Abstract
Purpose:
In individuals aged >50 years, age-related macular degeneration (AMD) is the leading cause of irreversible blindness. Intravitreal injections of antivascular endothelial growth factor (VEGF) agents (bevacizumab, ranibizumab, and aflibercept) show good efficacy and similar incidences of systemic adverse events (SAEs). However, comparative studies between agents are limited. Our study aimed to compare the real-world SAE risks of bevacizumab, ranibizumab, and aflibercept users.
Methods:
This retrospective cohort study identified new bevacizumab, ranibizumab, and aflibercept users in a multi-institutional database in Taiwan between 2014 and 2019. Inverse probability of treatment weights (IPTW) with propensity scores was conducted to achieve homogeneity among groups. The Fine and Gray model was utilized to estimate the subdistribution hazard ratio and 95% confidence interval.
Results:
This study included 701 bevacizumab, 463 ranibizumab, and 984 aflibercept users. After IPTW, all covariates were well-balanced. All three anti-VEGF agents had a low and comparable number per 100 person-years of major adverse cardiac events, heart failure, thromboembolic events, major bleeding, all-cause admission, and all-cause death (all P > 0.05). No significant differences in long-term change of systolic and diastolic blood pressure, low-density lipoprotein, estimated glomerular filtration rate, and alanine transaminase (all P for interaction > 0.05) were observed among groups.
Conclusion:
Bevacizumab, ranibizumab, and aflibercept had a good systemic safety profile in this study. All groups showed a low and similar SAE risk and no differences in their long-term change of laboratory data. Therefore, these anti-VEGF agents could be prescribed safely to patients with AMD.
Keywords: Aflibercept, age-related macular degeneration, bevacizumab, ranibizumab, systemic adverse event
Age-related macular degeneration (AMD) is the leading cause of irreversible blindness in individuals aged >50 years.[1] AMD impairs vision by upregulating the vascular endothelial growth factor (VEGF) and inducing angiogenesis. Anti-VEGF agents can counter this upregulation to prevent pathological neovascularization and reduce vascular permeability.[2] Intravitreal injection (IVI) of anti-VEGF agents has transformed the standard care for AMD. The three most common anti-VEGF agents are aflibercept, bevacizumab, and ranibizumab. Bevacizumab, originally used to treat breast and colorectal cancer, is a recombinant humanized monoclonal antibody against VEGF-A.[3] With a beneficial cost-effectiveness ratio, bevacizumab is broadly used off-label as an IVI for AMD after extemporaneous compounding. Ranibizumab, the first agent indicated for AMD, is a recombinant humanized monoclonal antibody fragment against VEGF-A.[4] Aflibercept, the newest agent approved for AMD, is a fusion protein of human VEGF receptors 1 and 2 that binds to the fragment crystallizable part of human immunoglobulin-G1 antibodies against placental growth factor, VEGF-A, and VEGF-B.[5] While the benefits for AMD are notable, critical issues remain on the systemic safety of these anti-VEGF agents.
Given their extensive use, the safety profiles of anti-VEGF agents shall be established. AMD is treated effectively by IVI, which, in combination with the blood-ocular barrier, ensures that anti-VEGF agents are provided at sufficient concentration and duration.[6] While the dosage of anti-VEGF agents in IVI is much smaller than that of intravenous infusions used for chemotherapy, systemic diffusion may still occur.[7] Furthermore, repeated injections are often necessary as AMD is a chronic condition. Therefore, an accumulated dosage could affect metabolic control and cause systemic adverse events (SAEs) over time.[8] In oncology, when anti-VEGF agents were used during intravenous chemotherapy, reported SAEs included arterial thromboembolism, hypertension, and proteinuria.[9,10,11] In contrast, anti-VEGF agents used in IVI have shown good systemic safety profiles in most published investigations.[12,13] However, limited studies have compared SAE risks between patients receiving different agents.
Each anti-VEGF agent has a distinct structure and pharmacokinetics, suggesting that the risks of different SAEs may vary.[7,14] High systemic exposure to bevacizumab significantly reduces the serum VEGF concentration.[15] In addition, aflibercept has a fast and long binding action on the VEGF receptor due to its high affinity.[16] The risks of SAEs may increase given the propensity for these anti-VEGFs to have a longer duration and higher concentration. Nevertheless, these anti-VEGFs have each shown favorable systemic safety outcomes compared with placebo in multiple meta-analyses of randomized controlled trials (RCTs).[17,18,19] Furthermore, RCTs that have compared the anti-VEGF agents have reported similar systemic safety profiles.[20,21] However, differences in therapeutic protocols and selected populations limit the application of these outcomes. The RCTs were found to be underpowered to identify significant differences among patients receiving the different anti-VEGF agents. Moreover, there are substantially different population characteristics and treatment patterns in administering anti-VEGF agents in real-world situations and RCTs. Observational studies in real-world settings remain limited, and most compare only two of the three anti-VEGF agents.[22,23,24] Therefore, this study aimed to investigate SAEs among the three common anti-VEGF agents in a multi-institutional database in Taiwan.
Methods
Data collection
This retrospective cohort study analyzed data from a multi-institutional database in Taiwan, which comprised 1.3 million people in seven institutes.[25] The database contains patients’ clinical diagnoses, medication usage, operation records, laboratory data, and imaging reports. Notably, the self-paid intervention, which is beyond the coverage of the Taiwan National Health Insurance program, could be identified by its payment code. The validation of this database was demonstrated in the previous publication.[25] Before 2015 and after 2016, AMD was diagnosed based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and ICD-10-CM diagnostic codes, respectively. One inpatient department diagnosis or two outpatient department diagnoses indicated the occurrence of comorbidities and outcomes. The protocol of this study complied with the Declaration of Helsinki. Due to the utilization of deidentified data, written informed consent was waived. The institutional review board approved this study (IRB No.: 202200606B1).
Patient inclusion
Patients diagnosed with AMD and receiving IVI of bevacizumab, ranibizumab, or aflibercept between January 1, 2014, and December 31, 2019, were collected. The first anti-VEGF agent utilized for IVI, bevacizumab, was in 2014. Therefore, the starting time at which we collected the data was 2014. We applied a new-user design to minimize bias.[26] In addition, an active-comparator design was used based on the three agents’ similar indications. The day of the first prescription of IVI of anti-VEGF agents served as the index date. This study excluded patients aged <50 years, diagnosed with diabetes mellitus, with other indications for IVI of anti-VEGF agents, prior utilization of studied drugs, history of corticosteroid IVI use, or a pre-existing malignancy. The remaining patients were followed until their primary outcome, last visit in the database, switch among studied agents, death, or December 31, 2019. According to the National Health Insurance guideline in Taiwan, patients aged more than 50 years are eligible for the application of anti-VEGF agents. Therefore, the cut-off age is set at 50 years for patients with AMD.
Outcome definition
This study's outcomes contained clinical events and laboratory data. Primary outcomes were SAEs among three anti-VEGF agents. These SAEs contained major adverse cardiac events (MACE), heart failure (HF), thromboembolic events, major bleeding, all-cause admission, and all-cause death, and overall complications indicated any one of the above events. Taiwan Death Registry provided the cause, place, and date of death, which were presented in this study. MACE indicated the composite of myocardial infarction, ischemic stroke, and cardiovascular death. Thromboembolic events represented the composite of myocardial infarction, ischemic stroke, transient ischemic attack, extremity thromboembolism, and systemic thromboembolism. The laboratory data, including systolic blood pressure (SBP), diastolic blood pressure (DBP), low-density lipoprotein (LDL), estimated glomerular filtration rate (eGFR), and alanine transaminase (ALT), were retrieved every 6 months.
Covariate assessment
Covariates included demographics, comorbidities, medications, laboratory data, which eye was injected, and the number of visits to the outpatient department of ophthalmology last year. Demographics comprised age, sex, body mass index (BMI), smoking, and alcohol. Comorbidities collected within 6 months before the index date included hypertension, dyslipidemia, myocardial infarction, ischemic stroke, HF, and chronic kidney disease. Ocular history comprised cataracts, glaucoma, retinal laser, vitrectomy, and the number of visits to the outpatient department of ophthalmology last year; the disease burden was assessed according to Charlson's comorbidity index scores.[27] Medications collected within 6 months before the index date included antihypertensive and other medications. The values of SBP, DBP, LDL, eGFR, and ALT measured on the day closest to the index date indicated the baseline.
Statistical analysis
We utilized inverse probability of treatment weights (IPTW) with propensity scores to create an additional adjustment cohort that compared the risk of SAEs among the three groups. Based on 50,000 regression trees, the generalized boosted model was used to estimate the propensity score because this study contained more than two groups. Through multivariable logistic regression, the covariates were calculated for propensity scores. We truncated the weight at the 97th percentile and utilized stabilized weight to avoid the impact of the extreme outlier weight.[28] Maximum absolute standardized differences (MASDs) <0.10 and >0.20 indicated negligible and substantial differences, respectively. The covariates with MASD > 0.1 would be adjusted with the multivariable analysis.
The incidence of events was presented as their number per 100 person-years. Death was regarded as a competing risk in this study; thus, the Fine and Gray subdistribution hazard model was applied to compare the incidence of non-fatal outcomes. In addition, we adopted the Cox proportional hazard model to assess the rate of fatal outcomes. A linear mixed model was utilized to evaluate the change in laboratory data. Single expectation maximization was used to input missing laboratory data. In the survival analysis, the study drug was the only explanatory variable. A two-sided P value of <0.05 indicated statistical significance. The analyses in our study were performed with the SAS (version 9.4; SAS Institute Inc., Cary, NC, USA).
Results
Patient enrollment and baseline characteristics
This study collected 5616 patients with AMD receiving IVIs of anti-VEGF agents between and 2019 [Fig. 1]. With the exclusion of unqualified patients, our analysis included 701 bevacizumab, 463 ranibizumab, and 984 aflibercept users. Before IPTW, males accounted for most patients in the aflibercept group (bevacizumab (58.1%) vs. ranibizumab (60.5%) vs. aflibercept (62.9%), MASD = 0.10; Table 1 and Supplementary Table 1). Bevacizumab users were the oldest (71.1 ± 10.8 years vs. 70.3 ± 10.6 years vs. 70.1 ± 9.7 years, MASD = 0.10). The three groups had similar Charlson's comorbidity index scores and prevalence of systemic comorbidities.
Figure 1.
The study population's selection process. Key: AMD, age-related macular degeneration; IVI, intravitreal injection; VEGF, vascular endothelial growth factor
Table 1.
Systemic characteristics of patients with AMD receiving IVI of bevacizumab, ranibizumab, and aflibercept before IPTW
| Variable | Bevacizumab (n=701) | Ranibizumab (n=463) | Aflibercept (n=984) | MASD |
|---|---|---|---|---|
| Male | 407 (58.1) | 280 (60.5) | 619 (62.9) | 0.10 |
| Age, years | 71.1±10.8 | 70.3±10.6 | 70.1±9.7 | 0.10 |
| Comorbidity | ||||
| Hypertension | 158 (22.5) | 98 (21.2) | 219 (22.3) | 0.03 |
| Dyslipidemia | 95 (13.6) | 58 (12.5) | 133 (13.5) | 0.03 |
| Myocardial infarction | 4 (0.57) | 6 (1.30) | 6 (0.61) | 0.08 |
| Ischemic stroke | 13 (1.9) | 11 (2.4) | 17 (1.7) | 0.05 |
| Heart failure | 5 (0.71) | 4 (0.86) | 7 (0.71) | 0.02 |
| Chronic kidney disease | 48 (6.8) | 27 (5.8) | 65 (6.6) | 0.04 |
| Charlson's Comorbidity Index score | 0.67±1.2 | 0.63±1.1 | 0.65±1.1 | 0.04 |
| Laboratory data | ||||
| SBP, mmHg | 136.1±21.1 | 139.2±21.8 | 136.5±18.7 | 0.15 |
| DBP, mmHg | 75.0±12.0 | 76.6±14.0 | 76.3±12.1 | 0.12 |
| LDL, mg/dL | 83.7±43.2 | 65.2±41.6 | 77.2±44.1 | 0.44 |
| eGFR, mg/dL | 78.4±26.1 | 75.7±28.9 | 79.6±27.2 | 0.14 |
| ALT, U/L | 25.0±16.7 | 29.1±25.6 | 27.1±19.5 | 0.19 |
| Follow up years | 2.3±1.8 | 2.7±1.8 | 1.8±1.4 | 0.78 |
AMD, age-related macular degeneration; ALT, alanine aminotransferase; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; IVI, intravitreal injection; IPTW, inverse probability of treatment weighting; LDL, low-density lipoprotein cholesterol; MASD, maximum absolute standardized difference; SBP, systolic blood pressure. *Data were presented as frequency (percentage) or mean±standard deviation
Supplementary Table 1.
Ocular characteristics patients with AMD receiving IVI of bevacizumab, ranibizumab, and aflibercept before IPTW
| Variable | Bevacizumab (n=701) | Ranibizumab (n=463) | Aflibercept (n=984) | MASD |
|---|---|---|---|---|
| Injected eye | ||||
| O.D. | 276 (39.4) | 200 (43.2) | 450 (45.7) | |
| O.S. | 252 (35.9) | 197 (42.5) | 406 (41.3) | |
| O.U. | 53 (7.6) | 29 (6.3) | 37 (3.8) | |
| Unknown | 120 (17.1) | 37 (8.0) | 91 (9.2) | |
| Mean injection number until the end of follow up | 3.0±3.7 | 4.7±3.4 | 4.5±2.9 | 0.48 |
| Ocular history | ||||
| Cataract | 361 (51.5) | 257 (55.5) | 565 (57.4) | 0.12 |
| Glaucoma | 22 (3.1) | 8 (1.7) | 17 (1.7) | 0.10 |
| Retinal laser | 55 (7.8) | 59 (12.7) | 35 (3.6) | 0.36 |
| Vitrectomy | 38 (5.4) | 9 (1.9) | 16 (1.6) | 0.23 |
| Number of OPD visits at ophthalmology last year | 2.7±2.5 | 3.1±2.8 | 2.7±1.9 | 0.20 |
AMD, age-related macular degeneration; IVI, intravitreal injection; IPTW, inverse probability of treatment weighting; MASD, maximum absolute standardized difference; O.D., oculus dexter; O.S., oculus laevus; O.U., oculi utrigue; OPD, outpatient department. *Data were presented as frequency (percentage) or mean±standard deviation
Regarding laboratory data, the bevacizumab group had the highest LDL (83.7 ± 43.2 vs. 65.2 ± 41.6 vs. 77.2 ± 44.1 mg/dL, MASD = 0.44). The ranibizumab group had the highest SBP (136.1 ± 21.1 vs. 139.2 ± 21.8 vs. 136.5 ± 18.7 mmHg, MASD = 0.15), DBP (75.0 ± 12.0 vs. 76.6 ± 14.0 vs. 76.3 ± 12.1 mmHg, MASD = 0.12), and ALT (25.0 ± 16.7 vs. 29.1 ± 25.6 vs. 27.1 ± 19.5 U/L, MASD = 0.19) and the lowest eGFR (78.4 ± 26.1 vs. 75.7 ± 28.9 vs. 79.6 ± 27.2 mL/min/1.73m2, MASD = 0.15).
Regarding ocular history, bevacizumab users had the highest prevalence of vitrectomy (5.4% vs. 1.9% vs. 1.6%, MASD = 0.23) and glaucoma (3.1% vs. 1.7% vs. 1.7%, MASD = 0.10). Ranibizumab users had the highest prevalence of retinal laser (7.8% vs. 12.7% vs. 3.6%, MASD = 0.36). Aflibercept users had the highest prevalence of cataracts (51.5% vs. 55.5% vs. 57.4%, MASD 0.12). Ranibizumab receivers had the most visits to outpatient department of ophthalmology last year (2.7 ± 2.5 vs. 3.1 ± 2.8 vs. 2.7 ± 1.9, MASD = 0.20). The ranibizumab group had the highest number of IVIs of anti-VEGF agents until the end of follow-up (3.0 ± 3.7 vs. 4.7 ± 3.4 vs. 4.5 ± 2.9, MASD = 0.48).
After IPTW, most covariates became comparable among the three groups [Supplementary Table 2]. Non-substantial differences (MASD > 0.10) were only observed with males (58.1% vs. 59.3% vs. 62.7%, MASD = 0.10), cataracts (53.0% vs. 58.8% vs. 56.0%, MASD = 0.12), retinal laser (6.4% vs. 7.3% vs. 4.5%, MASD = 0.11), vitrectomy (3.6% vs. 1.2% vs. 1.9%, MASD = 0.14), and LDL (77.9 ± 16.9 vs. 76.0 ± 13.7 vs. 77.5 ± 16.2 mg/dL, MASD = 0.12), which were adjusted for in the subsequent multivariable analysis. The median follow-up times of three users were 2.1 ± 1.7, 2.1 ± 1.7, and 2.0 ± 1.5 years, respectively.
Supplementary Table 2.
Baseline characteristics patients with AMD receiving IVI of bevacizumab, ranibizumab, and aflibercept after IPTW
| Variable | Bevacizumab (n=701) | Ranibizumab (n=463) | Aflibercept (n=984) | MASD |
|---|---|---|---|---|
| Male | 58.1 | 59.3 | 62.7 | 0.10 |
| Age, years | 70.5±10.4 | 70.5±10.2 | 70.3±10.0 | 0.02 |
| BMI, kg/m2 | 24.5±2.3 | 24.6±2.0 | 24.7±2.4 | 0.08 |
| Smoking | 5.2 | 4.3 | 5.8 | 0.06 |
| Comorbidity | ||||
| Hypertension | 21.6 | 19.2 | 22.4 | 0.08 |
| Dyslipidemia | 13.3 | 11.3 | 13.8 | 0.07 |
| Myocardial infarction | 0.47 | 0.86 | 0.58 | 0.05 |
| Ischemic stroke | 1.9 | 2.2 | 1.7 | 0.04 |
| Heart failure | 0.58 | 0.87 | 0.88 | 0.04 |
| Chronic kidney disease | 6.0 | 4.8 | 6.5 | 0.07 |
| Charlson's Comorbidity Index score | 0.64±1.1 | 0.60±1.1 | 0.65±1.1 | 0.05 |
| Anti-hypertensive medications | ||||
| ACEis/ARBs | 6.6 | 5.7 | 5.2 | 0.06 |
| Beta-blockers | 5.3 | 4.4 | 5.0 | 0.04 |
| Calcium channel blockers | 5.4 | 5.1 | 6.1 | 0.04 |
| Thiazides | 0.19 | 0.61 | 0.30 | 0.06 |
| Other medications | ||||
| Anti-platelets | 4.6 | 5.9 | 5.0 | 0.06 |
| Anti-coagulants | 0.73 | 0.29 | 0.67 | 0.05 |
| Statins | 5.6 | 5.0 | 5.3 | 0.03 |
| Fibrates | 0.33 | 0.00 | 0.42 | 0.08 |
| Laboratory data | ||||
| SBP, mmHg | 137.6±12.1 | 137.6±11.5 | 137.1±11.6 | 0.04 |
| DBP, mmHg | 76.2±6.9 | 76.4±7.5 | 76.4±7.3 | 0.03 |
| LDL, mg/dL | 77.9±16.9 | 76.0±13.7 | 77.5±16.2 | 0.12 |
| eGFR, mg/dL | 81.7±14.0 | 81.3±13.6 | 82.1±14.3 | 0.05 |
| ALT, U/L | 24.9±9.1 | 25.4±13.5 | 25.2±9.8 | 0.06 |
| Injected eye | ||||
| O.D. | 37.5 | 44.4 | 46.7 | |
| O.S. | 35.7 | 43.9 | 40.7 | |
| O.U. | 7.8 | 4.9 | 3.6 | |
| Unknown | 19.0 | 6.7 | 9.0 | |
| Mean injection number until the end of follow up | 3.4±3.8 | 4.1±3.0 | 4.6±3.2 | 0.33 |
| Ocular history | ||||
| Cataract | 53.0 | 58.8 | 56.0 | 0.12 |
| Glaucoma | 2.5 | 1.9 | 1.8 | 0.05 |
| Retinal laser | 6.4 | 7.3 | 4.5 | 0.11 |
| Vitrectomy | 3.6 | 1.2 | 1.9 | 0.14 |
| Number of OPD visits at ophthalmology last year | 2.7±2.3 | 2.9±2.3 | 2.7±1.9 | 0.08 |
| Follow up years | 2.1±1.7 | 2.2±1.7 | 2.0±1.5 | 0.23 |
AMD, age-related macular degeneration; ALT, alanine amino transferase; ACEi, angiotensin converting enzyme inhibitors; ARBs, angiotensin receptor blockers; BMI, body mass index; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; IVI, intravitreal injection; IPTW, inverse probability of treatment weighting; LDL, low density lipoprotein cholesterol; MASD, maximum absolute standardized difference; O.D., oculus dexter; O.S., oculus laevus; O.U., oculi utrigue; OPD, outpatient department; SBP, systolic blood pressure. *Data were presented as percentage or mean±standard deviation
Systemic events and laboratory data
Systemic outcomes among bevacizumab, ranibizumab, and aflibercept users with AMD are listed in Table 2. A low number of events per 100 person-years were observed for MACE, HF, thromboembolic events, major bleeding, all-cause admission, and all-cause death in bevacizumab, ranibizumab, and aflibercept. Risks for MACE, HF, thromboembolic events, major bleeding, all-cause admission, and all-cause death were comparable in the bevacizumab and aflibercept groups (subdistribution hazard ratio [SHR] =1.39, 95% confidence interval [CI]: 0.85–2.28; SHR = 1.29, 95% CI: 0.56–2.96; SHR = 1.05, 95% CI: 0.50–2.21; SHR = 0.85, 95% CI: 0.53–1.36; SHR = 1.12, 95% CI: 0.92–1.36; and SHR = 1.27, 95% CI: 0.94–1.72; respectively), ranibizumab and aflibercept groups (SHR = 1.07, 95% CI: 0.65–1.76; SHR = 0.79, 95% CI: 0.33–1.92; SHR = 0.84, 95% CI: 0.39–1.78; SHR = 1.05, 95% CI: 0.70–1.58; SHR = 1.19, 95% CI: 0.99–1.43; and SHR = 1.12, 95% CI: 0.84–1.50; respectively), and bevacizumab and ranibizumab groups (SHR = 1.30, 95% CI: 0.79–2.13; SHR = 1.63, 95% CI: 0.67–3.98; SHR = 1.25, 95% CI: 0.57–2.76; SHR = 0.81, 95% CI: 0.51–1.29; SHR = 0.94, 95% CI: 0.77–1.14; and SHR = 1.13, 95% CI: 0.85–1.52; respectively) [Table 3].
Table 2.
Incidence of systemic outcomes in patients with AMD receiving IVI of bevacizumab, ranibizumab, and aflibercept
| Outcome | Incidence of Bevacizumab | Incidence of Ranibizumab | Incidence of Aflibercept |
|---|---|---|---|
| Major cardiac adverse events† | 0.97 (0.63–1.31) | 0.73 (0.47–0.98) | 0.72 (0.47–0.97) |
| Myocardial infarction | 0.14 (−0.01–0.30) | 0.00 (0.00–0.00) | 0.08 (−0.02–0.19) |
| Ischemic stroke | 0.40 (0.14–0.66) | 0.17 (0.02–0.32) | 0.39 (0.17–0.60) |
| Cardiovascular death | 0.69 (0.41–0.97) | 0.60 (0.37–0.84) | 0.41 (0.22–0.59) |
| Heart failure | 0.49 (0.20–0.78) | 0.29 (0.10–0.49) | 0.39 (0.17–0.60) |
| Thromboembolic events‡ | 0.54 (0.24–0.85) | 0.43 (0.20–0.67) | 0.52 (0.27–0.77) |
| Major bleeding | 1.3 (0.8–1.8) | 1.6 (1.1–2.0) | 1.5 (1.1–1.9) |
| All-cause admission | 8.8 (7.5–10.1) | 9.3 (8.1–10.5) | 7.6 (6.6–8.6) |
| All-cause death | 2.6 (2.0–3.1) | 2.3 (1.9–2.8) | 1.9 (1.5–2.3) |
AMD, age-related macular degeneration; CI, confidence interval; IVI, intravitreal injection; IPTW, inverse probability of treatment weighting; N/A, not applicable. *Number of events per 100 person-years. †Composite of myocardial infarction, ischemic stroke, and cardiovascular death. ‡Composite of myocardial infarction, ischemic stroke, transient ischemic attack, extremity thromboembolism, and systemic thromboembolism
Table 3.
Time-to-event analysis of systemic outcomes of patients with AMD receiving IVI of bevacizumab, ranibizumab, and aflibercept
| Outcome | SHR (95% CI) of Bevacizumab vs. Aflibercept | SHR (95% CI) of Ranibizumab vs. Aflibercept | SHR (95% CI) of Bevacizumab vs. Ranibizumab |
|---|---|---|---|
| Major cardiac adverse events† | 1.39 (0.85–2.28) | 1.07 (0.65–1.76) | 1.30 (0.79–2.13) |
| Myocardial infarction | 1.84 (0.37–9.07) | N/A | N/A |
| Ischemic stroke | 0.99 (0.41–2.37) | 0.42 (0.15–1.23) | 2.34 (0.77–7.10) |
| Cardiovascular death | 1.62 (0.88–3.00) | 1.41 (0.77–2.56) | 1.16 (0.66–2.02) |
| Heart failure | 1.29 (0.56–2.96) | 0.79 (0.33–1.92) | 1.63 (0.67–3.98) |
| Thromboembolic events‡ | 1.05 (0.50–2.21) | 0.84 (0.39–1.78) | 1.25 (0.57–2.76) |
| Major bleeding | 0.85 (0.53–1.36) | 1.05 (0.70–1.58) | 0.81 (0.51–1.29) |
| All-cause admission | 1.12 (0.92–1.36) | 1.19 (0.99–1.43) | 0.94 (0.77–1.14) |
| All-cause death | 1.27 (0.94–1.72) | 1.12 (0.84–1.50) | 1.13 (0.85–1.52) |
AMD, age-related macular degeneration; CI, confidence interval; IVI, intravitreal injection; IPTW, inverse probability of treatment weighting; N/A, not applicable; SHR, subdistribution hazard ratio. †Composite of myocardial infarction, ischemic stroke, and cardiovascular death. ‡Composite of myocardial infarction, ischemic stroke, transient ischemic attack, extremity thromboembolism, and systemic thromboembolism
Long-term changes in laboratory data among the three treatment groups with AMD were assessed every 6 months [Fig. 2] No significant differences in changes in SBP (P > 0.05), DBP (P > 0.05), LDL (P > 0.05), and ALT (P > 0.05) were observed among the three groups. While eGFR levels persistently reduced in all three users, the change did not differ among the three groups with significant difference (P > 0.05).
Figure 2.
Long-term changes in (a) SBP, (b) DBP, (c) LDL, (d) eGFR, and (e) ALT among patients with AMD receiving IVIs of bevacizumab, ranibizumab, and aflibercept in the IPTW-adjusted cohort. Key: AMD, age-related macular degeneration; ALT, alanine transaminase; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; IVI, intravitreal injection; IPTW, inverse probability of treatment weights; LDL, low-density lipoprotein; SBP, systolic blood pressure
Discussion
In patients diagnosed with AMD, IVI of anti-VEGF agents showed favorable efficacy and a good overall safety profile. Nonetheless, few studies have compared the incidence of SAEs between anti-VEGF agents. This study intended to investigate the risks of SAEs in patients diagnosed with AMD and treated with aflibercept, bevacizumab, or ranibizumab. The results showed that all patients had low and similar incidences of MACE, HF, thromboembolic events, major bleeding, all-cause admission, and all-cause death, irrespective of the anti-VEGF agent. In addition, long-term changes in SBP, DBP, LDL, eGFR, and ALT did not differ significantly among the three groups. These findings suggest that aflibercept, bevacizumab, and ranibizumab were all efficacious and well-tolerated agents.
While studies have already been published on this topic, they had limited statistical power to compare the anti-VEGF agents’ safety profiles. SAEs due to the three anti-VEGF agents have been reported in comparative RCTs and network meta-analyses of RCTs. The DRCR Protocol T and LEAVO trials have previously reported low and comparable risks for all SAEs between the three anti-VEGF agents.[20,21] Similarly, network meta-analyses of RCTs found no significant differences in all SAEs.[12,18] However, the relevance of these studies to clinical practice may be limited as such RCTs excluded patients with high risk at baseline. As discussed, most retrospective research only reported comparisons between two of the three anti-VEGF agents except for one retrospective cohort study. That study, which analyzed an extensive US administrative claims database, found that all anti-VEGF users shared similar risks of acute MI, all-cause hospitalization, cerebral vascular disease, and major bleeding.[13] As bevacizumab and ranibizumab were approved earlier, more studies have identified and compared their risks of SAEs. Three retrospective cohort studies from Australia, Canada, and the USA reported a low and comparable incidence of cardiovascular diseases.[24,29,30] With aflibercept therapy increasing, SAE comparisons between aflibercept and ranibizumab have recently been performed and have reported similar risks. Patients had comparable rates of non-ocular hemorrhage and MACE in two retrospective cohort studies using large databases from Italy and France.[31,32] However, the reported observational studies had small populations, making them underpowered to detect significant differences in rare SAEs. Our study used the largest multi-institutional database in Taiwan to compare the risk of SAEs among three anti-VEGF agents, ensuring that the analysis was sufficiently powered.
In this retrospective study, all anti-VEGF agent users had low rates of all-cause admission, all-cause death, HF, MACE, HF, major bleeding, and thromboembolic events. Systemic diffusion of IVIs has formed the basis of the hypothesis for anti-VEGFs inducing SAEs.[15] Arterial thromboembolism, hypertension, and proteinuria have occasionally been reported in intravenous chemotherapy anti-VEGF users.[9,10,11] However, anti-VEGF concentrations are 400 times greater in intravenous infusions than in IVIs.[7] In IVI, the protection provided by the blood-ocular barrier ensures that systemic anti-VEGF concentrations are too low to trigger SAEs.[6] In our outcomes, the three anti-VEGF agents possessed similar profiles of systemic safety. The three treatment groups showed comparable differences in the changes of SBP, DBP, LDL, eGFR, and ALT in the long term. Therefore, their low concentrations of anti-VEGF agents in systemic circulation appeared to protect anti-VEGF-treated patients from increased risks of SAEs.
To the best of our knowledge, our study was the first to investigate the risks of SAEs among aflibercept, bevacizumab, and ranibizumab users with a large-population database with multiple institutions in the Asian population. Our outcomes showed similar and favorable safety profiles of three IVI anti-VEGF agents. This study had several strengths. First, the real-world database with a sufficiently large sample size allowed for the effective detection of rare SAEs. This post-marketing and large-scale study complements data presented in RCTs. Second, active comparator and new user designs were adopted to minimize residual confounding variables. All covariates were well-balanced after the IPTW adjustment. Third, only patients with the indication of AMD were included to specifically clarify the association between anti-VEGF agents and SAEs. One assessed indication was that patients diagnosed with AMD, diabetic macular edema, or retinal vein occlusion might have different risks of SAEs at baseline. Fourth, the database contained information on self-paid interventions beyond the coverage of the Taiwan National Health Insurance program. Therefore, this study had good coverage for three common anti-VEGF agents, especially bevacizumab due to its off-label use in clinical practice. Finally, the half-life of anti-VEGF agents ranged from 2.5 to 7.3 days, with a mean of 4.9 days. Sufficient follow-up of approximately 2 years was powered enough to identify potential differences in the long-term risk of SAEs.
Our study also had some limitations. First, the population was mainly Asian. While anti-VEGF agents have shown good efficacy globally, the conclusions arrived at in this study on the safety profile of IVIs of anti-VEGF agents require further study in other countries to strengthen their generalizability. Second, the actual incidence of SAEs could be underestimated as data outside the database was not captured. Nevertheless, based on the findings, SAEs would occur non-differentially among the three treatment groups, and their relative effects should remain unbiased. Third, our study only analyzed the most common anti-VEGFs: aflibercept, bevacizumab, and ranibizumab. In addition, it only enrolled AMD patients. Therefore, additional studies are required to explore the incidence of SAEs in patients receiving other anti-VEGFs, such as pegaptanib, or diagnosed with other indications, such as DME and RVO. Finally, although IPTW was applied to balance multiple variables, some residual confounders, such as dietary patterns and exercise habits, which served as important factors of SAEs, could not be captured from this database. A confounding effect was unable to be entirely eliminated because of this study's retrospective nature.
Conclusion
Aflibercept, bevacizumab, and ranibizumab showed a good safety profile in our routine care. The three treatment groups showed a low and comparable incidence of all-cause admission, all-cause death, HF, MACE, major bleeding, and thromboembolic events. In addition, comparable differences in changes in SBP, DBP, LDL, eGFR, and ALT in the long term were observed among the three groups. Therefore, these anti-VEGFs were effective and well-tolerated medications that could be prescribed to patients with AMD safely.
Conflicts of interest
There are no conflicts of interest.
Acknowledgments
This study was supported by Chang Gung Memorial Hospital. The funder had no role in this study's conduct or result interpretation. The authors thank Alfred Hsing-Fen Lin, Ben Yu-Lin Chou, and Jane Yan-Jen Shiu for their assistance with the statistical analyses during the manuscript's completion.
Funding Statement
This study was supported by Chang Gung Memorial Hospital research grants (CMRPG3N1001, CMRPG3M0321 and CMRPG3M0322) and Ministry of Science and Technology, Taiwan, research grants (MOST 110-2314-B-182A-118).
References
- 1.Thomas CJ, Mirza RG, Gill MK. Age-related macular degeneration. Med Clin North Am. 2021;105:473–91. doi: 10.1016/j.mcna.2021.01.003. [DOI] [PubMed] [Google Scholar]
- 2.Solomon SD, Lindsley K, Vedula SS, Krzystolik MG, Hawkins BS, et al. Anti-vascular endothelial growth factor for neovascular age-related macular degeneration. Cochrane Database Syst Rev. 2019;3:Cd005139. doi: 10.1002/14651858.CD005139.pub4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Avery RL, Pieramici DJ, Rabena MD, Castellarin AA, Nasir MA, Giust MJ. Intravitreal bevacizumab (Avastin) for neovascular age-related macular degeneration. Ophthalmology. 2006;113:363–72.e5. doi: 10.1016/j.ophtha.2005.11.019. [DOI] [PubMed] [Google Scholar]
- 4.Rosenfeld PJ, Brown DM, Heier JS, Boyer DS, Kaiser PK, Chung CY, et al. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med. 2006;355:1419–31. doi: 10.1056/NEJMoa054481. [DOI] [PubMed] [Google Scholar]
- 5.Papadopoulos Z. Aflibercept: A review of its effect on the treatment of exudative age-related macular degeneration. Eur J Ophthalmol. 2019;29:368–78. doi: 10.1177/1120672119832432. [DOI] [PubMed] [Google Scholar]
- 6.Del Amo EM, Rimpelä AK, Heikkinen E, Kari OK, Ramsay E, Lajunen T, et al. Pharmacokinetic aspects of retinal drug delivery. Prog Retin Eye Res. 2017;57:134–85. doi: 10.1016/j.preteyeres.2016.12.001. [DOI] [PubMed] [Google Scholar]
- 7.Avery RL, Castellarin AA, Steinle NC, Dhoot DS, Pieramici DJ, See R, et al. Systemic pharmacokinetics and pharmacodynamics of intravitreal aflibercept, bevacizumab, and ranibizumab. Retina (Philadelphia, Pa) 2017;37:1847–58. doi: 10.1097/IAE.0000000000001493. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Reibaldi M, Fallico M, Avitabile T, Marolo P, Parisi G, Cennamo G, et al. Frequency of intravitreal anti-VEGF injections and risk of death: A systematic review with meta-analysis. Ophthalmology Retina. 2022;6:369–76. doi: 10.1016/j.oret.2021.12.019. [DOI] [PubMed] [Google Scholar]
- 9.Scappaticci FA, Skillings JR, Holden SN, Gerber HP, Miller K, Kabbinavar F, et al. Arterial thromboembolic events in patients with metastatic carcinoma treated with chemotherapy and bevacizumab. J Natl Cancer Inst. 2007;99:1232–9. doi: 10.1093/jnci/djm086. [DOI] [PubMed] [Google Scholar]
- 10.Robinson ES, Khankin EV, Karumanchi SA, Humphreys BD. Hypertension induced by vascular endothelial growth factor signaling pathway inhibition: Mechanisms and potential use as a biomarker. Semin Nephrol. 2010;30:591–601. doi: 10.1016/j.semnephrol.2010.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Izzedine H, Massard C, Spano JP, Goldwasser F, Khayat D, Soria JC. VEGF signalling inhibition-induced proteinuria: Mechanisms, significance and management. Eur J Cancer. 2010;46:439–48. doi: 10.1016/j.ejca.2009.11.001. [DOI] [PubMed] [Google Scholar]
- 12.Plyukhova AA, Budzinskaya MV, Starostin KM, Rejdak R, Bucolo C, Reibaldi M, et al. Comparative safety of bevacizumab, ranibizumab, and aflibercept for treatment of neovascular age-related macular degeneration (AMD): A systematic review and network meta-analysis of direct comparative studies. J Clin Med. 2020;9:1522. doi: 10.3390/jcm9051522. doi: 10.3390/jcm9051522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Maloney MH, Payne SR, Herrin J, Sangaralingham LR, Shah ND, Barkmeier AJ. Risk of systemic adverse events after intravitreal bevacizumab, ranibizumab, and aflibercept in routine clinical practice. Ophthalmology. 2021;128:417–24. doi: 10.1016/j.ophtha.2020.07.062. [DOI] [PubMed] [Google Scholar]
- 14.Platania CB, Di Paola L, Leggio GM, Romano GL, Drago F, Salomone S, et al. Molecular features of interaction between VEGFA and anti-angiogenic drugs used in retinal diseases: A computational approach. Front Pharmacol. 2015;6:248. doi: 10.3389/fphar.2015.00248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.García-Quintanilla L, Luaces-Rodríguez A, Gil-Martínez M, Mondelo-García C, Maroñas O, Mangas-Sanjuan V, et al. Pharmacokinetics of intravitreal anti-VEGF drugs in age-related macular degeneration. Pharmaceutics. 2019;11:365. doi: 10.3390/pharmaceutics11080365. doi: 10.3390/pharmaceutics11080365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Yang J, Wang X, Fuh G, Yu L, Wakshull E, Khosraviani M, et al. Comparison of binding characteristics and in vitro activities of three inhibitors of vascular endothelial growth factor A. Mol Pharm. 2014;11:3421–30. doi: 10.1021/mp500160v. [DOI] [PubMed] [Google Scholar]
- 17.Ngo Ntjam N, Thulliez M, Paintaud G, Salvo F, Angoulvant D, Pisella PJ, et al. Cardiovascular adverse events with intravitreal anti-vascular endothelial growth factor drugs: A systematic review and meta-analysis of randomized clinical trials. JAMA Ophthalmol. 2021;139:1–11. doi: 10.1001/jamaophthalmol.2021.0640. doi: 10.1001/jamaophthalmol. 2021.0640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ye L, Jiaqi Z, Jianchao W, Zhaohui F, Liang Y, Xiaohui Z. Comparative efficacy and safety of anti-vascular endothelial growth factor regimens for neovascular age-related macular degeneration: Systematic review and Bayesian network meta-analysis. Ther Adv Chronic Dis. 2020;11:2040622320953349. doi: 10.1177/2040622320953349. doi: 10.1177/2040622320953349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Reibaldi M, Fallico M, Avitabile T, Bonfiglio V, Russo A, Castellino N, et al. Risk of death associated with intravitreal anti-vascular endothelial growth factor therapy: A systematic review and meta-analysis. JAMA Ophthalmol. 2020;138:50–7. doi: 10.1001/jamaophthalmol.2019.4636. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wells JA, Glassman AR, Ayala AR, Jampol LM, Bressler NM, Bressler SB, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema: Two-year results from a comparative effectiveness randomized clinical trial. Ophthalmology. 2016;123:1351–9. doi: 10.1016/j.ophtha.2016.02.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hykin P, Prevost AT, Vasconcelos JC, Murphy C, Kelly J, Ramu J, et al. Clinical effectiveness of intravitreal therapy with ranibizumab vs aflibercept vs bevacizumab for macular edema secondary to central retinal vein occlusion: A randomized clinical trial. JAMA Ophthalmol. 2019;137:1256–64. doi: 10.1001/jamaophthalmol.2019.3305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sangroongruangsri S, Chaikledkaew U, Kumluang S, Wu O, Geue C, Ratanapakorn T, et al. Real-world safety of intravitreal bevacizumab and ranibizumab treatments for retinal diseases in Thailand: A prospective observational study. Clin Drug Investig. 2018;38:853–65. doi: 10.1007/s40261-018-0678-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Etminan M, Maberley DA, Babiuk DW, Carleton BC. Risk of myocardial infarction and stroke with single or repeated doses of intravitreal bevacizumab in age-related macular degeneration. Am J Ophthalmol. 2016;163:53–8. doi: 10.1016/j.ajo.2015.11.030. [DOI] [PubMed] [Google Scholar]
- 24.Curtis LH, Hammill BG, Schulman KA, Cousins SW. Risks of mortality, myocardial infarction, bleeding, and stroke associated with therapies for age-related macular degeneration. Arch Ophthalmol. 2010;128:1273–9. doi: 10.1001/archophthalmol.2010.223. [DOI] [PubMed] [Google Scholar]
- 25.Shao SC, Chan YY, Kao Yang YH, Lin SJ, Hung MJ, Chien RN, et al. The Chang Gung research database-A multi-institutional electronic medical records database for real-world epidemiological studies in Taiwan. Pharmacoepidemiol Drug Saf. 2019;28:593–600. doi: 10.1002/pds.4713. [DOI] [PubMed] [Google Scholar]
- 26.Yoshida K, Solomon DH, Kim SC. Active-comparator design and new-user design in observational studies. Nat Rev Rheumatol. 2015;11:437–41. doi: 10.1038/nrrheum.2015.30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47:1245–51. doi: 10.1016/0895-4356(94)90129-5. [DOI] [PubMed] [Google Scholar]
- 28.Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med. 2015;34:3661–79. doi: 10.1002/sim.6607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Campbell RJ, Gill SS, Bronskill SE, Paterson JM, Whitehead M, Bell CM. Adverse events with intravitreal injection of vascular endothelial growth factor inhibitors: Nested case-control study. BMJ. 2012;345:e4203. doi: 10.1136/bmj.e4203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kemp A, Preen DB, Morlet N, Clark A, McAllister IL, Briffa T, et al. Myocardial infarction after intravitreal vascular endothelial growth factor inhibitors: A whole population study. Retina (Philadelphia, Pa) 2013;33:920–7. doi: 10.1097/IAE.0b013e318276e07b. [DOI] [PubMed] [Google Scholar]
- 31.Sultana J, Giorgianni F, Scondotto G, Ientile V, Cananzi P, Leoni O, et al. Association between intravitreal aflibercept and serious non-ocular haemorrhage compared with intravitreal ranibizumab: A multicentre observational cohort study. Drug Safety. 2020;43:943–52. doi: 10.1007/s40264-020-00956-x. [DOI] [PubMed] [Google Scholar]
- 32.Billioti de Gage S, Bertrand M, Grimaldi S, Zureik M. Risk of myocardial infarction, stroke, or death in new users of intravitreal aflibercept versus ranibizumab: A nationwide cohort study. Ophthalmol Ther. 2022;11:587–602. doi: 10.1007/s40123-021-00451-1. [DOI] [PMC free article] [PubMed] [Google Scholar]