Abstract
Purpose:
The purpose of this study was to evaluate the efficacy and safety of bevacizumab-awwb in the off-label treatment of neovascular age-related macular degeneration (n-AMD) and diabetic macular edema (DME).
Methods:
All patients with n-AMD and DME treated in the maintenance phase according to the “treat and extend” strategy, who underwent forced drug substitution from bevacizumab to bevacizumab-awwb from October 2022 to April 2023 at the Tor Vergata Polyclinic in Rome, were evaluated in a retrospective study. The primary outcome was changes in central retinal thickness (CRT) over time following drug substitution. The secondary outcomes were variations in drug durability, best corrected visual acuity (BCVA) and retinal fluid, and the incidence of drug-related local and systemic serious adverse events.
Results:
Of 80 eyes of 76 patients with n-AMD and 55 eyes of 44 patients with DME included, before and after drug substitution, the average CRT did not statistically differ; the proportion of patients within time intervals of q8, q12, and q16 was not different; and the mean BCVA remained constant. Of a cumulative 3496 bevacizumab-awwb treatments (2154 for patients with n-AMD and 1342 for patients with DME), no local severe complications were detected. Out of a total of 544 patients (342 affected by n-AMD and 202 affected by DME), no serious adverse events were reported.
Conclusions:
In our cohort of patients with n-AMD and DME in the maintenance phase, bevacizumab-awwb seems to represent a viable and cost-effective intravitreal therapy with comparable efficacy and safety to the originator.
Translational Relevance:
This study provides a preliminary assessment of the efficacy and safety of intravitreal bevacizumab-awwb, which is widely used off-label in retinal vascular diseases.
For many years, biologics with anti-vascular endothelial growth factor (VEGF) activity have revolutionized the therapy of retinal vascular diseases, including neovascular age-related macular degeneration (n-AMD), diabetic macular edema (DME), and diabetic retinopathy.
1,2
Patent expiration on such biologic products allowed the development and diffusion of biosimilars, defined by the World Health Organization as biotherapeutic products that are similar in terms of quality, safety, and efficacy to an already licensed reference product (originator).
3
Biologics have very complex quaternary structures with unique proprietary manufacturing processes. Even small variations in these processes can modify the drugs’ tertiary and quaternary structure, potentially changing their efficacy, safety, and immunogenicity.
4 Produced through a different biotechnological process, approved biosimilar molecules are not identical to their originator and may have minimal structural differences not affecting their efficacy and safety profile. The registration process of biosimilars results faster than that of the originator.
5
One of the main issues regarding biosimilar clinical use is related to their interchangeability with the originator. Both the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) have released guidance regarding interchangeability.
6,7
The EMA assumes that all approved biosimilars are interchangeable, but leaves the decision of the interchangeability at the pharmacy level to member countries, based on postmarketing surveillance data. The EMA defines the term “switching” when the change biosimilar/originator follows the physician’s request and “substitution” when it is done at the pharmacy level.
7
Another peculiarity in biosimilar registration is the “extrapolation of indications,” that is the approval of a biosimilar for use in an indication held by the originator, not specifically studied by head-to-head clinical trials. Extrapolation is therefore a rationale process that bridges the biosimilar product to all the indications approved for the originator. This process eliminates the need for duplicative clinical trials but must be justified with appropriate data.
8
The Italian Pharmaceutical Agency (AIFA) has recently adopted the concept of extrapolation of indications for biosimilar drugs whose originator is currently used “off-label” in ophthalmology. In fact, despite not being approved for the treatment of retinal diseases, the anti-VEGF drug bevacizumab (AVASTIN) has been widely used off-label for intravitreal use due to its well-established efficacy and safety profile.
9–13
In Italy, off-label use of bevacizumab has been formally authorized by AIFA (Note 98, December 31, 2020) for n-AMD and DME with best corrected visual acuity (BCVA) better than 0.4 logMAR and recommended as first-line therapy.
14
After the patent expiration of bevacizumab, various bevacizumab biosimilars have been registered for intravenous infusion in oncology. The first one approved in Italy was bevacizumab-awwb (MVASI).
Like the originator, bevacizumab-awwb was not approved for ophthalmic use and it was not ever injected into the vitreous. Despite this, AIFA stated to allow the use of off-label bevacizumab-awwb in the therapy of n-AMD and DME by extrapolating the off-label indications of the originator.
14
Because the regional health system must purchase the cheapest drug commercially available, in many regions of Italy, we had the abrupt and unplanned replacement at the pharmacy level of bevacizumab originator with the less expensive bevacizumab-awwb. This happened in our center on October 29, 2022. This change, fully justified in oncology, impacted several ophthalmological departments that had to use a compounded drug never tested for the intraocular route. In fact, biosimilar systemic pharmacokinetics are not predictive of safety or efficacy for intraocular use, so, in addition to analytic similarity evidence, a clinical trial would be essential.
The study was approved by the Institutional Review Board of Tor Vergata University and was conducted in adherence to the tenets of the Declaration of Helsinki. Informed consent was obtained from the subjects after explanation of the nature and possible consequences of the study.
To assess the effects of bevacizumab-awwb administered intravitreally, we evaluated the records of all patients with n-AMD and DME in our center who underwent drug substitution at the pharmacy level from bevacizumab to its biosimilar in a retrospective, nonrandomized, consecutive case series study.
The study’s primary outcome was the changes in central retinal thickness (CRT) over time, which represents a well-established biomarker of VEGF activity suppression, following drug substitution.
15
Prespecified secondary outcomes included changes in drug durability, BCVA, and retinal fluid. The incidence of serious adverse events (SAEs), both local and systemic, related to the drug was also assessed with particular attention to intraocular inflammation (IOI).
All the records of patients with n-AMD and DME present in the database (EYES-UP) of the Retina Unit of the Tor Vergata Polyclinic, treated by bevacizumab-awwb from October 29, 2022, to April 30, 2023, were considered.
To assess the effects of drug substitution, we selected patients who were responders to a bevacizumab loading phase and then treated in the maintenance phase according to an optical coherence tomography (OCT)-based “treat and extend” strategy. Only patients who received at least three maintenance intravitreal injections (IVTs) of bevacizumab, followed by three consecutive IVTs of bevacizumab-awwb using the same treatment strategy, were included in the study.
13,16
Only patients having a complete ophthalmologic examination with standardized BCVA assessment by Early Treatment Diabetic Retinopathy Study (ETDRS) charts, slit-lamp examination of the anterior and posterior segment, and spectral domain optical coherence tomography (SD-OCT; Spectralis; Heidelberg Engineering, Heidelberg, Germany) scans performed on the same day of the injection, were included. Only the measurements acquired on the same days of the three IVTs preceding and the three IVTs following the drug substitution were considered. The time interval between injections was also evaluated.
The same experienced operator reviewed the follow-up OCT scans of all patients to assess CRT and the presence of intraretinal (IRF), subretinal (SRF), and sub-retinal pigment epithelium (RPE) fluid (pigment epithelial detachment [PED]). We adopted the following arbitrary fluid scoring (0 = absent, 1 = mild, 2 = moderate, and 3 = severe) to evaluate fluid changes over time.
We excluded from the study: (I) patients with incomplete clinical records, poor quality OCT images, BCVA worse than 1.2 logMAR, and lesions characterized by extensive fibrosis or atrophy judged incapable of functional recovery or worsening; (II) patients with retinal comorbidities, such as central or branch retinal vein occlusion, diabetic retinopathy concomitant with n-AMD, macular pucker, full-thickness macular hole or lamellar hole, and who had previously undergone intraocular surgery, except for uncomplicated phacoemulsification with intraocular lens (IOL) implantation; (III) patients who had intervals between injections not in line with those prescribed in their treatment plan or who had switched drug due to a medical decision; and (IV) patients without a BCVA and OCT evaluation later than 16 weeks after the last bevacizumab-awwb injection.
The incidence of local and systemic SAEs was assessed, including all patients receiving at least one bevacizumab-awwb injection.
Clinical data were entered into an Excel spreadsheet (Microsoft, Redmond, WA, USA). Statistical analysis was then carried out using the statistical package for the social sciences (SPSS version 15.0, Chicago, IL, USA). Specifically, the quantitative continuous variables were CRT and BCVA, and the discrete variables were intervals between administrations (weeks, q). The distribution of the variables was assessed by the Kolmogorov-Smirnov test. Normally distributed variables were presented as mean ± standard deviation and range (minimum and maximum). Frequency data were presented as percentages. Between-group comparisons of the “normal distributed” variables were conducted by 1-way ANOVA and/or univariate ANOVA tests and/or by repeated measures. Multiple comparisons were conducted by Bonferroni’s post hoc test. Categorical occurrence data between groups were compared by the chi-square test or Fisher's exact test (if cells < 5). Pearson’s test was used to determine the correlation coefficient and statistical significance of changes in the CRT and the free intervals between injections over time. Multiple regression analysis was used to compare the linear correlations between CRT and time (measured in weeks). The P value < 0.05 was considered as the threshold for statistical significance.
A total of 342 records of patients with n-AMD referred to the Retina Unit of the Tor Vergata Polyclinic were evaluated. All patients were of Caucasian ethnicity. Among them, 80 eyes of 76 patients (36 men and 40 women; average age = 81.27 ± 7.61 years; range = 66–96 years) met the inclusion criteria. Sixty-five eyes had type 1 macular neovascularization (MNV; 81%) and 15 eyes had type 2 MNV (19%). On average, patients received 19.50 ± 11.89 anti-VEGF injections before drug substitution (range = 3–52).
The mean CRT attributable at the last treatment with bevacizumab, measured at the time of the first bevacizumab-awwb injection, was 314.90 ± 90.22 µ and did not change substantially at the first control after the third IVT with the biosimilar (300.75 ± 72.70 µ, P = 0.83). Average CRT measurements during bevacizumab injections also did not differ significantly from those assessed during bevacizumab-awwb injections (315.39 ± 86.51 µ vs. 307.36 ± 81.52 µ, P = 0.23).
Linear correlation of CRT toward time (weeks) over the course of bevacizumab therapy showed a negative trend (slope = −0.71) reflecting the overall drug’s therapeutic action. This trend remained essentially constant (slope = −0.98) and did not significantly differ (P = 0.43) even during subsequent bevacizumab-awwb IVTs, confirming its therapeutic activity.
Concerning the free interval between consecutive injections, statistical analysis showed that the mean free interval between doses was not affected by drug substitution (bevacizumab = 11.65 ± 4.82 weeks and bevacizumab-awwb = 12.00 ± 5.07 weeks, P = 0.72) suggesting the substantial overlapping of the two drugs also in terms of durability.
Additionally, the proportion of patients within time intervals of q8, q12, and q16 was not statistically different before and after drug substitution (
P = 0.73;
Fig. 1).
Mean BCVA also remained constant throughout the observation period (0.52 ± 0.31 logMAR vs. 0.54 ± 0.30 logMAR, P = 0.31). Furthermore, at the administration of the first bevacizumab-awwb, the mean BCVA value was 0.54 ± 0.30 logMAR and remained unchanged at the first observation after the third bevacizumab-awwb injection. Overall, BCVA was stable within ± 2 logMAR units in all patients.
The retinal fluid change scoring was also stable in most eyes. Specifically, the IRF score remained constant in 78.82% of cases, improved in 18.82%, and worsened in 2.36% (P = 0.33); the SRF score remained unchanged in 80.00% of cases, improved in 11.34%, and worsened in 8.66% (P = 0.95); the PED score remained constant in 96.28% of cases, improved in 2.48%, and worsened in 1.24% (P = 0.97).
A total of 202 records of patients with DME were evaluated. All patients were of Caucasian ethnicity. Among them, 55 eyes of 44 patients (27 men and 17 women; average age = 67.76 ± 11.69 years; range = 41–86 years) met the inclusion criteria. On average, patients received 15.44 ± 7.58 anti-VEGF injections before drug substitution (range = 3–36).
Mean CRT at the end of treatment with bevacizumab, measured at the time of the first bevacizumab-awwb injection, was 307.38 ± 75.81 µ and did not change substantially after the third IVT with the biosimilar (297.31 ± 77.50 µ, P = 0.74). Average CRT during bevacizumab injections also did not differ significantly from those registered during bevacizumab-awwb therapy (315.34 ± 81.53 µ vs. 304.93 ± 76.67 µ, P = 0.17).
The linear correlation pattern of CRT toward time (weeks) over the course of bevacizumab therapy showed a negative trend (slope = −0.56) reflecting the overall drug’s therapeutic activity. This correlation showed a similar negative trend (slope = −0.84) during subsequent IVTs of bevacizumab-awwb and there was no statistically significant difference in CRT trend over time following drug substitution (P = 0.52).
Statistical analysis showed that the mean free interval between consecutive injections remained constant for both drugs (bevacizumab = 11.60 ± 3.92 weeks and bevacizumab-awwb = 12.32 ± 4.78 weeks,
P = 0.07), suggesting, also for DME, the substantial overlap of the two drugs in terms of durability. Additionally, the proportion of patients within time intervals of q8, q12, and q16 was not statistically different after the drug substitution (
P = 0.85;
Fig. 2).
Mean BCVA also remained constant throughout the observed period (0.35 ± 0.23 logMAR vs. 0.37 ± 0.22 logMAR, P = 0.32). Furthermore, at the administration of the first bevacizumab-awwb, the mean BCVA value was 0.36 ± 0.21 logMAR and remained stable at the first assessment after the third bevacizumab-awwb injection (0.39 ± 0.24 logMAR, P = 0.95). Even in patients with DME, BCVA remained stable within ± 2 logMAR units in all cases.
Retinal fluid scoring was stable in most eyes. Specifically, as evidenced by the evaluation of follow-up OCT scans, IRF remained constant in 61.75% of cases, improved in 20.03%, and worsened in 18.22% (P = 0.52). In addition, in terms of SRF, there was no overall difference after drug substitution (P = 0.98).
A cumulative of 3496 bevacizumab-awwb treatments (2154 for n-AMD and 1342 for DME) were given in our center during the observation period.
No local complications other than minor procedure-related adverse events (AEs), normally accompanying intravitreal therapy, were detected. In particular, neither side effects attributable to drug immunogenicity nor cases of IOI or vasculitis were reported.
Concerning systemic side effects, out of a total of 544 patients (342 patients affected by n-AMD and 202 patients affected by DME), no SAEs were reported (in particular arterial thromboembolic events), but the number of injections performed was certainly below the number necessary to accurately assess their incidence.
Among the included patients, 12 out of 80 n-AMD eyes and 3 out of 55 DME eyes needed to switch to an on-label anti-VEGF for ineffectiveness after 3 repeated monthly doses of bevacizumab-awwb.
The question of the efficacy and safety of biosimilars has long been controversial and has been further complicated by the introduction of biosimilars as substitutes for off-label biological drugs in the therapy of retinal diseases.
3
In 2021, to address a supply shortage of the originator bevacizumab, in the United States, some insurance companies have suggested the intravitreal use of bevacizumab biosimilars approved for metastatic colorectal cancer. This suggestion was contested by the American Academy of Ophthalmology (AAO) and the American Society of Retina Specialists because these biosimilars had never been tested for intraocular use. The AAO declared: “The Academy does not oppose the use of biosimilars or the use of off-label drugs in general. We are against the use of drugs in the eye that have never been tested for ophthalmic diseases.”
17
In Italy, as already mentioned, off-label use of bevacizumab has been authorized by AIFA (Note 98, December 31, 2020) and encouraged as first-line therapy in n-AMD and DME.
14
This AIFA statement appears to be in contrast, from a regulatory point of view, with the decision of the EMA concerning the registration of intravitreal bevacizumab biosimilars.
In fact, the EMA refused marketing approval for an intravitreal bevacizumab biosimilar (Ipique), because the company produced only literature concerning the originator bevacizumab to support its registration.
9–13 The EMA stated that there was no evidence concerning the efficacy and safety of Ipique in comparison to the originator when used intravitreally in the treatment of n-AMD. Therefore, the EMA considered that the safety and efficacy of this drug had not been adequately demonstrated and that the risks of Ipique registration outweighed its benefits, refusing marketing authorization.
18
Outlook Therapeutics has recently obtained EMA approval for a bevacizumab biosimilar (bevacizumab-vikg) for exclusive intraocular use, adding further issues to the already complex history of bevacizumab’s use in ophthalmology. The drug, compared for efficacy and safety to ranibizumab in NORSE studies, could be approved by the FDA with a paradoxical increase in the sale price, compared with bevacizumab and its cheaper biosimilars.
19 In our opinion, it is highly improbable that a registered biosimilar could replace the off-label use of bevacizumab in ophthalmology without any regulatory effort to modify the actual deregulation affecting this class of drugs, characterized by the mixing of “off-label use,” arbitrary indications’ extrapolation, substitution at pharmacy level, and centralized cheaper drug purchase.
Therefore, information about new bevacizumab biosimilars is necessary and this retrospective study, despite being limited, may be useful to share pilot data concerning the safety and efficacy profile of off-label intravitreal bevacizumab-awwb in patients affected by n-AMD and DME.
Regarding the safety profile, no significant local or systemic AEs were reported, although the number of injections was certainly lower than those required to assess statistically significant results. Furthermore, retrospective analyses are often burdened by the under-reporting of side effects that are either not reported or not attributed to anti-VEGF therapy by patients.
Concerning efficacy, the absence of changes in mean CRT suggests that the intravitreal use of bevacizumab-awwb ensures a clinically measurable anti-VEGF effect. CRT values showed a slightly negative trend over time following bevacizumab administration. This trend did not change after the drug substitution both in n-AMD and DME suggesting the overlapping of the drugs in terms of efficacy. Despite the limitation of a subjective initial assessment, we can also infer the efficacy of the biosimilar in VEGF blockade from the substantial stability of retinal fluid that represents, alongside CRT, a biomarker of VEGF activity, as widely reported in the literature.
15,20
Concerning functional results, BCVA remained almost stable in both diseases, confirming the capability of bevacizumab-awwb to stabilize visual function in these patients.
Nevertheless, we must highlight that only patients in the maintenance phase, injected following an OCT-based treat and extend strategy, were included in this study. According to this administration regimen, patients underwent follow-up OCT scans before each IVT to customize the next interval between treatments according to the anatomic response. Therefore, we were able to have information about the durability of bevacizumab-awwb by measuring the free interval changes between injections after the drug substitution. In n-AMD eyes, the average free interval between injections was about 12 weeks and did not change significantly after drug substitution. This interval is in the range of the values reported in the literature.
13 A similar trend was observed in patients with DME, suggesting that both drugs had virtually overlapping characteristics even in terms of durability.
16 Moreover, the proportion of patients in q8, q12, and q16 was not affected by drug substitution, both in patients with n-AMD and patients with DME.
Consequently, bevacizumab-awwb seems to represent a viable and cost-effective therapeutic alternative to the originator drug, at least in patients with n-AMD and patients with DME in the maintenance phase when strict OCT-basted “treat and extend” monitoring is warranted.
This study has some limitations due to its retrospective nature and the small sample size. Furthermore, the study did not include naïve patients, but only subjects in the maintenance phase and previously treated with different drugs and with a variable number of injections and disease duration.
In conclusion, our work has allowed a preliminary assessment of the efficacy and safety of bevacizumab-awwb, which, due to an intricate regulatory affair, is widely used off-label in retinal vascular diseases in our country. Therefore, the results of our study represent one of the few evidences available in the literature regarding the intravitreal use of bevacizumab-awwb.