Modern medicine is predicated on making clinical decisions based upon the best available systematic research in what is termed evidence-based medicine. In many fields, much of this evidence emerges from clinical trials, which routinely provide the highest quality data available to guide decision-making by policymakers, patients, and clinicians — all of whom represent the end users of these data.1 A critical feature of clinical trial design is selecting endpoints, which, if appropriate, render the trial valid, relevant, and therefore useful to the end users. In late-phase clinical trials, endpoints are generally chosen such that they may be statistically analyzed to help determine the efficacy of an intervention being studied.
In some instances, a single clinical trial endpoint would be inadequate to fully encapsulate the effects of the intervention to all end users. In other cases, measuring only a single endpoint would be a wasted opportunity because the trial has a significant financial cost, involves a relatively risky intervention, or involves a rare patient population.1,2 In instances where multiple endpoints are specified, the primary endpoint is used to determine whether the intervention met its goal and is the primary basis on which regulatory approval is granted. Care must be taken when multiple primary endpoints are used and meeting any one of them is proposed as sufficient to determine an intervention’s efficacy, as this may lead to a type I error.3
In contrast, secondary endpoints are supportive: they offer further support of an intervention’s efficacy or the mechanism of therapeutic action. Tertiary, or exploratory, endpoints are chosen to explore a novel hypothesis related to the intervention, but generally involve effects too rare for robust statistical analysis. Endpoints may also be classified as clinical (how a patient “feels, functions, or survives”) or nonclinical, the latter of which includes biomarkers. These are patient characteristics that are “objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention.”1-3 Review of clinical trials in medical retina for neovascular (“wet”) age-related macular degeneration (nAMD) illustrates many of these facets, as well as the need for changing clinical trial endpoints as therapeutic interventions evolve.
Fundus photograph of a right eye with new submacular hemorrhage from neovascular macular degeneration.
A History of Wet AMD Trials
Historically, the first widely used treatment for nAMD was laser photocoagulation, which was supported by 3 multicenter randomized controlled clinical trials sponsored by the National Eye Institute under the umbrella of the Macular Photocoagulation Study.4 Because these studies largely took part in the 1980s, eligibility was based on clinical examination and angiographic features of choroidal neovascular membranes. Primary endpoints were the change from baseline in best-corrected visual acuity (BCVA) using a Bailey and Lovie chart developed in the 1970s.5
The first FDA-approved pharmacotherapy for nAMD was verteporfin photodynamic therapy (PDT; Visudyne; Novartis) in 1999. Preclinical experimentation suggested that parameters of 6 or 12 mg/m2 verteporfin dose, irradiance of 600 mW/cm2, and fluence of 50 to 150 J/cm2 might be effective while minimizing the risk of retinal damage. Phase 1/2 clinical trials relied on the primary endpoint of post-treatment “extent of fluorescein leakage” from choroidal neovascular membranes (CNVMs) on fluorescein angiography.6 Exploratory endpoints included “subretinal hemorrhage, RPE atrophy, and retinal arteriolar, venular, and capillary nonperfusion.”6 Change from baseline in best-corrected visual acuity (BCVA) was used to assess safety of PDT (as opposed to effectiveness) because the trials were uncontrolled, only 12 weeks in duration, and BCVA varied greatly even in untreated disease. Results of 1 treatment demonstrated safety and effectiveness at 1 to 4 weeks; however, leakage recurred by 12 weeks though often was less extensive when retreatment protocols were tested.6,7
Given these results, phase 3 clinical trials called the Treatment of Age-related Macular Degeneration with Photodynamic Therapy (TAP, for classic CNVMs) began in late 1996. Unlike phase 1/2 trials, the primary endpoint in these trials was the percentage of eyes with fewer than 3 lines of visual acuity loss from baseline on modified Early Treatment Diabetic Retinopathy Study (ETDRS) charts, as these were larger trials with longer duration and intent-to-treat analysis. At 12 months, 61% of PDT-treated eyes vs 46% of placebo eyes met this endpoint, with this effect reaching statistical significance in predominantly classic CNVM lesions. Secondary endpoints assessed extent of fluorescein leakage from the classic and occult areas of the CNVM lesions. In a corollary study, called Verteporfin in Photodynamic Therapy (VIP, for occult or presumed early classic CNVMs), the primary outcome was the same. Secondary outcomes similarly revolved around angiographic features of the CNVMs (including development of classic features) but also included aspects of visual acuity such as drop below 34 ETDRS letters (~20/200).8
It is notable that several aspects of VIP and TAP influenced the pivotal trials that heralded the beginning of the anti-VEGF era for nAMD treatment. Phase 3 clinical trials for pegaptanib sodium (Macugen; Pfizer), a nucleic acid aptamer that binds to vascular endothelial growth factor 165 (VEGF165), also used percentage of patients losing less than 15 ETDRS letters at 54 weeks of treatment as a primary endpoint (67% to 73% of treatment vs 53% to 60% of placebo).9 Secondary endpoints related to vision in this trial also demonstrated a positive effect though were not particularly robust. For example, 33% vs 23% of patients did not lose any vision in treatment vs placebo arms, respectively.9
Because none of these treatments yielded “clinically significant improvements in vision,”10 there was a move toward the modern era of treatment with more effective anti-VEGF drugs. ANCHOR was a 24-month trial comparing intravitreal ranibizumab (Lucentis; Genentech) with PDT in classic CNVMs with the primary endpoint again being percentage of eyes avoiding 15 letters or more loss of vision at 12 months (over 94% of ranibizumab vs ~64% of PDT patients achieved this).10 MARINA was an analogous study for minimally classic or occult CNVMs with the same primary endpoint and nearly identical cohort outcomes.11 Although the primary endpoints were relatively impressive in and of themselves, the secondary endpoints with respect to visual acuity provided overwhelming evidence of ranibizumab’s effectiveness in treating CNVMs: visual acuity improved by 15 letters or more in over 40% vs 5.6% of the patients in the 0.5 mg ranibizumab arm vs PDT arms, respectively.10
Extension Studies and Noninferiority Trials
The effectiveness of ranibizumab even in the phase 1/2 trials changed clinical trials for nAMD going forward. Specifically, after completion of phase 1/2 trials, extension studies for patients emerged, though a lag sometimes existed between their last ranibizumab injection and enrollment in the extension with subsequent ranibizumab injections. During that lag, patients were monitored with optical coherence tomography (OCT) and fluorescein angiography with OCT proving more sensitive to detecting recurrent disease activity.12 This laid the groundwork for the more widespread adoption of OCT, as a subsequent clinical trial, termed PrONTO, launched to examine variable dosing regimens (pro re nata, or PRN). Although it was not a registration study, this was the first major study where treatment was guided by OCT findings and used primary endpoints of a) change in visual acuity from baseline, b) OCT data, and c) the number of intraocular injections performed over 2 years. Also, PRN retreatment also partially relied on OCT retreatment guidelines (initially worsening of the central retinal thickness by 100 µm, later modified to quantitative OCT changes “that suggested recurrent fluid in the macula”).12
Another change in clinical trials following MARINA and ANCHOR stemmed from the relatively high bar set by these trials in treating nAMD. Because no FDA-approved drug to date has demonstrated a statistically significant improvement over monthly ranibizumab in preventing a 15-letter loss of visual acuity, clinical trials pivoted toward demonstrating noninferiority. Although the FDA recognizes 4 types of “well-controlled” studies (no treatment, placebo, dose-response, and noninferiority), only noninferiority uses an active control of known effectiveness and is not designed to show superiority of an experimental treatment.13 Like many fields pivoting towards noninferiority clinical trials, not using active control trials in patients with nAMD became unethical following MARINA and ANCHOR. Importantly, noninferiority studies are designed to demonstrate that a new treatment is effective, not that it is as effective as the active control.13 Specifically, a new drug is considered effective if it is not less effective than the active control intervention by a prespecified amount termed the noninferiority margin, M.13
Because a placebo arm is not present in noninferiority studies, the effect of the active control relative to placebo is inferred on past randomized controlled trials, in this case MARINA and ANCHOR. For example, in the VEGF Trap-Eye: Investigation of Efficacy and Safety in Wet AMD phase 3 clinical trial studies (VIEW 1 and VIEW 2 for aflibercept [Eylea; Regeneron]), investigators worked with regulators, including the FDA, to determine appropriate noninferiority margins and “clinical equivalence” based on MARINA and ANCHOR.14 The primary clinical trial endpoint was noninferiority to ranibizumab in the proportion of patients losing less than 15 ETDRS letters at 52 weeks with an M of 10% establishing noninferiority and of 5% establishing “clinical equivalence.” This 10% level of noninferiority was thought to represent two-thirds of the treatment effect of ranibizumab above untreated disease. Secondary endpoints included functional changes, such as mean change in visual acuity and change in NEI VFQ-25 score from baseline, but also anatomic changes such as CNV size on fluorescein angiography and retinal thickness on OCT. Notably, aflibercept demonstrated noninferiority as well as clinical equivalence.
Comparison of Age-related Macular Degeneration Treatments Trials (CATT) was a similar noninferiority study comparing regimens of bevacizumab (off-label Avastin; Genentech) to similar ones of ranibizumab, though its primary endpoint was mean change in visual acuity at 1 year. Secondary outcomes included number of injections and change in fluid accumulation on OCT.15
Durability of Treatment
A departure from monthly ranibizumab as a comparator occurred in the registration trials for faricimab (Vabysmo; Genentech) as well as brolucizumab (Beovu; Novartis). By way of background, phase 2 trials for faricimab (AVENUE and STAIRWAY) did use monthly ranibizumab as the active comparator against varied dosing and frequency (including monthly) of faricimab. Primary endpoints in these trials were mean change in BCVA from baseline to week 36 (AVENUE) and week 40 (STAIRWAY) and from week 12 to week 36 (AVENUE). Secondary endpoints included mean central macular thickness on OCT, changes in CNVM characteristics on fluorescein angiography, and percentage of patients gaining or not losing at least 15 ETDRS letters.16,17
It is notable that AVENUE was not designed as a noninferiority study, and therefore did not meet its primary endpoint of demonstrating faricimab’s superiority to monthly ranibizumab. However, despite this, investigators deemed a phase 3 clinical trial to be warranted, because anatomic outcomes and differences in BCVA were not statistically different than monthly ranibizumab in AVENUE and that extended dosing regiments of faricimab in STAIRWAY were similarly comparable to monthly ranibizumab.16,17
Despite faricimab’s phase 2 clinical trial design and historical precedent, investigators of faricimab’s phase 3 clinical trials (LUCERNE and TENAYA) decided to use 8-week intervals (Q8W) of aflibercept as the active comparator. They explained this decision as rooted in aflibercept’s dosing (2 mg every 8 weeks following a loading phase) being standard worldwide, thus satisfying regulatory requirements.18 They further noted that aflibercept as a comparator aligned well with the practice patterns of most retina specialists19 and that faricimab was designed for extended dosing (up to 16 weeks) to address a desire by clinicians and patients for more durable and more effective treatment options, so it was most appropriate to use a comparator with a standard dosing closer in duration.18,19 Phase 3 investigators for brolucizumab also cite the move toward extended dosing intervals as using Q8W aflibercept as a control. By this rationale, such clinical trial designs were appropriately tailored to the end users of the data: regulatory agencies, clinicians, and patients.18
As clinical trials continue to move toward testing agents designed to be more durable and effective, trial design continues to change. In ASCENT, a phase 3 clinical trial for ABBV-RGX-314 gene therapy (Regenxbio), investigators determined that a primary outcome of noninferiority in BCVA change at 1 year from baseline compared to biweekly aflibercept was most suitable, though rescue therapy was permissible. In contrast, combination therapies must meet primary endpoints of superior visual outcomes as opposed to simply noninferiority. For example, sozinibercept’s (Opthea) phase 3 trials COAST (sozinibercept plus aflibercept) and ShORe (sozinibercept plus ranibizumab) use biweekly aflibercept and monthly ranibizumab, respectively, as active comparators, with their primary endpoint being mean change in baseline BCVA at 52 weeks. Because Opthea investigators cite a desire to improve upon the visual acuity decline during the second year of treatment in TENAYA and LUCERNE, patients in COAST and ShORe will be followed until week 100.20
It remains to be seen how future attempts to create even more durable and effective therapies will continue this trend of clinical trial design and endpoint selection changing to meet the needs of regulators, clinicians, and patients. RP
References
1. McLeod C, Norman R, Litton E, Saville BR, Webb S, Snelling TL. Choosing primary endpoints for clinical trials of health care interventions. Contemp Clin Trials Commun. 2019;16:100486. doi:10.1016/j.conctc.2019.100486
2. Strimbu K, Tavel JA. What are biomarkers? Curr Opin HIV AIDS. 2010;5(6):463-466. doi:10.1097/COH.0b013e32833ed177
3. O’Neill RT. Secondary endpoints cannot be validly analyzed if the primary endpoint does not demonstrate clear statistical significance. Control Clin Trials. 1997;18(6):550-567. doi:10.1016/s0197-2456(97)00075-5
4. Argon laser photocoagulation for senile macular degeneration. Results of a randomized clinical trial. Arch Ophthalmol. 1982;100(6):912-918. doi:10.1001/archopht.1982.01030030920003
5. Bailey IL, Lovie JE. New design principles for visual acuity letter charts. Am J Optom Physiol Opt. 1976;53(11):740-745. doi:10.1097/00006324-197611000-00006
6. Schmidt-Erfurth U, Miller JW, Sickenberg M, et al. Photodynamic therapy with verteporfin for choroidal neovascularization caused by age-related macular degeneration: results of retreatments in a phase 1 and 2 study [published correction appears in Arch Ophthalmol 2000;118(4):488]. Arch Ophthalmol. 1999;117(9):1177-1187. doi:10.1001/archopht.117.9.1177
7. Treatment of Age-related Macular Degeneration with Photodynamic Therapy (TAP) Study Group. Photodynamic therapy of subfoveal choroidal neovascularization in age-related macular degeneration with verteporfin: 1-year results of 2 randomized clinical trials — TAP report. [published correction appears in Arch Ophthalmol 2000;118(4):488]. Arch Ophthalmol. 1999;117(10):1329-1345.
8. Verteporfin in Photodynamic Therapy Study Group. Verteporfin therapy of subfoveal choroidal neovascularization in age-related macular degeneration: 2-year results of a randomized clinical trial including lesions with occult with no classic choroidal neovascularization — verteporfin in photodynamic therapy report 2. Am J Ophthalmol. 2001;131(5):541-560. doi:10.1016/s0002-9394(01)00967-9
9. Gragoudas ES, Adamis AP, Cunningham ET Jr, Feinsod M, Guyer DR; VEGF Inhibition Study in Ocular Neovascularization Clinical Trial Group. Pegaptanib for neovascular age-related macular degeneration. N Engl J Med. 2004;351(27):2805-2816. doi:10.1056/NEJMoa042760
10. Brown DM, Kaiser PK, Michels M, et al. Ranibizumab vs verteporfin for neovascular age-related macular degeneration. N Engl J Med. 2006;355(14):1432-1444. doi:10.1056/NEJMoa062655
11. Rosenfeld PJ, Brown DM, Heier JS, et al. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med. 2006;355(14):1419-1431. doi:10.1056/NEJMoa054481
12. Lalwani GA, Rosenfeld PJ, Fung AE, et al. A variable-dosing regimen with intravitreal ranibizumab for neovascular age-related macular degeneration: year 2 of the PrONTO Study. Am J Ophthalmol. 2009;148(1):43-58.e1. doi:10.1016/j.ajo.2009.01.024
13. US Department of Health and Human Services, et al. Noninferiority clinical trials to establish effectiveness: guidance for industry. November 2016. Accessed September 17, 2024. https://www.fda.gov/media/78504/download
14. Heier JS, Brown DM, Chong V, et al. Intravitreal aflibercept (VEGF trap-eye) in wet age-related macular degeneration [published correction appears in Ophthalmology. 2013;120(1):209-10]. Ophthalmology. 2012;119(12):2537-2548. doi:10.1016/j.ophtha.2012.09.006
15. Comparison of Age-related Macular Degeneration Treatments Trials (CATT) Research Group, Martin DF, Maguire MG, et al. Ranibizumab and bevacizumab for treatment of neovascular age-related macular degeneration: 2-year results. Ophthalmology. 2012;119(7):1388-1398. doi:10.1016/j.ophtha.2012.03.053
16. Sahni J, Dugel PU, Patel SS, et al. Safety and efficacy of different doses and regimens of faricimab vs ranibizumab in neovascular age-related macular degeneration: the AVENUE phase 2 randomized clinical trial. JAMA Ophthalmol. 2020;138(9):955-963. doi:10.1001/jamaophthalmol.2020.2685
17. Khanani AM, Patel SS, Ferrone PJ, et al. Efficacy of every 4 monthly and quarterly dosing of faricimab vs ranibizumab in neovascular age-related macular degeneration: the STAIRWAY phase 2 randomized clinical trial [published correction appears in JAMA Ophthalmol. 2020;138(9):1006]. JAMA Ophthalmol. 2020;138(9):964-972. doi:10.1001/jamaophthalmol.2020.2699
18. Khanani AM, Guymer RH, Basu K, et al. TENAYA and LUCERNE: rationale and design for the phase 3 clinical trials of faricimab for neovascular age-related macular degeneration. Ophthalmol Sci. 2021;1(4):100076. doi:10.1016/j.xops.2021.100076
19. American Society of Retina Specialists. ASRS 2020 preferences and trends membership survey. December 5, 2021. Accessed September 17, 2024. https://www.asrs.org/asrs-community/pat-survey
20. LifeSci Events. Opthea key opinion leader event on sozinibercept (OPT-302) in wet age-related macular degeneration. April 4, 2024. Accessed September 11, 2024. https://lifescievents.com/event/opthea/
