Intravitreal anti-VEGF drug use in industrialized nations: why are newly introduced medications causing us to inject more and not less?

Advances in our understanding of the pathophysiology responsible for the most common blinding diseases in developed nations – neovascular age-related macular degeneration (nAMD) and diabetic retinopathy (DR) – have led to significant advances in the treatment of these conditions. Since the discovery of vascular endothelial growth factor (VEGF) in 1989^1,2, three VEGF inhibitory drugs (pegaptanib, ranibizumab, and aflibercept) have been approved for administration via intravitreal injections and a fourth (bevacizumab) is commonly used in an off-label manner. These drugs have been approved in over 80 nations and have become standard of care for the treatment of blinding chorioretinal vascular conditions.

The phase III registration trials for ranibizumab featured monthly intravitreal injections for 6 months (macular edema due to retinal vein occlusions)^3,4, 2 years (nAMD)^5,6, and 3 years (diabetic macular edema)⁷. Monthly injections are the gold standard for the treatment of most patients but physicians recognize that such frequent injections may not be required to produce excellent structural and functional responses in all patients, so many physicians individualize therapy with pro re nata (PRN) and treat-and-extend (T&E) regimens, in which patients are injected at each clinic visit and the intervals between visits are extended until disease activity recurs. The American Society of Retina Specialists annual Preferences and Trends Survey has reported that the majority of physicians use individualized therapy with injections given less frequently than monthly.

Ranibizumab has been approved for monthly injections and only more recently for PRN and T&E injections, whereas aflibercept has been approved for q8wk (every eight weeks) injections for the treatment of both nAMD⁸ and diabetic macular edema (DME)⁹. The comparative phase III trials showed that injections of aflibercept every 8 weeks works as well as q4wk aflibercept for DME and q4wk ranibizumab for nAMD, a feature that some authors have attributed to aflibercept’s high binding affinity for VEGF-A and its prolonged intraocular half-life. Aflibercept’s efficacy and durability in the phase III trials was consistent with a longer duration of action than both ranibizumab and bevacizumab but, other than the 12 week capped PRN regimen in year 2 of the VIEW trials and modified PRN regimen in the DRCR.net Protocol T trial¹⁰, these latter two drugs were not used with longer treatment intervals in head-to-head trials against aflibercept. Nevertheless, many physicians believed that widespread use of aflibercept would decrease the frequency of injections in most patients. Extended duration of action promised to decrease the number of injections, the frequency of clinic visits, and both direct and indirect costs of therapy, which could slow the cost increases that are challenging many national healthcare systems.

To determine whether approval of aflibercept actually led to a decrease in the number of injections administered to patients with newly diagnosed, treatment-naïve chorioretinal vascular conditions during the first year, in this issue of Current Medical Research and Opinion Vorum et al. examined the Danish National Patient Registry and the Civil Registration System¹¹. This large database review reported the mean number of injections per patient but was unable to associate this with the treatment strategies that were employed or the visual changes that resulted.

The authors found that the mean number of injections given to patients with nAMD after aflibercept became available increased by 0.69 during the first year of treatment. This finding is consistent with that reported by several other large database studies and does not support the previously held notion that aflibercept possesses a longer duration of action than ranibizumab, but neither does it support the conclusion that aflibercept has a shorter duration of action that requires more frequent injections.

Without more granular data, we can only speculate as to why aflibercept’s attractive pharmacokinetic properties have not resulted in fewer intravitreal injections for patients with nAMD throughout Denmark and other countries, but a couple of possibilities deserve consideration. Recently published short-term and long-term studies support the notion that visual acuity improvements correlate positively with the frequency of injections. Two-year results from CATT and IVAN show that patients “give up” two letters of best corrected visual acuity (BCVA) when they are treated according to a less intensive PRN regimen instead of monthly¹². Visual acuities from long-term (>5 years) treatment of nAMD vary considerably among studies but better results are generally associated with more aggressive therapy¹³. As a result of these findings, review papers¹⁴, meeting presentations, and CME symposia have been emphasizing the need for frequent injections to not under-treat patients and cause irreversible loss of vision. Physicians may be reacting to these messages by treating more aggressively with a higher frequency of monthly regimens or with stricter retreatment criteria within PRN regimens. Greater use of T&E regimens compared to PRN results in more injections during the first year of therapy, the timeframe assessed by Vorum et al. Excellent results with T&E as reported by the LUCAS trial¹⁵ may have encouraged an increasing number of physicians to adopt this treatment strategy.

If aflibercept, with its favorable pharmacokinetic attributes, cannot extend inter-treatment intervals across a population, can we expect better performance from any other drugs or devices? Let’s examine the current anti-VEGF development pipeline with an eye on the future state. Table 1 lists the drugs, extended release devices, and delivery systems discussed below with commentary regarding their investigational status and important results.

Table 1. Drugs, extended release devices, and delivery systems that prevent the actions of vascular endothelial growth factor (VEGF) are listed together with a summary of their current investigational status and commentary regarding their clinical strengths and weaknesses.

Drug or device	Current trial status	Commentary
Abicipar Pegol DARPin (2 mg/0.05 ml)	Recruiting: phase III AMD trials (CEDAR & SEQUOIA)	Phase I DME data: • Data from 4 patients suggests t1/2 of 13 days Phase II AMD data: • Mixed duration of action from different phase II trials: 1. Time to recurrence c.58 days 2. 2 month and 3 month dosing produced VA results inferior to ranibizumab
RTH258 (antibody fragment) (6 mg/0.05 ml)	Recruiting: phase III AMD trials (2 month and 3 month injection arms vs. aflibercept)	Phase II AMD data: • 12 week results non-inferior to aflibercept (each drug delivered q8weeks) • After week 32, both drugs given q12weeks. RTH eyes required fewer “rescue” injections than aflibercept (60 vs. 94)
Genentech/ForSight 4 refillable reservoir (ranibizumab)	Completed: phase I AMD trial	Phase I AMD data: • 4 significant implantation-related complications (cohort of 20 patients) • 1-year visual acuity results comparable to those from phase III ranibizumab trials • Required mean of 4.8 “refills” of reservoir
Neurotech encapsulated cell reservoir (aflibercept-like protein)	Terminated: multi-center phase II AMD trial due to under-performance vs. aflibercept	Phase I AMD data: • 2 earlier generation devices with lower synthesis rates produced poor results Phase II AMD study: • Third generation anti-VEGF device used in phase II trial (never previously tested)
Avalanche Biotechnologies adenovirus delivery system (AAV) (DNA insertion for soluble receptor)	Completed: phase II AMD trial	Phase II AMD study: • Visual acuity results in AAV group significantly worse than those treated with ranibizumab • Future is uncertain

Abbrev: AMD: age-related macular degeneration; DME: diabetic macular edema; VA, visual acuity; q8weeks, every 8 weeks; q12weeks, every 12 weeks.

For several years developers have worked on drugs and devices that they hoped would possess extended durations of actions. Two injectable anti-VEGF drugs, abicipar, a pegylated designed ankyrin repeat protein (DARPin), and RTH258, a single strand antibody fragment, are currently in phase III trials for the treatment of nAMD. The developers of both drugs hope to demonstrate success with 3 month dosing when compared to monthly ranibizumab or bimonthly aflibercept. Both drugs improved macular morphology and BCVA when administered frequently in phase II studies, but when injection intervals were extended to 3 months, the results appeared less promising. Neither molecule matches aflibercept’s pharmacokinetic characteristics because of their smaller sizes – intraocular half-lives of macromolecules are proportional to their molecular weights – and weaker binding affinities for VEGF₁₆₅. Developers are hoping that pegylation of abicipar will significantly prolong its intraocular half-life and the large dose of RTH258 (6 mg/0.05 ml) will compensate for its small size.

Extended release corticosteroid inserts usually last between 3 months (dexamethasone phosphate biodegradable insert, Ozurdex¹) and 3 years (fluocinolone acetonide non-biodegradable insert, Iluvien²) but sustained release anti-VEGF devices are much more difficult to produce. The molecular weights of anti-VEGF macromolecules (molecular weight of 48–149 kDa) are orders of magnitude greater than those of corticosteroids (0.39 kDa), making it difficult to pack a sufficient number of protein molecules into a device that easily fits into the eye. Compounding this problem is the tendency for large proteins to lose their necessary three-dimensional conformations when exposed to physiological environments for extended periods of time.

Genentech and ForSight 4 are developing a refillable, trans-scleral ranibizumab reservoir¹⁶. In a phase I study, 20 patients achieved improvements in BCVA that were comparable to those in the ranibizumab registration trials. Four surgery-related serious adverse events occurred but none after the implantation technique was modified. The implant had to be refilled a mean of 4.8 times during the 1 year study, but since patients generally receive 5–7 ranibizumab injections when treated PRN, the refill frequency with the reservoir provides little advantage over a PRN regimen.

Neurotech has developed an encapsulated cell device that contains genetically modified, immortalized retinal pigment cells that are programmed to synthesize a molecule resembling aflibercept. A phase II nAMD study with a third generation anti-VEGF implant was terminated early because patients failed to respond adequately¹⁷.

Avalanche Biotechnologies has developed an adenovirus vector delivery system (AVA-101) to insert the DNA for a naturally occurring slt-1 (soluble VEGF receptor-1) into retinal pigment epithelial cells. Infected cells synthesize the soluble VEGF inhibitory protein and excrete it into the outer retina and choriocapillaris. In a phase IIa trial, patients with nAMD received subretinal AVA-101 and 0.5 mg ranibizumab (at baseline and 1 month, and again as rescue therapy) compared to a control group that received monthly ranibizumab. At the 52 week endpoint, mean improvement in BCVA was +2.2 letters in the AVA-101 group compared to +9.3 letters in the ranibizumab group¹⁸.

The disappointing results from these recent studies suggest that extended duration anti-VEGF therapy may not be achievable anytime soon. Drugs that target other pro-angiogenic molecules (platelet-derived growth factor and angiopoietin 2) appear to be promising but these will be used in combination with VEGF inhibition. The pivotal role of VEGF in ocular angiogenesis means that anti-VEGF therapy will likely play a critical role in the treatment of nAMD for years to come.

Physicians, patients, and healthcare policy makers should resign themselves to the continuing need for frequent intravitreal injections. Innovative strategies to control rising costs and provide sufficient manpower to meet the burgeoning need for patient care will be required. Previously published data from Denmark first showed us the powerful ability of anti-VEGF therapy to lower national blindness rates due to AMD¹⁹, and Vorum and colleagues have now shown that this success results from intensive injection regimens from which we are unlikely to experience relief any time soon.

Transparency

Declaration of funding

This editorial was not funded.

Declaration of financial/other relationships

M.W.S. has disclosed that he is on the advisory board and receives institutional research support from Allergan and Regeneron, and is a consultant for Boehringer-Ingelheim and Momenta.

CMRO peer reviewer on this manuscript has no relevant financial relationships to disclose.

Notes

Notes

1 Ozurdex is a registered trade name of Allergan, Irvine, CA.

2 Iluvien is a registered trade name of Alimera, Alpharetta, GA.