Aflibercept 8 mg versus Faricimab Treat-and-Extend for Diabetic Macular Edema or Neovascular Age-Related Macular Degeneration: A Bayesian Fixed-Effect Network Meta-analysis of Clinical Trials

Abstract

Introduction

Reducing intravitreal injection frequency while maintaining efficacy is a critical goal in alleviating the burden associated with anti-vascular endothelial growth factor (VEGF) therapy in patients with diabetic macular edema (DME) or neovascular age-related macular degeneration (nAMD). In clinical trials, aflibercept 8 mg and faricimab 6 mg administered at extended dosing intervals have demonstrated similar efficacy compared with aflibercept 2 mg, but with fewer injections. This network meta-analysis (NMA) indirectly compared numbers of injections and efficacy between aflibercept 8 mg and faricimab 6 mg administered according to a treat-and-extend (T&E)-based regimen in patients with DME or nAMD.

Methods

A systematic literature review was conducted on 10 January 2025, to identify randomized controlled trials of aflibercept 8 mg or faricimab T&E with observation periods of approximately 2 years. Outcomes included number of injections, absolute change from baseline in best-corrected visual acuity (BCVA), and absolute and percentage change from baseline in central subfield thickness (CST). NMAs were performed with Bayesian statistical models. Injection numbers were adjusted to 104 weeks to account for differences in trial observation periods.

Results

The NMA included 2-year data from six trials: PHOTON, YOSEMITE, and RHINE (DME); and PULSAR, TENAYA, and LUCERNE (nAMD). Treatment with aflibercept 8 mg was associated with significantly fewer injections compared with faricimab T&E in patients with DME (mean difference −3.62 [95% credible interval −4.22, −3.02]) or nAMD (−1.47 [−1.90, −1.05]). Mean changes from baseline in BCVA (absolute) or CST (absolute and percentage) did not differ significantly between the two treatments.

Conclusion

This NMA indicated that aflibercept 8 mg required significantly fewer injections while maintaining similar efficacy over 2 years of treatment compared with faricimab T&E in patients with DME or nAMD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40123-025-01247-3.

Key Summary Points

Why carry out this study?

Frequent intravitreal injections can impose a substantial burden on patients with diabetic macular edema (DME) or neovascular age-related macular degeneration (nAMD) receiving anti-vascular endothelial growth factor (VEGF) therapy.

Aflibercept 8 mg and faricimab have been approved in several countries, including the USA and those of the European Union, and allow for extended dosing intervals in both DME and nAMD. However, dosing regimens with aflibercept 8 mg and faricimab have not been evaluated in head-to-head trials.

This network meta-analysis indirectly compared injection numbers and efficacy between aflibercept 8-mg and faricimab treat-and-extend (T&E) regimens in patients with DME or nAMD using 2-year data from randomized controlled trials.

What was learned from the study?

Results of this network meta-analysis indicated that aflibercept 8 mg was associated with significantly fewer injections and similar efficacy over 2 years compared with faricimab T&E in both DME and nAMD, although differences in aflibercept 8-mg and faricimab trial designs influenced the number of injections observed in the trials. However, owing to the limitation of this network meta-analysis, prospectively designed head-to-head clinical trials or real-world studies would be required to confirm the findings.

Introduction

Diabetic macular edema (DME) and neovascular age-related macular degeneration (nAMD) are major causes of vision loss worldwide [1–3], and the prevalence of both is expected to rise owing to the increasing prevalence of diabetes and aging populations, respectively [4, 5]. Intravitreal anti-vascular endothelial growth factor (VEGF) therapy has significantly improved visual and anatomic outcomes in clinical trials and is the standard of care for DME and nAMD [6–12]. By improving vision, intravitreal anti-VEGF therapy has been projected to provide direct and indirect economic benefits [13, 14] provided that real-world treatment outcomes reflect those achieved in clinical trials [13]. However, visual and anatomic improvements reported for clinical trials have not always been observed in real-world clinical practice [15–22]. This can be partly attributed to the substantial burden associated with frequent intravitreal anti-VEGF injections [23–28]. Aflibercept 8 mg and faricimab allow for extended dosing intervals without compromising efficacy [29–32], which may help to alleviate the burden associated with treatment of VEGF-driven retinal diseases.

Aflibercept acts as a soluble decoy receptor that binds to VEGF-A and placental growth factor, leading to inhibition of VEGF receptors [33]. Aflibercept 8 mg delivers a fourfold higher molar dose than the original 2-mg dose, prolonging VEGF inhibition and allowing for extended dosing intervals. In the pivotal 96-week PHOTON [29, 34] and PULSAR [30, 35] trials, which compared aflibercept 8 mg every 12 or 16 weeks (8q12 or 8q16) with aflibercept 2 mg every 8 weeks (2q8) following initial monthly dosing in patients with DME (PHOTON [NCT04429503]) or nAMD (PULSAR [NCT04423718]), patients in the 8q12 and 8q16 groups could qualify for dosing intervals of up to every 24 weeks beginning at week 52. Over 2 years of treatment, aflibercept 8 mg at extended dosing intervals achieved similar best-corrected visual acuity (BCVA) gains and demonstrated a similar safety profile compared with aflibercept 2q8, with fewer injections [29, 30, 34, 35].

Faricimab is a bispecific antibody that binds to and inhibits the activity of angiopoietin-2 and VEGF-A [36]. Faricimab 6 mg treat-and-extend (T&E) regimens up to every 16 weeks were compared with faricimab 6 mg every 8 weeks or aflibercept 2q8 following initial monthly dosing in pivotal trials that included patients with DME (YOSEMITE [NCT03622580] and RHINE [NCT03622593] [32, 37]) or nAMD (TENAYA [NCT03823287] and LUCERNE [NCT03823300] [31, 38]). Over 2 years of treatment, faricimab T&E demonstrated similar visual improvements and safety compared with faricimab 6 mg every 8 weeks and aflibercept 2q8, with fewer injections [31, 32, 37, 38].

Aflibercept 8 mg and faricimab 6 mg are approved in several countries, including the USA [33, 36] and in Europe [39, 40], for the treatment of DME or nAMD with dosing intervals up to every 16 weeks. To guide decision-making in the management of patients with DME and nAMD in real-world clinical practice, it is important to understand potential differences in treatment exposure and efficacy with novel interventions. In the absence of head-to-head trials, this network meta-analysis (NMA) was conducted to indirectly compare numbers of injections and efficacy between aflibercept 8 mg (8q12 or 8q16) and faricimab T&E (including regimens from every 12 weeks [q12w] to every 16 weeks [q16w]) in patients with DME or nAMD using 2-year follow-up data from randomized controlled trials identified from a systematic literature review.

Methods

Systematic Literature Review

The systematic literature review was designed and conducted in accordance with National Institute for Health and Care Excellence (NICE) guidelines [41–43], recommendations of the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) Task Force on Indirect Treatment Comparisons [44], and the Cochrane Handbook [45, 46]. The systematic literature review was not prospectively registered in any database.

Initially, two broad systematic literature reviews encompassing MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov were conducted in September 2023 to identify randomized controlled trials of any anti-VEGF agent in patients with DME or nAMD. Studies reporting data for at least 40 patients were included to reduce potential publication bias and uncertainty associated with small sample sizes. Only articles published in English were included. No additional restrictions were imposed on the study comparator, patient characteristics, or geographical locations. However, no articles from this initial search were used to inform the NMA. A third systematic literature review encompassing MEDLINE, Embase, and ClinicalTrials.gov was performed on 10 January 2025, to identify randomized controlled trials that had been published since September 2023. The third systematic literature review limited anti-VEGF therapy to aflibercept 8 mg or faricimab T&E and only included trials with observation periods of approximately 2 years to account for variations in loading dose regimens that occur in the first year of treatment.

The searches were conducted using Medical Subject Heading terms and keywords relevant to the populations, interventions, and study designs of interest (Supplementary Table S1). Two authors independently screened the retrieved records to remove duplicates and identify eligible studies. Data describing study design and characteristics, populations, interventions, comparators, and outcomes of interest were extracted and verified by three authors. Outcomes of interest were mean injection frequencies, and mean changes from baseline in BCVA and central subfield thickness (CST). For the purposes of this analysis, central retinal thickness (reported in the aflibercept 8 mg trials) and CST were considered to be equivalent. Risk of bias within eligible studies was evaluated by two authors using the Cochrane risk-of-bias tool version 2 [47].

Network Meta-analysis

NMAs were performed with Bayesian statistical models, consistent with NICE guidelines [41, 48] and recommendations of the ISPOR Task Force on Indirect Treatment Comparisons [44], using R: A language and environment for statistical computing (version 4.4.1) [49, 50]. Both fixed-effect and random-effects models were considered for the analyses. On the basis of the established guidelines, the decision of the most suitable model was informed by deviance information criterion statistics and model convergence (see Supplementary Methods) [48]. The Bayesian fixed-effect NMA protocol was not prospectively registered.

For continuous outcomes, the input data were point estimates and standard errors at the treatment-arm level. Where multiple results were reported for the same outcome and treatment arm, but with different approaches to impute missing data, the following hierarchy was applied for the selection of inputs: (1) results based on imputation of missing data (as used in the mixed model for repeated measures approach), (2) results based on the last observation carried forward approach, and (3) observed cases (no imputation).

Differences in the mean number of injections between aflibercept 8 mg and faricimab T&E in patients with DME or nAMD were estimated over 2 years of treatment. Injection numbers were adjusted to 104 weeks to account for differences in observation periods across trials (see Supplementary Methods). Efficacy was evaluated via estimation of differences in mean absolute changes from baseline in BCVA and both mean absolute and percentage changes from baseline in CST between aflibercept 8 mg and faricimab T&E in patients with DME or nAMD over 2 years of treatment. Percentage changes in CST were calculated using reported baseline and mean change from baseline values of CST (details in the Supplementary Methods). For the purposes of this analysis, the aflibercept 8q12 and 8q16 groups were combined to create a single aflibercept 8-mg group. The faricimab T&E group includes q12w and q16w dosing intervals. Results were reported as mean differences with 95% credible intervals. Continuous outcome differences were considered significant when 95% credible intervals did not cross 0.

Ethics

This NMA includes data from previously conducted clinical trials. No new studies with human participants were conducted by any of the authors.

Results

Systematic Literature Review

The initial systematic literature reviews identified two articles reporting on the YOSEMITE, RHINE, TENAYA, and LUCERNE trials of faricimab T&E versus aflibercept 2 mg; however, these articles reported 1-year data only [31, 32] and were not eligible for inclusion in the NMA.

The third systematic literature review identified two articles reporting 2-year data for the YOSEMITE and RHINE [37] and TENAYA and LUCERNE [38] trials of faricimab T&E versus aflibercept 2 mg, which were included in the NMA (Supplementary Fig. S1). Excluded articles included those reporting data for trials of aflibercept 2 mg versus an anti-VEGF agent other than faricimab or aflibercept 8 mg, and two articles reporting 1-year data for the PHOTON and PULSAR trials of aflibercept 8 mg versus aflibercept 2 mg [29, 30]. Two-year data for PHOTON and PULSAR were obtained from clinical study reports (data on file; PHOTON clinical study report, Regeneron Pharmaceuticals, Inc., 2023; PULSAR clinical study report, Bayer AG, 2023), which were included in the NMA. These six identified trials were generally of high quality with low risk of bias per the Cochrane risk-of-bias tool version 2 (Supplementary Fig. S2).

Network Meta-analysis

The six trials included in the NMA are summarized in Table 1. The random-effects models led to very similar deviance information criterion values, but with unstable results, compared with the fixed-effect models (Supplementary Methods). As such, and since only three trials were included in each network comparison (DME and nAMD) (Supplementary Fig. S3), it was determined that fixed-effect models, rather than random-effects models, were the most appropriate. Note, as the networks did not include any closed loops, no assessment of inconsistency was required.

Table 1.

Summary of clinical trials included in the NMA

Study (NCT number)	Patients/indication	Intervention	N	Injection schedule	Mean age/sex	Primary outcome
PHOTON (NCT04429503) [29, 34]	Adults aged ≥ 18 years with treatment-naive or previously treated DME	Aflibercept 8q16	164	Three initial monthly doses then 8q12 or 8q16 until week 96 If prespecified criteria were met, dosing intervals could be shortened from week 16a or extended from week 52b	62.3 years/61% male	Aflibercept 8q12 and 8q16 versus aflibercept 2q8 demonstrated: Noninferior BCVA changes from baseline to week 48 (mean differences −0.57 [95% CI −2.26, 1.13; p < 0.0001] and −1.44 [95% CI −3.27, 0.39; p = 0.0031] letters, respectively)c Similar BCVA gains from baseline to week 96 (mean differences 0.45 [95% CI −1.55, 2.45; p < 0.0001]d and −1.11 [95% CI −3.27, 1.05; p = 0.0044]d letters, respectively)c
		Aflibercept 8q12	329
		Aflibercept 2q8	167	Five initial monthly doses then 2q8 until week 96
YOSEMITE (NCT03622580) [32, 37]	Adults aged ≥ 18 years with treatment-naive or previously treated DME	Faricimab 6 mg T&E	313	Four initial monthly doses then adjustable dosing intervals up to every 16 weeks (if prespecified criteria were mete) until week 96	61.6–62.8 years/57–63% male	Faricimab 6 mg T&E and 6 mg every 8 weeks versus aflibercept 2q8 demonstrated: Noninferior BCVA gains from baseline through year 1 (mean differences 0.7 [97.52% CI −1.1, 2.5] and −0.2 [97.52% CI −2.0, 1.6] letters, respectively)c,f Maintenance of noninferior BCVA changes through year 2 (mean differences −0.7 [95.04% CI −2.5, 1.2] and −0.7 [95.04% CI −2.6, 1.2] letters, respectively; both p > 0.05d)
		Faricimab 6 mg every 8 weeks	315	Six initial monthly doses then every 8 weeks until week 96
		Aflibercept 2q8	312	Five initial monthly doses then 2q8 until week 96
RHINE (NCT03622593) [32, 37]	Adults aged ≥ 18 years with treatment-naive or previously treated DME	Faricimab 6 mg T&E	319	Four initial monthly doses then adjustable dosing up to every 16 weeks (if prespecified criteria were mete) until week 96	61.6–62.5 years/59–62% male	Faricimab 6 mg T&E and 6 mg every 8 weeks versus aflibercept 2q8 demonstrated: Noninferior BCVA changes from baseline through year 1 (mean differences 0.5 [97.52% CI −1.1, 2.1] and 1.5 [97.52% CI −0.1, 3.2] letters, respectively)c,f Maintenance of noninferior BCVA changes through year 2 (mean differences 0.7 [95.04% CI −1.3, 2.7] and 1.5 [95.04% CI −0.5, 3.6] letters, respectively; both p > 0.05d)
		Faricimab 6 mg every 8 weeks	317	Six initial monthly doses then every 8 weeks until week 96
		Aflibercept 2q8	315	Five initial monthly doses then 2q8 until week 96
PULSAR (NCT04423718) [30, 35]	Adults aged ≥ 50 years with treatment-naive CNV secondary to nAMD	Aflibercept 8q16	338	Three initial monthly doses then 8q12 or 8q16 until week 96 Dosing intervals were shortened from week 16 or extended from week 52 if prespecified criteria were metg,h	74.5 years/54.5% female	Aflibercept 8q12 and 8q16 versus aflibercept 2q8 demonstrated: Noninferior BCVA changes from baseline to week 48 (mean differences −0.97 [95% CI −2.87, 0.92; p = 0.0009] and −1.14 [−2.97, 0.69; p = 0.0011] letters, respectively)c Similar BCVA changes from baseline to week 96 (mean differences −1.01 [95% CI −2.82, 0.80; p = 0.0006]d and −1.08 [−2.87, 0.71; p = 0.0007]d letters, respectively)c
		Aflibercept 8q12	336
		Aflibercept 2q8	337	Three initial monthly doses then 2q8
TENAYA (NCT03823287) [31, 38]	Adults aged ≥ 50 years with treatment-naive CNV secondary to nAMD	Faricimab 6 mg T&E	334	Four initial monthly doses then adjustable dosing intervals up to every 16 weeks (if prespecified criteria were meti) until week 60 All patients received an active dose of faricimab at week 60 then T&E until week 108j	75.9–76.7 years/57–63% female	Faricimab 6 mg T&E versus aflibercept 2q8 demonstrated: Noninferior BCVA changes from baseline through year 1 (mean difference 0.7 [95% CI −1.1 to 2.5] letters)c,f Maintenance of noninferior BCVA changes through year 2 (mean difference 0.4 [95% CI −1.9, 2.8] letters)c,f
		Aflibercept 2q8	337	Three initial monthly doses then 2q8 until week 108
LUCERNE (NCT03823300) [31, 38]	Adults aged ≥ 50 years with treatment-naive CNV secondary to nAMD	Faricimab 6 mg T&E	331	Four initial monthly doses then adjustable dosing intervals up to every 16 weeks (if prespecified criteria were meti) until week 60 All patients received an active dose of faricimab at week 60 then T&E until week 108j	74.8–76.1 years/57–61% female	Faricimab 6 mg T&E versus aflibercept 2q8 demonstrated: Noninferior BCVA changes from baseline through year 1 (mean difference 0.0 [95% CI −1.7, 1.8] letters)c,f Maintenance of noninferior BCVA changes through year 2 (mean difference −0.2 [95% CI −2.4, 2.1] letters])c,f
		Aflibercept 2q8	327	Three initial monthly doses then 2q8 until week 108

2q8 aflibercept 2 mg every 8 weeks, 8q12 aflibercept 8 mg every 12 weeks, 8q16 aflibercept 8 mg every 16 weeks, BCVA best-corrected visual acuity, CNV choroidal neovascularization, CRT central retinal thickness, CST central subfield thickness, DME diabetic macular edema, nAMD neovascular age-related macular degeneration, NMA network meta-analysis, OCT optical coherence tomography, T&E treat and extend

^a > 10-letter loss in BCVA from week 12 due to persistent or worsening DME and > 50-µm increase in CRT from week 12 (patients meeting these criteria at weeks 16 or 20 had their aflibercept 8-mg dosing interval shortened to every 8 weeks; patients meeting these criteria at week 24 had their aflibercept 8-mg dosing interval shortened by 4-week increments to a minimum interval of every 8 weeks)

^b < 5-letter loss in BCVA from week 12 and CRT < 300 µm or < 320 μm on Spectralis (patients meeting these criteria had their aflibercept 8-mg dosing interval extended by 4-week increments to a maximum of every 24 weeks)

^cBased on a noninferiority margin of 4 letters

^dNominal p-value

^eCST < 325 µm at or after week 12 (once achieved, faricimab dosing intervals were first extended to every 8 weeks, then could be maintained, or extended by 4-week increments to a maximum of 6 mg every 16 weeks, or shortened by 4- or 8-week increments to a minimum of 6 mg every 4 weeks, based on prespecified CST and BCVA criteria)

^fNoninferiority was established by the lower bounds of the CIs for the mean differences being more than −4 letters

^g > 5-letter loss in BCVA from week 12 due to persistent or worsening nAMD and > 25-µm increase in CST from week 12 or new-onset foveal neovascularization or foveal hemorrhage (patients meeting these criteria at weeks 16 or 20 had their aflibercept 8-mg dosing interval shortened to 8 mg every 8 weeks; patients meeting these criteria at week 24 had their aflibercept 8-mg dosing interval shortened by 4-week increments to a minimum interval of every 8 weeks)

^h < 5-letter loss in BCVA from week 12 and no fluid at the central subfield on OCT and no new-onset foveal hemorrhage or foveal neovascularization (patients meeting these criteria had their aflibercept 8-mg dosing interval extended by 4-week increments to a maximum of every 24 weeks)

ⁱPatients with active disease at week 20 received faricimab 6 mg every 8 weeks, patients with active disease at week 24 received faricimab 6 mg every 12 weeks, and patients without active disease at week 28 received faricimab 6 mg every 16 weeks

^jFaricimab dosing intervals were extended by 4-week increments to a maximum of 6 mg every 16 weeks (if a patient achieved stable anatomic features, vision, and no new macular hemorrhage), shortened by 4- or 8-week increments to a minimum of 6 mg every 8 weeks (if worsening vision or anatomic features were found, or new macular hemorrhage), or maintained if neither the extension nor shortening criteria were met

In patients with DME, the NMA indicated that aflibercept 8 mg was associated with significantly fewer injections compared with faricimab T&E (−3.62 [95% credible interval: −4.22, −3.02]). Similarly, in patients with nAMD, aflibercept 8 mg was associated with significantly fewer injections compared with faricimab T&E (−1.47 [95% credible interval: −1.90, −1.05]) (Fig. 1). The mean change from baseline in BCVA did not differ significantly between aflibercept 8 mg compared with faricimab T&E in patients with DME (−0.16 [95% credible interval: −2.58, 2.28]) or nAMD (−1.14 [95% credible interval: −3.53, 1.25]) (Fig. 2A). The mean change from baseline in CST also did not differ significantly between aflibercept 8 mg and faricimab T&E in patients with DME (15.65 [95% credible interval: −6.06, 37.60]) or nAMD (1.17 [95% credible interval: −9.60, 11.92]) (Fig. 2B). Similarly, the percentage change from baseline in CST did not differ significantly between aflibercept 8 mg compared with faricimab T&E in patients with DME (2.66 [95% credible interval: −0.98, 6.32]) or nAMD (0.96 [95% credible interval: −0.71, 2.63]) (Fig. 2C).

Fig. 1.

Fixed-effect NMA of the mean number of injections estimated over a 104-week treatment period^a. Group [1] and [2] refer to the first and second group in each pair of comparisons, with mean differences (95% Crl) to the left of the dashed line indicating Group [1] is favored versus Group [2]; e.g., in the first comparison, aflibercept 8 mg (Group [1]) is favored versus faricimab T&E (Group [2]). CrI credible interval, DME diabetic macular edema, nAMD neovascular age-related macular degeneration, NMA network meta-analysis, T&E treat and extend. ^aInjection numbers were adjusted to 104 weeks to account for differences in the clinical trials’ observation periods. ^bComparison was statistically significant (95% Crls did not cross 0)

Fig. 2.

Fixed-effect NMA of (a) mean changes from baseline in BCVA, (b) mean changes from baseline in CST, and (c) mean percentage changes from baseline in CST over 2 years of treatment. Group [1] and [2] refer to the first and second group in each pair of comparisons; e.g., aflibercept 8 mg (Group [1]) versus faricimab T&E (Group [2]). BCVA best-corrected visual acuity, CrI credible interval, CST central subfield thickness, DME diabetic macular edema, nAMD neovascular age-related macular degeneration, NMA network meta-analysis, T&E treat and extend. ^aComparison was statistically significant (95% Crls did not cross 0)

Discussion

Informed by 2-year follow-up data from pivotal randomized controlled trials, this NMA found that treatment with aflibercept 8 mg was associated with significantly fewer injections and similar efficacy outcomes compared with faricimab T&E in patients with DME or nAMD. These findings align with a previous NMA, which showed that aflibercept 8 mg-treated patients with nAMD achieved similar efficacy outcomes with fewer injections compared with patients who were treated with fixed or flexible regimens of other anti-VEGF agents, including faricimab [51]. However, this earlier NMA was limited to 1 year of follow-up data, whereas the present NMA included 2 years of follow-up data. The indirect comparison of injection frequencies including the first and second years of treatment is more insightful than only including the first year of treatment, as the trial designs typically only permitted dosing interval extensions in the second year. The difference in number of injections that was reported by the present NMA reflects the higher number of initial monthly injections required for faricimab T&E in the YOSEMITE, RHINE, TENAYA, and LUCERNE trials (four injections) compared with aflibercept 8 mg in the PHOTON and PULSAR trials (three injections) as well as the longer dosing interval extensions that were permitted for aflibercept 8 mg (up to 24 weeks beginning at week 52) compared with faricimab T&E (up to 16 weeks) [29–32, 34, 35, 37, 38].

The findings of the present NMA may have key implications for real-world clinical practice. Owing to the frequency of intravitreal injections, anti-VEGF therapy imposes a substantial burden on patients, caregivers, and healthcare systems [23]. A less intensive, flexible dosing regimen may have additional benefits for patients, including less time lost attending appointments [52], less disruption to normal life [24], and less anxiety [25–27]. Similarly, the economic burden of anti-VEGF therapy may be alleviated by reducing work absenteeism among employed caregivers, who dedicate substantial time to helping patients attend appointments and may themselves experience anxiety and other mental health issues [28]. Fewer injections would also likely mean reduced costs for the healthcare system. This economic model, based on the number of injections administered in clinical trials, estimated that a reduced aflibercept 8-mg dosing frequency compared with faricimab would translate to lower direct and indirect costs for patients with DME and nAMD (Kuznik et al, Ophthalmol Ther 2025, in press).

NMAs allow for indirect comparisons of interventions from different trials [53]. However, as with all indirect treatment comparisons, NMAs have inherent limitations and cannot replace head-to-head trials [54, 55]. Limitations of this NMA include the impact of cross-trial differences, including trial design, patient eligibility criteria, dosing interval modification criteria, outcome measurement methods, and baseline patient characteristics. With respect to trial designs, efficacy outcomes were evaluated at week 96 in PULSAR and PHOTON but were averaged over weeks 92, 96, and 100 in YOSEMITE and RHINE and over weeks 104, 108, and 112 in TENAYA and LUCERNE. This reflects differences in the observation period across the trials, which ranged from 96 weeks in PHOTON and PULSAR to 112 weeks in TENAYA and LUCERNE. To account for this latter difference, the mean numbers of injections were assessed over two periods and standardized to 104 weeks. With respect to baseline patient characteristics, the proportion of patients with DME who were previously treated with other intravitreal anti-VEGF agents ranged from 20% to 22% in YOSEMITE and RHINE to 44% in PHOTON. Adjustment by prior treatment was not performed, although any bias was anticipated to favor faricimab T&E given that previously treated patients tend to be less responsive to treatment than treatment-naive patients. Changes in CST may have been limited by the ceiling effect if mean CST was high at baseline, such as in the PHOTON (454 µm), YOSEMITE (486 µm), and RHINE (471 µm) trials. To account for this, percentage changes from baseline were calculated and analyzed; however, as percentage changes in CST were not reported for trials of faricimab T&E, assumptions were employed in their calculation and caution should therefore be exercised when interpreting these results. Furthermore, the included trials reported values for CST (YOSEMITE, RHINE, TENAYA, and LUCERNE) and central retinal thickness (PHOTON and PULSAR), which for the purpose of this analysis were considered to be equivalent. However, subtle differences in definitions or measurement protocols between trials may exist [56, 57] and represent a potential limitation of this NMA. With regards to the generalizability of the findings, data from trials conducted in a controlled setting with a homogeneous patient population may not necessarily be representative of real-world clinical practice. This analysis does not include patient-reported outcomes or qualitative data, which could provide additional context on the practical significance of fewer injections. Further research is needed to evaluate how such differences in injection frequency are perceived by patients and whether they translate into meaningful improvements in quality of life or treatment satisfaction. This would provide a more comprehensive understanding of the clinical relevance of these findings. Although this analysis was funded by Regeneron Pharmaceuticals, Inc., this analysis and the drafting of the technical report on which this article was based were performed independently of the sponsor. The NMA included 2-year data from PULSAR and PHOTON that were available as clinical trial reports from the sponsor but were not publicly available and not yet published in a peer-reviewed journal, which was confirmed by the systematic literature review. Although the full 2-year results from the faricimab trials (YOSEMITE, RHINE, TENAYA, and LUCERNE) have been published and were included in the NMA, it is possible that additional relevant but unpublished clinical trial data for faricimab exist.

Conclusions

Results of this NMA show that patients with DME or nAMD who were treated with aflibercept 8 mg achieved similar visual and anatomic outcomes with significantly fewer injections compared with patients treated with faricimab T&E over 2 years. However, owing to limitations of this NMA, prospectively designed head-to-head clinical trials or real-world studies would be required to confirm the findings.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Andreas Kuznik, Yingxin Xu, Steven Sherman, and Nimesh Patel were involved in the conception and design of the study and interpretation of data. Ali Mojebi, Sam Keeping, and Keith Chan conducted the systematic literature review, developed and conducted the NMA, and wrote the technical report, independently from the sponsor (Regeneron Pharmaceuticals, Inc.). All authors were involved in critical review of manuscript drafts and approved the final version for submission.

Funding

This analysis and its publication, including the journal’s Rapid Service Fee, was funded by Regeneron Pharmaceuticals, Inc. Theodore Leng received funding in part from Research to Prevent Blindness and NIH grant P30-EY026877.

Data Availability

The 2-year follow-up data used to inform this NMA were obtained from published manuscripts for the YOSEMITE, RHINE, TENAYA, and LUCERNE trials, and from clinical study reports reporting data through week 96 of the PHOTON and PULSAR trials. Individual anonymized participant data will be considered for sharing (1) once the product and indication has been approved by major health authorities (e.g., US Food and Drug Administration, European Medicines Agency, Pharmaceuticals and Medical Devices Agency, etc.) or development of the product has been discontinued globally for all indications on or after April 2020 and there are no plans for future development, (2) if there is legal authority to share the data, and (3) if there is not a reasonable likelihood of participant re-identification. Submit requests to https://vivli.org/.

Declarations

Conflict of Interest

Scott Friedman has received grants or contracts from Bayer, JAEB Center, Ophthea, Regeneron Pharmaceuticals, Inc., Genentech, Amgen, Clearside Biomedical, EyePoint, ONL Therapeutics, REGENXBIO, Kowa, Kodiak Sciences, and Belite Bio, has participated on a Data Safety Monitoring Board or Advisory Board for Amgen, Genentech, and Ocular Therapeutics, and has stock or stock options in REvOpsis. Yingxin Xu, Andreas Kuznik, and Steven Sherman are employees and stockholders of Regeneron Pharmaceuticals, Inc. Ali Mojebi, Sam Keeping, and Keith Chan are employees of Precision AQ, which received funding from Regeneron Pharmaceuticals, Inc. for conducting this analysis. Theodore Leng has received funding from Astellas and served as a consultant for Astellas, Boehringer Ingelheim, Roche/Genentech, Topcon, Toku, and Virtual Field. Nimesh Patel has served as an advisor for Alcon, Alimera, Allergan, Apellis, Biogen, DORC, EyePoint, Genentech, Kyoto Drug Company, Regeneron Pharmaceuticals, Inc., and REGENXBIO.

Ethical Approval

This NMA uses data from previously conducted clinical trials and does not report any new data from trials conducted in human participants by any of the authors.