Randomized Trial of Biosimilar ABP 938 Compared with Reference Aflibercept in Adults with Neovascular Age-Related Macular Degeneration

Randomized, double-masked, active-controlled,  multicenter, multiregional, comparative clinical study

N= 579

To evaluate the clinical efficacy, safety, and immunogenicity of biosimilar ABP 938 (Pavblu; aflibercept-ayyh) compared with the aflibercept reference product (RP) in adults with neovascular (wet) age-related macular degeneration (nAMD)

ABP 938 (n= 288)

Aflibercept RP (n= 288)

Treatment-naïve nAMD involving the center of the fovea in ≥1 eye, age ≥50 years, best-corrected visual acuity (BCVA) between 34 and 73 Early Treatment Diabetic Retinopathy Study (ETDRS) letters, intraretinal and/or subretinal fluid, central retinal thickness >270 μm in the study eye at screening.

If both eyes were eligible at screening, the eye with the worse BCVA was selected for study.

Total lesion size >12-disc areas, active choroidal neovascularization (CNV) area <50% of total lesion area, scar/fibrosis/atrophy involving the center of the fovea in the study eye

Patients were randomized 1:1 to ABP 938 or aflibercept RP, stratified by geographic region and disease severity. Both groups received 2 mg/0.05 mL by intravitreal injection every 4 weeks for the first 3 doses, then every 8 weeks. At week 16, patients receiving ABP 938 continued ABP 938, while those receiving aflibercept RP were rerandomized either to continue aflibercept RP or transition to ABP 938 through the end of the study. Ophthalmic assessments included BCVA, slitlamp examination, intraocular pressure (IOP), indirect ophthalmoscopy, and spectral-domain optical coherence tomography (SD-OCT) at each visit; fundus photography and fluorescein angiography were performed at screening and weeks 8, 16, 24, and 52. Safety, immunogenicity, and a pharmacokinetic (PK) substudy of serum-free drug concentrations were also assessed.

Enrollment: June 2020 to January 2023

Study period: 52 weeks, consisting of a screening period (up to 4 weeks before randomization), a parallel arm period (from day 1 to week 16), a posttransition period (from rerandomization at week 16 to week 48), and a safety follow-up period (through week 52)

Primary: Least squares (LS) mean difference in change in BCVA from baseline to week 8

Secondary: Proportion of patients who maintained vision, mean change from baseline in BCVA, proportion of patients who gained ≥10 or ≥15 letters of vision, change from baseline in CNV area, change from baseline in central subfield thickness (CST)

ABP 938

(n= 288)

Aflibercept RP

(n= 288)

Total

(N= 576)

76.0 ± 7.9

151 (52.4%)

250 (86.8%)

58.9 ± 10.68

438.4 ± 129.05

At week 8, mean BCVA improved by 6.4 ETDRS letters with ABP 938 and 6.5 letters with aflibercept RP; the LS mean treatment difference was 0.1 letters, with 95% CI -1.3 to 1.5 and 90% CI -1.1 to 1.3, meeting the prespecified similarity margins.

Secondary efficacy outcomes were also comparable, including ≥10-letter gain at week 8 in 29.4% versus 32.7% of patients, respectively. After rerandomization at week 16, efficacy was maintained through week 52, with vision maintained in 95.6% of patients continuing ABP 938, 97.6% continuing aflibercept reference product, and 95.9% transitioning from aflibercept RP to ABP 938; ≥15-letter gains occurred in 24.3%, 29.6%, and 24.4%, respectively.

Safety findings were similar between treatments, with adverse events through week 16 reported in 39.2% and 37.2% of patients, respectively, most commonly conjunctival hemorrhage. No clinically meaningful differences in ocular or nonocular adverse events, events of interest, or antidrug antibody incidence were observed, supporting comparable efficacy, safety, and immunogenicity of ABP 938 and aflibercept reference product in nAMD. 

This study supports the conclusion of no clinically meaningful differences in efficacy, safety, and immunogenicity between ABP 938 and aflibercept RP in patients with nAMD. Transitioning from aflibercept RP to APB 938 at week 16 resulted in comparable efficacy, safety, and immunogenicity between treatment groups.

This was a robust biosimilar-comparability study with a randomized, double-masked, active-controlled design, adequate sample size, prespecified similarity margins, and a clinically relevant transition arm. Its main limitation for the broader Eylea-biosimilar question is that the clinical population was limited to treatment-naïve nAMD, so the study supports Pavblu’s similarity to aflibercept for nAMD but does not directly test its other FDA-approved aflibercept indications, such as retinal vein occlusion (RVO)-related macular edema, diabetic macular edema (DME), or diabetic retinopathy.

Prospective, phase 3, randomized, double-masked, parallel group, multicenter, international clinical trial

N= 449

SB15 (n= 224)

AFL (n= 225)

Participants were randomized 1:1 to receive 2 mg of SB15 or AFL every 4 weeks for the first 12 weeks (3 injections), followed by dosing every 8 weeks up to week 48, with final assessments at week 56.

At week 32, participants were rerandomized to either continue receiving SB15 or AFL or be transitioned from AFL to SB15 to evaluate the impact of switching from AFL to SB15. The week 32 time point was chosen for switching as it is the first dosing time point after week 24, which, according to regulatory authorities, is when the near-maximum serum concentrations under steady state should be measured.

Enrollment: June 2020 to March 2022

Treatment and follow-up: 56 weeks

Primary: Change in BCVA from baseline to week 8

Secondary: Changes in BCVA and central subfield thickness up to week 32, safety, pharmacokinetics, and immunogenicity

SB15

(n= 224)

AFL

(n= 225)

Race 

    Asian

    White

52 (23.2%)

170 (75.9%)

51 (22.7%)

172 (76.4%)

SB15

(n= 224)

AFL

(n= 225)

Biosimilar SB15 versus reference aflibercept in neovascular age-related macular degeneration: 1-year and switching results of a phase 3 clinical trial

Prospective, double-masked, randomised, phase 3 trial

N= 449

To evaluate efficacy, safety, pharmacokinetics (PK) and immunogenicity of SB15 versus reference aflibercept (AFL), and switching from AFL to SB15 in neovascular age-related macular degeneration (nAMD)

Initial randomization:

SB15 (n= 224)

AFL (n= 225)

Week 32 groups

SB15/SB15 (n= 219)

AFL/AFL (n= 108)

AFL/SB15 (n= 111)

Presence of a treatment-naïve subfoveal choroidal neovascularisation (CNV) lesion secondary to AMD occupying at least 50% of the total lesion, a total lesion area of ≤9.0 disc areas, and a best-corrected visual acuity (BCVA) score of 20/40–20/200 using Early Treatment Diabetic Retinopathy Study (ETDRS) charts

Not specified in the provided text

Participants were enrolled at 56 centers in 10 countries, including Europe, the United States, Russia, South Korea, and Japan. Participants were initially randomized 1:1 to intravitreal SB15 or AFL, 2 mg/0.05 mL, at weeks 0, 4, and 8, followed by dosing every 8 weeks through week 48. At week 32, AFL recipients were rerandomized to continue AFL or switch to SB15; SB15 recipients continued SB15 through a rerandomization procedure that maintained masking. Ophthalmic examinations, ETDRS visual acuity testing, and optical coherence tomography were performed at each visit; fluorescein angiography and fundus photography were performed at screening and weeks 32 and 56. Retinal images were assessed centrally. Continuous efficacy outcomes were analyzed using analysis of covariance, and categorical outcomes using a stratified Cochran–Mantel–Haenszel test. Efficacy analyses used available data without imputation; switching-period safety analyses included participants receiving ≥1 dose after rerandomization.

June 2020 to February 2021

Changes from baseline to week 56 in BCVA, central subfield thickness, total retinal thickness, and CNV area; proportions of participants losing <15 or gaining ≥15 BCVA letters; presence of intraretinal/subretinal fluid, sub-retinal pigment epithelial fluid, and active CNV leakage at week 56; change in NEI VFQ-25 composite score; and switching-period safety, immunogenicity, and serum trough drug concentrations

SB15 (n= 224)

73.7 ± 8.05

118 (52.7%)

Race

Asian

White

52 (23.2%)

170 (75.9%)

51 (22.7%)

172 (76.4%)

26 (23.4%)

84 (75.7%)

24 (22.2%)

83 (76.9%)

59.5 ± 10.60

353.3 ± 95.61

445.2 ± 140.06

SB15/SB15 (n= 219)

Change in BCVA from baseline to Week 56; LS mean, letters 

7.4 ± 0.93

Study participants with <15 letters loss from baseline at week 56

207 (95.8%)

Study participants with ≥15 letters gain from baseline at Week 56

57 (26.4%)

Change in CST from baseline to Week 56; LSmean, μm 

-119.2 ± 4.36

Change in TRT from baseline to week 56; LSmean, μm 

-132.4 ± 7.19

At week 56, intraretinal or subretinal fluid was present in 47.2%, 44.0%, and 48.5% of participants in the SB15/SB15, AFL/SB15, and AFL/AFL groups, respectively; sub-retinal pigment epithelial fluid was present in 24.1%, 27.5%, and 24.8%, and active choroidal neovascularization (CNV) leakage in 77.8%, 84.0%, and 79.6%.

LS mean changes from baseline in CNV area were −1.26, −0.68, and −1.25 mm², respectively, while mean NEI VFQ-25 composite score improvements were 4.1, 4.9, and 4.1 points.

After week 32, treatment-induced antidrug antibodies developed in no participants continuing SB15 or switching to SB15, compared with 1/96 (1.0%) participants continuing AFL.

Serum trough concentrations at weeks 40 and 56 were below the quantification limit of 5.00 ng/mL in all treatment groups.

Incidence and characteristics of treatment-emergent adverse events (TEAEs) were comparable between groups. Most TEAEs were mild or moderate and not related to the study drug. Ocular TEAEs included cataract and visual acuity reduced. Non-ocular TEAEs included hypertension. No intraocular inflammation was reported during the switching period.

Efficacy, safety, PK and immunogenicity were comparable between SB15 and AFL and between switched and non-switched participants.

Randomized, masked switching with central image assessment and substantial European enrollment strengthens this study’s relevance to aflibercept biosimilar switching. However, the week 56 analyses were not powered to establish equivalence, and the single switch with 24-week follow-up in initially treatment-naïve nAMD patients limits conclusions about repeated switching, rare adverse events, and other retinal indications.

20-week, multicenter, open-label extension study in India

N= 52

Continuation arm (n= 29)

Switch arm (n= 23)

Primary: Incidence of TEAEs

Secondary: Change in visual acuity (BCVA), change in CST, immunogenicity (antidrug antibodies)

Eye disorders

Diabetic retinal edema

Diabetic retinopathy

26 (89.66)

15 (51.72)

18 (78.26)

11 (47.83)

44 (84.62)

26 (50)

The adjusted mean difference in BCVA change between baseline of parent study and week 76 (end of extension study) was −1.20 ± 2.95 ETDRS letters (90% CI −6.15 to 3.75) and from baseline of extension study to week 20 (week 76—parent study) was 1.59 ± 1.1 (90% CI −0.26 to 3.43).

The adjusted mean difference in CST from baseline of parent study to week 76 was −15.9 ± 38.07 µm (90% CI −80.47 to 48.62) and from baseline of extension study to its end was 13.53 ± 16.13 µm (90% CI −13.86 to 40.92).

No participant reported treatment-induced or boosted antidrug antibodies.

Multicenter, retrospective, real-world study in South Korea

N= 984 patients (1083 eyes)

Treatment-naïve eyes (n= 188)

Switching eyes (n= 895)

Primary: Change in best-corrected visual acuity (BCVA) and central foveal thickness (CFT)

Secondary: BCVA and CFT changes at 1, 2, 3, 4, 5, and 6 months post-treatment, fluid resolution rates, adverse events