Abstract
Background:
In eyes with diabetic macular edema (DME), the relative efficacy of administering aflibercept monotherapy versus bevacizumab first and then switching to aflibercept if the eye is not improving sufficiently (a form of step therapy) is unknown.
Methods:
At 54 clinical sites, 312 eyes (270 adults) with DME involving the macular center and visual acuity letter score (range, 0 to 100, with higher scores indicating better visual acuity) of 69 to 24 (Snellen equivalent 20/50-20/320) were randomized to 2.0 mg intravitreous aflibercept (158 eyes) or 1.25 mg intravitreous bevacizumab (154 eyes). Beginning at 12 weeks, bevacizumab-group eyes were switched to aflibercept if protocol-specified criteria were met. The primary outcome was mean change in visual acuity over 2 years.
Results:
Over 2 years, 70% of eyes in the bevacizumab-first group switched to aflibercept; mean improvement in visual acuity was 15.0 letters in the aflibercept-monotherapy group and 14.0 letters in the bevacizumab-first group (adjusted difference: +0.8 [95%CI, −0.9 to +2.5] letters; P=0.37). Retinal central subfield thickness and visual acuity mean changes from baseline at 2 years were similar in both groups.
Conclusions:
We found no evidence of difference in visual outcomes over 2 years between treatment with aflibercept monotherapy versus bevacizumab first with switching to aflibercept for suboptimal response. Initiating therapy with bevacizumab and switching to aflibercept as done in this trial is a safe and effective alternative to aflibercept monotherapy in eyes with moderate vision loss due to DME involving the center of the macula.
Trial Registration Number:
Introduction
Diabetic macular edema (DME) is a leading cause of vision loss in working-age adults.1,2 Current standard DME treatment involves intravitreous injections of anti-vascular endothelial growth factor (VEGF) agents into the eye.1 Aflibercept and ranibizumab were FDA-approved for DME treatment and bevacizumab, a lower cost alternative, is used off-label.3,4
These three drugs are effective in improving visual acuity.5 Although a comparative effectiveness study of these drugs showed similar outcomes in patients with baseline visual acuity of 20/40 or better,5 aflibercept treatment resulted in better vision in eyes with baseline visual acuity 20/50 or worse, compared with bevacizumab at two years (mean improvement of 18.1 letters with aflibercept vs 13.3 with bevacizumab), and better vision through 1 year compared with ranibizumab.5 Nonetheless, many eyes with baseline vision of 20/50 or worse in the bevacizumab group had good vision outcomes, with 68% achieving a final visual acuity of 20/40 or better at 2 years.
Because of the substantially higher costs of aflibercept and ranibizumab relative to bevacizumab, an increasing number of insurers require “step therapy,” wherein bevacizumab is used initially with a switch to another anti-VEGF agent if the clinical response is unsatisfactory.6 For eyes with baseline visual acuity 20/50 or worse, it is not known if this strategy compromises long-term visual acuity relative to aflibercept monotherapy.
We conducted a randomized clinical trial to compare two strategies for treating DME involving the center of the macula in eyes with visual impairment (20/50 or worse): aflibercept monotherapy versus bevacizumab first and then switching to aflibercept if the response was suboptimal.
Methods
We conducted Protocol AC, a multi-center, randomized clinical trial at 54 clinical sites in the United States. The manuscript was collaboratively written by the writing committee. The second, fourth and fifth members of the committee analyzed the data and vouch for the accuracy and completeness of the data and analyses, and the first and third member of the committee vouches for the fidelity of the study to the protocol (available with the full text of this article at NEJM.org). The sponsor has confidentiality agreements with the authors, or their institutions named in the credit lines.
The study adhered to the tenets of the Declaration of Helsinki and was approved by a central ethics board. Participants provided written informed consent. An independent data and safety monitoring committee provided oversight. We followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines. (Fig S1)
Study population
Participants were at least 18 years of age and had type 1 or 2 diabetes, at least one eye with best corrected Electronic-Early Treatment Diabetic Retinopathy Study visual acuity (VA) letter score between 69 and 24 (Snellen equivalent of 20/50 to 20/320), center-involved diabetic macular edema (DME) on ophthalmoscopic examination, and central subfield thickness (CST) values greater than optical coherence tomography (OCT) machine- and sex-specific thresholds.7,8 Eyes receiving anti-VEGF treatment for DME in the past 12 months or any DME treatment within the prior 4 months were excluded.
Study Design
Randomization schedules were generated using a permuted block design (random block sizes of 2 or 4) stratified by site. Patients with one study eye were randomly assigned 1:1 to receive 2.0 mg intravitreous aflibercept (Eylea, “aflibercept monotherapy”) or 1.25 mg intravitreous bevacizumab (Avastin) with a switch to 2.0 mg aflibercept if the eye met protocol criteria (“bevacizumab-first”)(Fig. S1). For patients with two study eyes, the eye with the greater VA score was randomly assigned to a treatment group and the other eye to the alternative treatment.
Visits occurred at baseline and every 4 weeks through 1 year, and every 4 to 16 weeks in year 2 depending on disease progression and retreatment. VA with refraction and OCT scans were obtained by certified technicians at each visit. Fundus photographs were collected annually. Study patients were masked to treatment assignment at clinic visits but could become unmasked through billing documents. Technicians at annual visits, central reading center graders, and the medical monitor were masked to treatment assignment. Study investigators were not masked. Adverse events were coded according to the Medical Dictionary for Regulatory Activities.
Treatment Protocol
Eyes were treated at randomization and thereafter according to the prespecified retreatment protocol (Fig S2). Initially, six monthly injections were administered, unless “success” criteria were met starting at month 4 (VA 20/20 or better and CST less than sex- and device-specific threshold). Starting at 24 weeks, the injection was deferred if there was “sustained stability” in VA and CST over two consecutive injections based on real-time data entry by the clinical site for these measures. (An automated algorithm provided treatment instructions.) Injections resumed if eyes worsened.
Beginning at 12 weeks, bevacizumab first eyes were switched to aflibercept if the following criteria were met: 1) persistent center-involved DME (as defined by CST being above the eligibility threshold); 2) an adequately treated eye (bevacizumab injections were given at the previous two consecutive visits); 3) no recent improvement (VA did not improve ≥ 5 letters and CST did not decrease by ≥ 10% compared to or between each of the prior two visits); and 4) vision was suboptimal (visual acuity was 20/50 or worse if before 24 weeks or 20/32 or worse if 24 weeks or later (Fig 1A). Switched eyes received two aflibercept injections, separated by one month, and then aflibercept injections according to the retreatment protocol.
Figure 1.
(A) Criteria for Switching to Aflibercept for Eyes Assigned to the Bevacizumab-First Group. (B)Time from Randomization to Meeting the Switch Criteria Among Eyes Randomly Assigned Bevacizumab-First. Eyes that did not meet the switch criteria were censored at the last completed visit.
Statistical Analysis
The primary outcome was the time-averaged change in VA letter score over 104 weeks, derived by calculating the area under the curve (AUC) over 104 weeks for change in VA from baseline and dividing by the length of follow-up. We calculated that we would require a sample size of 312 eyes to detect a difference between groups, assuming a true difference of at least 3.5 letters in mean change in VA AUC over 2 years, a standard deviation of 8.7 letters with adjustment for baseline VA,5 a type 1 error rate of 5%, a loss-to-follow-up rate of 15%, and a power of approximately 90%.
All analyses followed the intention-to-treat principle. Missing VA values at follow-up visits were imputed with Markov chain-Monte Carlo multiple imputation. Outlying values were truncated to ± 3 standard deviations (SD) from the mean of the VA distribution at 104 weeks (14 of 4685 observed values (0.3%)). The primary analysis of the time-averaged mean score used a linear mixed-effects model with a random intercept to account for the correlation in outcome between 2 eyes of a patient, adjustment for baseline VA and number of study eyes for the patient, and robust variance estimation. Pre-planned subgroup analyses evaluated the effects of baseline CST and VA. Secondary outcomes were compared using linear mixed models or logistic regression with a random intercept term, or an independent t-test (number of visits). Systemic safety outcomes were compared using Fisher’s exact test and ocular safety outcomes were compared using Barnard’s unconditional exact test.
Mean (±SD) or median (interquartile range [IQR]) are reported. All P values and 95% confidence intervals (CI) were 2-sided. As pre-specified, no P values are presented for secondary efficacy outcome measures. No adjustment for multiplicity in sensitivity, subgroup, or safety analyses was implemented. Analyses were performed using SAS/STAT software version 15.1 (SAS Institute Inc).
Results
Study Patients
Between December 8, 2017, and November 25, 2019, 312 eyes from 270 patients were assigned to aflibercept monotherapy (n=158 eyes) or bevacizumab-first (n=154 eyes). Median age was 61 (54-66) years; 48% were females; 53% were white, 26% were Hispanic or Latino, and 19% were Black or African American. Median baseline VA letter score was 60 (65-52; approximate Snellen equivalent 20/63), and median baseline CST was 488 (404-616) μm. Baseline characteristics were similar between groups (Table 1, S1). Excluding deaths, the 2-year completion rate was 88% (226 of 256; Fig S1). The mean number of visits completed over 2 years for patients with one study eye was 22.0±3.6 in the 98 patients in the aflibercept-monotherapy group and 22.5±3.4 in the 94 patients in the bevacizumab-first group (Table 2, S2).
Table 1.
Baseline Patient and Study Eye Characteristics
| Aflibercept- Monotherapy |
Bevacizumab-First | |
|---|---|---|
| (n = 158 eyes of 158 patients) |
(n = 154 eyes of 154 patients) |
|
| Patient characteristics, No. (%) | ||
| No. of study eyes | ||
| 1 | 116 (73%) | 112 (73%) |
| 2 | 42 (27%) | 42 (27%) |
| Sex, female | 76 (48%) | 74 (48%) |
| Age, median (IQR), y | 60 (55, 66) | 61 (54, 67) |
| Self-reported race/ethnicity a | ||
| White | 82 (52%) | 83 (54%) |
| Black or African American | 32 (20%) | 26 (17%) |
| Hispanic or Latino | 39 (25%) | 41 (27%) |
| Asian | 2 (1%) | 2 (1%) |
| Native Hawaiian or Other Pacific Islander | 1 (<1%) | 1 (<1%) |
| American Indian or Alaska Native | 0 | 0 |
| More than one race | 1 (<1%) | 1 (<1%) |
| Unknown or not reported | 1 (<1%) | 0 |
| Type of diabetes | ||
| 1 | 7 (4%) | 8 (5%) |
| 2 | 151 (96%) | 146 (95%) |
| Duration of diabetes, median (IQR), y | 15 (9, 21) | 14 (5, 21) |
| Hemoglobin A1c b, median (IQR), % | 8.0 (7.2, 9.4) | 8.0 (6.8, 9.4) |
| Study eye characteristics, No. (%) | ||
| Visual acuityc, letter score, median (IQR) | 61 (65, 54) | 60 (65, 51) |
| Visual acuity, Snellen equivalent, median (IQR) | 20/63 (20/65, 20/80) | 20/63 (20/50, 20/100) |
| 68 - 54 (Snellen equivalent 20/50 to 20/80) | 119 (75%) | 111 (72%) |
| 53 - 39 (Snellen equivalent 20/100 to 20/160) | 28 (18%) | 31 (20%) |
| 38 - 24 (Snellen equivalent 20/200 to 20/320) | 11 (7%) | 12 (8%) |
| OCT CST d, median (IQR), μm | 478 (401, 579) | 496 (407, 628) |
| Diabetic Retinopathy Severity Scale e | ||
| DR absent or questionable (levels 10, 12, 14, 15) | 0 | 1 (<1%) |
| Microaneurysms only (level 20) | 0 | 2 (1%) |
| Mild to moderate NPDR (levels 35, 43) | 58 (38%) | 42 (29%) |
| Moderately severe to severe NPDR (levels 47, 53) | 72 (48%) | 79 (54%) |
| Mild to moderate PDR (levels 61, 65) | 15 (10%) | 16 (11%) |
| High-risk PDR (levels 71, 75) | 6 (4%) | 7 (5%) |
| Prior Anti-VEGF Treatment for DME | 25 (16%) | 29 (19%) |
Abbreviations: CST, central subfield thickness; DME, diabetic macular edema; IQR, interquartile range; NPDR, nonproliferative diabetic retinopathyOCT, optical coherence tomography; PDR, proliferative diabetic retinopathy; VEGF, vascular endothelial growth factor.
Patient-reported race/ethnicity were collected based on fixed categories.15
Unavailable for 10 eyes in each group
Measured with the Electronic-Early Treatment Diabetic Retinopathy Study visual acuity test on a scale from 100 letters (Snellen equivalent to 20/10) to 0 letter (Snellen equivalent of <20/800); higher scores indicate better vision. The data presented are the best-corrected visual acuity in the study eye following protocol-defined refraction.
Cirrus measurements were converted to Spectralis equivalents using the following formula: Spectralis = 40.78 + 0.95 × Cirrus.14 Unavailable for 2 eyes in the aflibercept-monotherapy group. The Reading Center confirmed the presence of DME in these eyes.
Level based on reading center assessment. Unavailable for 2 eyes in the aflibercept-monotherapy group and 1 eye in the bevacizumab-first group. Ungradable for 7 eyes in the aflibercept-monotherapy group and 6 eyes in the bevacizumab-first group. Percentage is based on eyes with gradable baseline DR severity.
Table 2.
Study Outcomes Through 2 Years of Follow-up
| Outcomes, N (%) | Aflibercept- Monotherapy Group (N = 132) |
Bevacizumab- First Group (N = 128) |
Adjusted Difference (95% CI) |
P Value |
|---|---|---|---|---|
| Visual acuity letter score a | ||||
| Primary Outcome: Area under the curve over 2 y | ||||
| Mean (SD) | 15.0 (8.5) | 14.0 (8.8) | +0.8 (−0.9, +2.5) | 0.37 |
| Median (IQR) | 14.8 (9.2, 20.9) | 13.2 (8.3, 18.7) | -- | -- |
| Visual acuity at 2 y | ||||
| Letter score, mean (SD) | 72 (15) | 73 (13) | -- | -- |
| Snellen equivalent, mean | 20/40 | 20/40 | -- | -- |
| Eyes with visual acuity 20/20 or greater | 29 (22%) | 28 (22%) | −1% (−11%, +8%) | -- |
| Eyes with visual acuity 20/40 or greater | 96 (73%) | 95 (74%) | −7% (−19%, +5%) | -- |
| Eyes with visual acuity 20/200 or worse | 6 (5%) | 2 (2%) | +3% (−1%, +7%) | -- |
| Change from baseline in letter score to 2 y | ||||
| Mean (SD) | 14.7 (14.5) | 15.9 (12.4) | −1.8 (−4.9, +1.2) | -- |
| Eyes improving ≥ 15 letters | 70 (53%) | 74 (58%) | −7% (−21%, +8%) | -- |
| Eyes improving ≥ 10 letters | 101 (77%) | 98 (77%) | −3% (−13%, +8%) | -- |
| Eyes worsening ≥ 10 letters | 9 (7%) | 5 (4%) | +4% (−2%, +9%) | -- |
| Eyes worsening ≥ 15 letters | 8 (6%) | 4 (3%) | +3% (−1%, +8%) | -- |
| OCT central subfield thickness b, μm | ||||
| At 2 y, mean (SD) | N = 126 315 (93) | N = 127 336 (100) | -- | -- |
| Eyes with OCT CST below gender-specific OCT threshold for CI-DME at 2 y c | 76 (60%) | 70 (55%) | +4% (−12%, +20%) | -- |
| Change from baseline to 2 y, mean (SD) | −192 (143) | −198 (160) | −16 (−39, +7) | -- |
| Number of study visits completed d | ||||
| Up to 2 y, mean (SD) | N = 98 22.0 (3.6) | N = 94 22.5 (3.4) | −0.5 (−1.5, 0.5) | -- |
| Number of study injections administered e | ||||
| Up to 24 wks, mean (SD) | N = 148 5.7 (0.7) | N = 146 5.8 (0.5) | -- | -- |
| Between 24 wks and <1 y, mean (SD) | N = 144 3.7 (1.8) | N = 142 4.2 (1.8) | -- | -- |
| Up to 1 y, mean (SD) | N = 144 9.3 (2.1) | N = 142 10.0 (2.0) | −0.6 (−1.0, −0.1) | -- |
| Between 1 y and <2 y, mean (SD) | 5.3 (2.8) | 6.1 (3.1) | -- | -- |
| Up to 2 y, mean (SD) | 14.6 (4.1) | 16.1 (4.1) | −1.5 (−2.4, −0.5) | -- |
Abbreviations: IQR, interquartile range; NPDR, nonproliferative diabetic retinopathy; OCT, optical coherence tomography; PDR, proliferative diabetic retinopathy; PRP, panretinal photocoagulation.
Measured with the Electronic-Early Treatment Diabetic Retinopathy Study visual acuity test on a scale from 100 letters (Snellen equivalent to 20/10) to 0 letter (Snellen equivalent of <20/800); higher scores indicate better vision. The data presented are the best-corrected visual acuity in the study eye following protocol-defined refraction. Missing data in visual acuity measurements during follow-up were imputed using multiple imputation for eyes not completing the 2-year visit or where data are otherwise noted as missing, assuming data were missing at random. Outlying values were truncated to ± 3 standard deviations from the observed 104-week mean. Area under the curve outcome was calculated after multiple imputation and data truncation were applied, using all observed and imputed data points for each patient. Continuous outcomes were compared between treatment groups using a linear mixed effects model with a random intercept term (for the correlation between two study eyes from the same patient) and robust variance estimation, adjusting for baseline visual acuity and number of study eyes. Binary outcomes were compared using logistic regression with a random intercept (for the correlation between two study eyes from the same patients), conditional standardization, and the delta method (for estimation of risk difference), adjusting for baseline visual acuity and number of study eyes. Adjustment for baseline visual acuity, number of study eyes, and the correlation between the two study eyes from the same patient were removed from the model only when there was a convergence issue.
Cirrus measurements were converted to Spectralis equivalents using the following formula: Spectralis = 40.78 + 0.95 × Cirrus.14 Only include eyes with available baseline CST. 104-week CST was unavailable for 5 eyes in the aflibercept monotherapy group and 1 eye in the bevacizumab first group. Missing data were imputed using multiple imputation. Treatment group comparison was performed using a linear mixed effects model with a random intercept term for patients and robust variance estimation, adjusting for baseline CST and number of study eyes.
For Zeiss Cirrus, ≥ 290 μm for females and ≥ 305 μm for males. For Heidelberg Spectralis, ≥ 305 μm for females and ≥ 320 μm for males.7,8 No imputation for this outcome because the threshold values are machine specific and not Spectralis equivalents, which are being imputed. Treatment group comparison was performed using logistic regression with a random intercept (for the correlation between two study eyes from the same patients), conditional standardization, and the delta method (for estimation of risk difference).
Limited to only unilateral patients (i.e., patients with one study eye) who completed 2-year of follow-up.
All intervals are closed on the left and open on the right. Limited to eyes that completed the visit or any later common visit. All study injections were counted, including aflibercept injections received among eyes in the bevacizumab first group. Estimated differences were adjusted for number of study eyes and the correlation between two study eyes from the same patients.
Study Treatments
Over 2 years, eyes in the aflibercept-monotherapy group received 14.6±4.1 injections on average while eyes in the bevacizumab-first group received 16.1±4.1 (adjusted difference [95% CI]: −1.5 [−2.4, −0.5]; Table 2). Eyes in the bevacizumab-first group received an average of 9.2±5.2 bevacizumab injections and 6.9±5.8 aflibercept injections over 2 years. Overall, 153 (97%) eyes in the aflibercept-monotherapy group and 147 (95%) eyes in the bevacizumab-first group were compliant with study injections at all visits (Table S3). The cumulative proportion of eyes in the bevacizumab-first group switched to aflibercept was 39% (95% CI, 32%-47%) up to 24 weeks, 60% (95% CI, 53%-68%) up to 52 weeks, and 70% (95% CI, 62%-77%) over 2 years (Fig 1B). The cumulative proportion of eyes in the aflibercept-monotherapy group that met the criteria (but continued aflibercept treatment as prespecified) was 13% (95% CI, 8.8%-20%) up to 24 weeks, 26% (95% CI, 20%-34%) up to 52 weeks, and 30% (95% CI, 23%-38%) over 2 years (Fig S3). Upon completing the 2-year visit, patients with one study eye were asked whether they believed they knew which drugs they had received, and if they did, to indicate which drug; 107 of 188 patients (57%) did not know or refused to answer; and 46 of the 80 (58%) patients who provided a full response selected the correct study-arm.
Primary Outcome
The mean change in VA from baseline over 2 years (AUC) was 15.0±8.5 letters in the aflibercept-monotherapy group and 14.0±8.8 letters in the bevacizumab-first group (adjusted difference: +0.8 [95% CI, −0.9 to +2.5]; P=0.37; Table 2 and Fig 2A). Sensitivity analyses, including a comparison involving only the 42 patients with one eye assigned to each treatment group, yielded results consistent with the primary analysis (Table S4). The adjusted mean letter score difference was −1.6 [95% CI, −4.4 to +1.2] for eyes with baseline CST < 400 μm and +2.4 [95% CI, +0.2 to +4.7] for eyes with baseline CST ≥ 400 μm (P=0.03 for interaction). We found no evidence of interaction effects for other pre-specified subgroups (Table S5).
Figure 2.
Mean Change in Visual Acuity Through 2 Years (A); Mean Change in Central Subfield Thickness from Baseline Through 2 Years (B). Error bars represent 95% confidence intervals. Visual acuity was measured with the Electronic-Early Treatment Diabetic Retinopathy Study visual acuity test on a scale from 100 letters (Snellen equivalent to 20/10) to 0 letter (Snellen equivalent of <20/800)13; higher scores indicate better vision. The data presented are the best-corrected visual acuity in the study eye following protocol-defined refraction. Central subfield thickness was measured from optical coherence tomography scans. Cirrus measurements were converted to Spectralis equivalents using the following formula: Spectralis = 40.78 + 0.95 × Cirrus.14
Secondary Outcomes
At 2 years, mean change in VA from baseline was 14.7±14.5 letters in the aflibercept -monotherapy group (N=132) and 15.9±12.4 letters in the bevacizumab-first group (N=128; adjusted difference: −1.8 [95% CI, −4.9 to +1.2]; Table 2). In each group, 77% of eyes had visual acuity improvement of ≥10 letters (adjusted difference: −3% [95% CI, −13% to +8%]), and 22% had visual acuity of 20/20 or better (adjusted difference: −1% [95% CI, −11% to +8%]). Visual acuity of 20/40 or better at 2 years was found in 73% of the aflibercept-monotherapy group and 74% of the bevacizumab-first group (adjusted difference: −7% [95% CI, −19% to +5%]).
Mean change in CST from baseline to 2 years was −192±143 μm in the aflibercept-monotherapy group (N=126), and −198±160 μm in the bevacizumab-first group (N=127; adjusted difference: −16 [95% CI, −39 to +7]; Table 2 and Figure 2B). Although greater reduction in CST was observed in the aflibercept-monotherapy group at 24 and 52 weeks (Table S6), the percentages of eyes at 2 years with CST below DME thresholds were similar: 60% with aflibercept-monotherapy and 55% with bevacizumab-first (adjusted difference: +4% [95% CI, −12% to +20%]).
At 2 years, a ≥2-step improvement in diabetic retinopathy severity occurred in 53 (50%) of 105 eyes in the aflibercept-monotherapy group and in 60 (56%) of 108 eyes in the bevacizumab-first group (adjusted difference: −3% [95% CI, −23% to +17%]). Few eyes experienced ≥2-step worsening (4% in both groups; adjusted difference: 0% [95% CI, −5% to +5%]; Table S2).
Adverse Events
One eye in the aflibercept-monotherapy group developed endophthalmitis. Prespecified ocular adverse events were infrequent in both treatment groups (Table 3). At least one serious systemic adverse event occurred in 60 of 116 patients in the aflibercept-monotherapy group (52%), 40 of 112 in the bevacizumab-first group (36%), and 18 of 42 in the group of patients with two study eyes (43%; P=0.05; Table 3). Occurrence of any Antiplatelet Trialists’ Collaboration adverse event was 12% in the aflibercept-monotherapy group, 8% in the bevacizumab-first group, and 7% in patients with two study eyes; (P=0.57; Table 3). All adverse events are listed in Tables S7-11.
Table 3.
Systemic and Ocular Adverse Events of Special Interest Through 2 Years of Follow-up
| Events of special interest through 2 y a | Aflibercept Monotherapy Group (N = 116 patients) |
Bevacizumab First Group (N = 112 patients) |
Bilateral (N = 42 patients) |
P Value |
|---|---|---|---|---|
| Systemic adverse events, No. (%) b | ||||
| Vascular events defined by APTC criteria c | ||||
| Nonfatal myocardial infarction | 3 (3%) | 1 (<1%) | 0 | 0.52 |
| Nonfatal stroke | 3 (3%) | 6 (5%) | 1 (2%) | 0.56 |
| Death due to potential vascular cause or unknown cause | 8 (7%) | 2 (2%) | 2 (5%) | 0.14 |
| Any event | 14 (12%) | 9 (8%) | 3 (7%) | 0.57 |
| Prespecified events | ||||
| Death from any cause | 10 (9%) | 4 (4%) | 3 (7%) | 0.28 |
| Hospitalization | 56 (48%) | 36 (32%) | 18 (43%) | 0.04 d |
| Serious adverse event | 60 (52%) | 40 (36%) | 18 (43%) | 0.05 |
| Hypertension | 19 (16%) | 10 (9%) | 11 (26%) | 0.02 e |
| Events of special interest through 2 y a | Aflibercept Monotherapy Group (N = 158 eyes) |
Bevacizumab First Group (N = 154 eyes) |
-- | P Value |
| Ocular adverse events, No. (%) f | ||||
| Endophthalmitis | 1 (<1%) | 0 | -- | 0.52 |
| Any retinal detachment | 0 | 3 (2%) | -- | 0.08 |
| Rhegmatogenous retinal detachment | 0 | 2 (1%) | -- | -- |
| Traction retinal detachment | 0 | 1 (<1%) | -- | -- |
| Traumatic cataract | 0 | 0 | -- | -- g |
| Vitreous hemorrhage | 7 (4%) | 10 (6%) | -- | 0.53 |
| Ocular Inflammation h | 3 (2%) | 1 (<1%) | -- | 0.53 |
| Intraocular pressure elevation i | 15 (9%) | 14 (9%) | -- | 0.92 |
| Increase in IOP ≥10 mmHg from baseline | 10 (6%) | 6 (4%) | -- | -- |
| IOP ≥30 mmHg | 1 (<1%) | 1 (<1%) | -- | -- |
| Initiation of medication to lower IOP that was not in use at baseline | 7 (4%) | 9 (6%) | -- | -- |
| Glaucoma procedure | 0 | 0 | -- | -- |
| Neovascularization of the iris | 2 (1%) | 0 | -- | 0.21 |
| Neovascular glaucoma | 1 (<1%) | 0 | -- | 0.52 |
Abbreviations: APTC, Antiplatelet Trialists’ Collaboration; IOP, intraocular pressure.
Unless otherwise specified adverse events were collected at any time during study follow-up. Some events may have been reported more than once (i.e., at multiple visits). Including patients/ study eyes that experienced at least one adverse event of interest. All adverse events were patient-reported, except for IOP events.
The proportion of patients with each systemic adverse event was compared among groups using Fisher’s exact test. Pairwise comparisons between groups were conducted only when the overall test has P ≤ 0.05, using Fisher’s exact test without adjustment for multiple comparisons.
Vascular events were defined according to the criteria of the Antiplatelet Trialists’ Collaboration.16
P values = 0.02 for aflibercept monotherapy vs. bevacizumab first, 0.59 for aflibercept monotherapy vs. bilateral, and 0.26 for bevacizumab first vs. bilateral
P values = 0.11 for aflibercept monotherapy vs. bevacizumab first, 0.17 for aflibercept monotherapy vs. bilateral, and 0.009 for bevacizumab first vs. bilateral.
The proportion of eyes experiencing each outcome was compared between treatment groups using Barnard’s unconditional exact test, considering the number of eyes in each treatment group fixed.
Treatment group comparison was not performed because no event occurred.
Inflammation included the presence of inflammatory cells or flare in the anterior chamber, pseudohypopyon, chorioretinitis, choroiditis, pseudoendophthalmitis, iritis, uveitis, and the presence of vitreal cells.
Intraocular pressure elevation was defined as an increase in intraocular pressure of 10 mmHg or more from baseline at any visit, an intraocular pressure of 30 mmHg or more at any visit, the initiation of medication to lower intraocular pressure that was not in use at baseline, or glaucoma procedure.
Discussion
In this randomized clinical trial of eyes with center-involved DME and moderately impaired vision of 20/50 or worse, there was no overall difference in mean change in vision over a 2-year period whether administering aflibercept monotherapy or giving bevacizumab first and switching to aflibercept based on prespecified criteria for suboptimal response. There was also no difference in retinal thickness between the groups at 2 years. About 70% of patients in the bevacizumab-first group were switched to aflibercept during the 2-year study. More than half (57%) of these patients switched between 12 and 24 weeks. These results expand our understanding of how and whether clinical outcomes differ using different anti-VEGF treatment strategies for DME.
Previously, DRCR Retina Network Protocol T demonstrated better visual and retinal thickness outcomes in eyes with visual acuity worse than 20/50 that were assigned to aflibercept versus bevacizumab monotherapy.5 Unsurprisingly, given these findings, visual acuity and retinal thickness outcomes in the current study appeared to favor aflibercept-monotherapy over bevacizumab-first throughout the first year. However, rescue treatment with aflibercept mitigated the average visual and anatomic differences that arose from initiating therapy with bevacizumab versus aflibercept. A post hoc sensitivity analysis of study eyes from the subset of participants who were treated in one eye with aflibercept-monotherapy and the other eye with bevacizumab-first did not identify significant differences in visual outcomes between the study eyes. The study cohort had higher proportions of minority participants, in the aggregate, compared to those of the US population (Table S12), but the proportions are similar enough that results are probably generalizable. Black persons are disproportionately affected by DME and over-represented in the study. A limitation, however, is an underrepresentation of Asian Americans and no representation of Native Americans or Alaskan natives.
In a pre-specified secondary analysis of the subgroup of eyes with baseline CST ≥400 μm, eyes treated with aflibercept monotherapy had better mean visual acuity over 2 years (+2.4 [95% CI, +0.2 to +4.7]). However, a difference of this magnitude, especially when the upper end of the 95% confidence interval does not include 5 letters, is of marginal clinical relevance.
Findings from this study are particularly relevant given the increasing frequency of insurers mandating “step therapy” with use of bevacizumab before use of other drugs approved by the US Food and Drug Administration (FDA). However, results from this study can be generalized only to patients who receive therapy according to the same switching criteria, anti-VEGF agents, and retreatment algorithm as used in this trial. Given the ~26-fold difference in cost between the drugs (average Medicare reimbursement is approximately $1830 USD for one dose of aflibercept and $70 USD for one dose of bevacizumab),9,10 initiating treatment with bevacizumab could have substantial cost reductions for the healthcare system.
There was a higher frequency of serious systemic adverse events in the aflibercept-monotherapy (52%) versus bevacizumab-first (36%) group; however, previous trials of aflibercept monotherapy for DME have demonstrated similar event rates compared to sham treatment and to bevacizumab monotherapy.5,11,12 Also, 70% of bevacizumab-first patients received aflibercept during this study, making assessment of differences in serious systemic adverse events difficult to interpret.
This study has limitations. First, it is unknown whether milder or stricter switching criteria would have led to different results. Second, although efforts were made to keep patients masked to their treatment assignment, the cost of aflibercept in this study was generally billed to the patients’ insurance when applicable: unmasking could occur if patients viewed billing information. However, when asked about their treatment assignment, 57% did not know or refused to answer and of the remainder, 43% of those who guessed, did so incorrectly. Third, besides aflibercept, this study did not include other anti-VEGF agents approved by US FDA for the treatment of DME.
In this trial, we found no evidence of difference in visual outcomes over 2 years in eyes treated with aflibercept monotherapy compared with eyes treated with bevacizumab first. Initiating therapy with bevacizumab and switching to aflibercept as done in this trial is a safe and effective alternative to aflibercept monotherapy in eyes with moderate vision loss due to center-involved DME.
Supplementary Material
Funding Support:
Research reported in this publication was supported by the National Eye Institute and the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Number UG1EY014231. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The DRCR Retina Network had complete control over the design of the protocol, ownership of the data, all editorial content of presentations and publications related to the protocol, and the decision to submit for publication
Role of the Sponsor:
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 decision to submit for publication or the preparation of the manuscript.
Contributor Information
Chirag D. Jhaveri, Retina Consultants of Austin; Austin Research Center for Retina (Austin TX).
Adam R. Glassman, Jaeb Center for Health Research, Tampa, FL.
Frederick L. Ferris, III, Ophthalmic Research Consultants, LLC. Waxhaw, NC. USA.
Danni Liu, Jaeb Center for Health Research, Tampa, FL.
Maureen G. Maguire, Jaeb Center for Health Research, Tampa, FL.
John B. Allen, Southeast Clinical Research Associates, LLC. Charlotte, NC.
Carl W. Baker, The Ophthalmology Group, LLC Paducah, KY.
David Browning, Charlotte Eye, Ear, Nose, and Throat Associates, Charlotte, NC.
Matthew A. Cunningham, Florida Retina Institute, Orlando, FL.
Scott M. Friedman, Florida Retina Consultants, Lakeland, FL.
Lee M. Jampol, Feinberg School of Medicine, Northwestern University, Chicago, Illinois.
Dennis M. Marcus, Southeast Retina Center, P.C. Augusta, GA.
Daniel F. Martin, Cole Eye Institute, Cleveland Clinic, Cleveland, Ohio.
Carin M. Preston, Jaeb Center for Health Research, Tampa, FL.
Cynthia R. Stockdale, Jaeb Center for Health Research, Tampa, FL.
DRCR Retina Network, Joslin Diabetes Center, Beetham Eye Institute, Harvard Department of Ophthalmology, Boston, MA.
References
- 1.Flaxel CJ, Adelman RA, Bailey ST, et al. Diabetic Retinopathy Preferred Practice Pattern(R). Ophthalmology. 2020;127(1):P66–P145. [DOI] [PubMed] [Google Scholar]
- 2.Lee R, Wong TY, Sabanayagam C. Epidemiology of diabetic retinopathy, diabetic macular edema and related vision loss. Eye Vis (Lond). 2015;2:17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Nguyen QD, Brown DM, Marcus DM, et al. Ranibizumab for diabetic macular edema: Results from 2 phase III randomized trials: RISE and RIDE. Ophthalmology. 2012;119(4):789–801. [DOI] [PubMed] [Google Scholar]
- 4.Brown DM, Schmidt-Erfurth U, Do DV, et al. Intravitreal aflibercept for diabetic macular edema: 100-week results from the VISTA and VIVID studies. Ophthalmology. 2015;122(10):2044–2052. [DOI] [PubMed] [Google Scholar]
- 5.Wells JA, Glassman AR, Ayala AR, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema: two-year results from a comparative effectiveness randomized clinical trial. Ophthalmology. 2016;123(6):1351–1359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sachs RE, Kyle MA. Step Therapy's Balancing Act - Protecting Patients while Addressing High Drug Prices. N Engl J Med. 2022;386(10):901–904. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chalam KV, Bressler SB, Edwards AR, et al. Retinal thickness in people with diabetes and minimal or no diabetic retinopathy: Heidelberg Spectralis optical coherence tomography. Invest Ophthalmol Vis Sci. 2012;53(13):8154–8161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Writing Committee for the Diabetic Retinopathy Clinical Research Network, Gross JG, Glassman AR, et al. Panretinal photocoagulation vs intravitreous ranibizumab for proliferative diabetic retinopathy: A randomized clinical trial. JAMA. 2015;314(20):2137–2146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Centers for Medicare and Medicaid Services. 2022 ASP Drug Pricing Files. https://www.cms.gov/medicare/medicare-part-b-drug-average-sales-price/2022-asp-drug-pricing-files. Accessed Accessed 24 February 2022.
- 10.Centers for Medicare & Medicaid Services. Medicare advantage prior authorization and step therapy for part B drugs. 2018; https://www.cms.gov/newsroom/fact-sheets/medicare-advantage-prior-authorization-and-step-therapy-part-b-drugs. Accessed Accessed 24 February 2022.
- 11.Diabetic Retinopathy Clinical Research Network, Wells JA, Glassman AR, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema. N Engl J Med. 2015;372(13):1193–1203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Heier JS, Korobelnik JF, Brown DM, et al. Intravitreal Aflibercept for Diabetic Macular Edema: 148-Week Results from the VISTA and VIVID Studies. Ophthalmology. 2016;123(11):2376–2385. [DOI] [PubMed] [Google Scholar]
- 13.Beck RW, Moke PS, Turpin AH, et al. A computerized method of visual acuity testing: adaptation of the Early Treatment of Diabetic Retinopathy Study testing protocol. Am J Ophthalmol. 2003;135(2):194–205. [DOI] [PubMed] [Google Scholar]
- 14.Sun JK, Josic K, Melia M, et al. Conversion of Central Subfield Thickness Measurements of Diabetic Macular Edema Across Cirrus and Spectralis Optical Coherence Tomography Instruments. Transl Vis Sci Technol. 2021;10(14):34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.National Institutes of Health. Racial and ethnic categories and definitions for NIH diversity programs and for other reporting purposes. https://grants.nih.gov/grants/guide/notice-files/not-od-15-089.html. Accessed Accessed 21 March 2022.
- 16.Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of antiplatelet therapy--I: Prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ. 1994;308(6921):81–106. [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.