ± SD)||78.6 ± 11.2||80.5 ± 11.9|
|Grid photocoagulation sessions (n) (mean ± SD)||1.41 ± 0.87||1.51 ± 0.78|
|Moderate or severe nonproliferative diabetic retinopathy||19||17|
|Diabetic retinopathy treated with PRP||13||11|
Best-corrected visual acuity
At baseline, mean BCVA (logMAR) ± standard error (SE) was 0.60 (Snellen equivalent: 20/80) ± 0.05 and 0.63 (Snellen equivalent: 20/85) ± 0.06 in the IV bevacizumab and IV ranibizumab groups, respectively ( P = .680). Intragroup significant improvement in mean BCVA compared with baseline was observed at all study follow-up visits ( P < .05). Maximum mean BCVA improvement occurred at weeks 44 and 48 (−0.23 ± 0.02 logMAR: ∼2.5 ETDRS lines) in the IV bevacizumab group and at week 48 (−0.29 ± 0.04 logMAR: ∼3 ETDRS lines) in the IV ranibizumab group. There was a significantly greater mean improvement in BCVA in the IV ranibizumab group compared with the IV bevacizumab group at weeks 8 ( P = .0318) and 32 ( P = .0415), with a trend towards significance at weeks 28, 36, and 40 ( P < .10) ( Table 2 , and Figure 1 , Top).
With respect to the proportion of eyes losing or gaining ≥10 or ≥15 ETDRS letters, no significant difference between IV bevacizumab and IV ranibizumab groups was observed ( P > .05).
In the IV bevacizumab group, the proportion of eyes losing ≥10 ETDRS letters was 6% at week 16 and from weeks 28-40, and 3% at weeks 12, 20, and 24. The proportion of eyes in the IV bevacizumab group that lost ≥15 letters was 3% at weeks 32 and 36. In the IV ranibizumab group, a loss of ≥10 ETDRS letters was not observed at any follow-up visit. A gain of ≥10 ETDRS letters was observed in 45% and 44% of eyes in the IV bevacizumab and IV ranibizumab groups, respectively, at week 16, and in 61% and 68% in the 2 groups, respectively, at week 48. A gain of ≥15 letters was observed in 15% and 16% of eyes in the IV bevacizumab and IV ranibizumab groups, respectively, at week 16, and in 39% and 48% in the 2 groups, respectively, at week 48 ( Figure 1 , Bottom).
Central subfield thickness
At baseline, mean ± SE central subfield thickness was 451 ± 22 μm and 421 ± 23 μm at baseline in the IV bevacizumab and IV ranibizumab groups, respectively ( P = .4062) ( Figure 2 , Top). Intragroup significant reduction in central subfield thickness compared with baseline was observed at all study follow-up visits ( P < .05). Maximum mean central subfield thickness reduction occurred at week 44 (−136 ± 23 μm) in the IV ranibizumab group and at week 48 (−126 ± 25 μm) in the IV bevacizumab group ( Table 2 , and Figure 2 , Bottom). There was no difference in mean central subfield thickness reduction between the IV bevacizumab and IV ranibizumab groups at any of the study follow-up visits. However, there was a significantly higher proportion of eyes with a central subfield thickness ≤275 μm in the IV ranibizumab group compared with the IV bevacizumab group at weeks 4 ( P = .0029; likelihood ratio), 28 ( P = .0077), 36 ( P = .0028), and 44 ( P = .0292) ( Figure 3 ).