Real World Outcomes for AMD: FRB! Registry

The prevalence of vision impairment and blindness in Australia due to neovascular age-related macular degeneration (nAMD) is on the rise due to the growth in the ageing population.¹ Individuals with the disease experience difficulties in their routine activities, such as reading, driving and recognising faces, which frequently result in severe impairment in quality of life and loss of independence. This leads to a substantial increase in demands on the Australian health system and in health care costs.

the Australian health system has much to be proud of in its management of nAMD

The advent of vascular endothelial growth factor (VEGF) inhibitors in the previous decade was a major breakthrough in the management of patients with nAMD. The landmark study, Minimally Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the treatment of Neovascular Age-related macular degeneration (MARINA), reported a mean gain in visual acuity (VA) of 6.5–7.2 letters (Early Treatment Diabetic Retinopathy Study chart) after two years of treatment in eyes that received four weekly ranibizumab treatment.² The other pivotal phase three study of ranibizumab, anti-VEGF antibody for the treatment of predominantly classic choroidal neovascularisation in AMD (ANCHOR), found a mean gain of 8.1–10.7 letters after two years of four weekly ranibizumab treatment.³ The results of these studies were promising, but the significant burden that monthly treatment imposes on patients, their caregivers and the treating physicians makes it difficult to achieve in routine clinical practice. Alternative regimens to monthly treatment, ‘pro re nata’ and ‘treat and extend’, have been developed. These regimens employ fewer injections so they are more feasible in routine clinical practice, however, their results are generally inferior to the monthly treatments used in the clinical trials.^4,5

Clinical trials assess the ‘efficacy’ of an intervention in a controlled environment for a highly selected group of patients that may not be representative of the general population with the disease.⁶ The ‘effectiveness’ of these interventions in the general population in routine clinical practice needs to be established by population based, postmarketing, real-world observational studies.6 There have been instances of subsequent withdrawal of drugs that had shown promising results in the clinical trials, such as the lipid lowering drug cerivastatin, based on post-marketing observational studies.⁷

‘Real world studies’ refers to information on healthcare obtained from non-clinical research settings such as electronic health records, product and disease registries, claims and billing data, and data gathered through personal devices and health applications.⁸ Real world evidence is a way to incorporate a wide array of evidence into information on health care. The evidence can inform outcomes research, drug development, patient care, quality improvement, safety surveillance, research on health care systems and well controlled effectiveness studies.⁹ There is a growing need for, and acceptance of, real world evidence with the transition of the healthcare system to evidencebased practice. Regulatory agencies, such as the United States Food and Drug Administration and European Medicines Agency, can request real world data for safety surveillance and risk assessment.¹⁰ 

The Fight Retinal Blindness! (FRB!) Project is designed as a tool to collect data on outcomes of treatment of macular diseases in routine clinical practice

Real world data can be collected using registries. The Agency for Healthcare Research and Quality defines a registry as “an organised system that uses observational study methods to collect data (clinical and other) to evaluate specified outcomes for a population defined by a particular disease, condition, or exposure, and that serves one or more predetermined scientific, clinical, or policy purposes.”¹¹ Registries capture data from existing administrative systems and databases, from medical records, or directly from the clinical staff using data collection forms. Registries vary greatly, but all depend on how their populations are defined. For instance, product registries include people exposed to a particular healthcare product or surgical device (e.g. a pharmaceutical product), health services registries include patients with particular clinical conditions, interventions or episodes of care (e.g. surgical procedures) and disease registries include patients with the same disease (e.g. retinitis pigmentosa). A combination of these can exist in the same registry.¹¹

The Fight Retinal Blindness! (FRB!) Project is designed as a tool to collect data on outcomes of treatment of macular diseases in routine clinical practice.¹² This webbased registry, developed for data entry, has been facilitating accurate, systematic and standardised data collection across various centres in Australia, New Zealand, Singapore and Switzerland since 2007. It has now been extended to other countries in Europe and Asia. The registry aims to monitor and assess whether the outcomes of treatment in routine practice of the new drugs for macular diseases are consistent with the promising results observed in the pivotal clinical trials. Data can be entered by practitioners or their assistant using any device with a browser during routine follow up clinical visits. It takes less than 15 seconds and generates an interactive graphical summary of a patient’s ‘treatment journey’. Different dosing regimens can be compared with this interface for their ocular safety, long-term effectiveness and patient acceptance. The outcomes achieved by individual drugs, in various clinical settings, can also be compared. The system allows individual physicians to review their own data for self-audit against national benchmarks. Clinicians can capture data for nAMD, choroidal neovascularisation other than nAMD, diabetic macular edema and retinal vein occlusion. Publications from the FRB! Registry have addressed many clinically relevant issues in treating nAMD in the real-world clinical setting. Some of these will be addressed here.

SHORT AND LONG-TERM OUTCOMES

Gillies et al. from the FRB! Project group provided one of the first observational studies on the treatment of nAMD with VEGF inhibitors.¹³ They described the 12 month treatment outcomes and adverse events in over 1,000 treatment-naive eyes (VA >20 letters [Snellen equivalent 6/120] that started VEGF inhibitors from January 2006 to September 2012 in Australia. They reported that VEGF inhibitors achieved reasonably good outcomes in real-world practice, with a mean VA gain of 4.7 letters at 12 months from baseline (mean VA: 57.1 letters) after a mean of seven injections. Lesion characteristics, particularly type and size of the lesion, did not significantly affect the outcomes, which was consistent with retrospective analyses of MARINA and ANCHOR.^3,14 Significant adverse events were infrequent. The mean VA gain observed in this study was higher than other reports around that time, including the Swedish Lucentis Quality Registry, the WAVE study, and an analysis of the German reinjection scheme, most likely because eyes in Australia received more VEGF inhibitor treatments.^15-17 The mean number of VEGF inhibitor treatments over 12 months in Australia was similar to those of the Comparison of Age- Related Macular Degeneration Treatment Trial, a major US clinical trial that used the pro re nata (PRN) regimen.⁴

The outcomes of a subset of the FRB! database participants, that fulfilled the entry criteria of the MARINA study (including VA of 20/20 to 20/320 Snellen equivalent), referred to as the ‘FRBMARINA’ cohort, were compared with those eyes that received ranibizumab in the MARINA study.¹⁸ The mean change in VA at 12 months of the FRB-MARINA cohort (+5.5 letters) was only 1.7 letters less than the actively treated group from the MARINA study (+7.2 letters) and the proportions of eyes losing fewer than 15 letters, the primary outcome of the study, were similar (94.6% MARINA, 91.3% FRB-MARINA). Eyes in the MARINA cohort had received 13 injections over the 13 months while the FRB-MARINA cohort received a mean of only 7.3 injections at 9.5 visits for this similar outcome. This study showed that the outcomes of phase 3 studies, such as MARINA, could be approached in real world routine practice with a modified treatment regimen.

Many of our patients have been treated for over 10 years now but there are few, if any, data on long-term outcomes of VEGF inhibitors for nAMD. A study from the FRB! Registry evaluated seven year outcomes in 1,212 eyes with a mean follow up of 53.5 months.¹⁹ Only 549 (45%) eyes that could have continued treatment through five years did so. The mean VA increased from 55.1 letters (approx. 6/24) at baseline to 61.4 letters (approx. 6/19+1) after six months of treatment and was maintained above the baseline level through five years (Figure 4). The group of 131 eyes that were followed for seven years lost a mean of 2.6 letters from the baseline level. The proportion of eyes with VA ≥70 letters (6/12, driving vision) was 40% and those with VA ≤35 letters (20/200) was 18% at seven years. Geographic atrophy (37%), subretinal fibrosis (31%) and structural damage (11%) were the main causes for a ≥10 letters loss after seven years. A median of six injections from a median of nine visits were recorded in the first year, which decreased to five treatments and seven to nine visits per year throughout the following six years. The high dropout rate of 50% at five years is commonly found in observational studies. The mean VA for these increased initially but decreased markedly to the baseline level around the time treatment was discontinued.

This study established that much better outcomes could be achieved than were reported by the SEVEN-UP study, the only other long-term data available, which reported nearly a 10 letter mean drop in vison at seven years with over 90% of eyes developing macular atrophy. The different outcomes appear to be related to different injection intensities, with the Australian cohort receiving approximately twice as many injections through a treat and extend regimen compared to the US study, which used PRN dosing.

REASONS FOR TREATMENT DISCONTINUATION

A study from the FRB! Registry evaluated the reasons for treatment discontinuation based at a single practice at which 248 patients received ranibizumab for nAMD.²⁰ It evaluated patients who had been treated for up to six years, which was longer than other reports of that time.^15,21 A total of 105 patients (42%) had discontinued treatment during this period.

Reasons unlikely to be related to poor outcomes, such as transfer to another physician (27 [27%]), patient declining treatment (26 [26%]), and death (11 [11%]), were recorded in most (64%) of the 102 eyes in whom the reasons were known.

Reasons that would probably be related to poor outcomes, such as further treatment considered to be futile (27 [27%]), were reported in a minority.

This study suggested that quality of life issues, such as financial burden, frequent treatment visits, and treatment associated pain and discomfort, were less common reasons for permanent discontinuation.

A study by Kruger et al., found a higher percentage (21%) of patients in whom treatment was discontinued because the lesions were believed to be inactive, whereas only 4% of patients declined further treatment. The proportion of patients in which further treatment was believed to be futile was similar.

different outcomes appear to be related to different injection intensities, with the Australian cohort receiving approximately twice as many injections through a treat and extend regimen compared to the US study, which used PRN dosing

TREATMENT REGIMEN

Various regimens of VEGF inhibitor treatments for nAMD have evolved to decrease the burden and risks associated with monthly treatment. These are based mainly on the practitioner’s judgment of underlying activity of the neovascular membrane being treated. One of the first analyses of the PRN regimen, in which eyes are seen regularly but treatment is withheld unless there are signs of lesion activity, reported that visual outcomes could approach those of a fixed monthly treatment schedule with fewer injections.²² Another approach, treat and extend (T&E), also aims to decrease the burden of both the number of treatments and clinic visits by progressively increasing the intervals between treatments after the CNV has been inactivated during the induction phase of treatment to keep the lesion inactive with the fewest possible treatments.

A study from the FRB! Registry reported the first two year outcomes of VEGF inhibitors in nAMD using the T&E approach in real world practice.²³ A mean gain of +5.3 letters at 24 months from baseline was found from a mean of 13 injections at 15 visits. The initial mean VA was similar for all years when treatment started, but the number of injections increased and mean VA gains improved with successive years, from +2.7 letters for eyes starting in 2007 after a mean of 9.7 injections in two years, to +7.8 letters for eyes that started in 2012 after a mean of 14.2 injections over two years. This study provided some of the first evidence that treating nAMD with a T&E regimen in routine clinical practice could produce good results from fewer visits than other approaches.

Aflibercept was first compared with ranibizumab directly in the VEGF Trap- Eye: Investigation of Efficacy and Safety in wet AMD studies (VIEW 1&2).²⁴ This phase 3 clinical trial reported that the 12 month visual outcomes of aflibercept injections every two months (following three initial monthly injections) were ‘non-inferior’ to monthly injections of ranibizumab or, indeed, aflibercept. Both drugs were reported to maintain VA during the second year of treatment under a variable dosing regimen.²⁵ A study from the FRB! Registry compared 12 month treatment outcomes of ranibizumab versus aflibercept in the 197 matched eyes in real world clinical practice.²⁶ The mean improvements in VA in the ranibizumab and aflibercept groups at 12 months were similar (3.0 vs. 4.9 letters, p = 0.26) with a similar mean number of injections (8.1 vs. 8.0 injections, p = 0.27). There were more treatment switches from ranibizumab to aflibercept than vice versa. Eyes that discontinued treatment were similar in both the groups. This study found that neither ranibizumab nor aflibercept was superior to the other in terms of VA outcomes and treatment frequency at 12 months for nAMD in Australian real world practice.

The treatment patterns of nAMD during the maintenance phase, which begins once the underlying lesion is inactivated, have been analysed by the registry.²⁷ The lesion became inactive in two thirds of the eyes after treatment with ≤3 VEGF inhibitor injections (short induction phase). The lesion reactivated in half of the eyes within the first year of the maintenance phase, the most common reactivation interval being eight weeks. The greatest risk of reactivation occurred when the interval between injections was extended beyond 12 weeks, reaching 37% per visit if the interval was ≥20 weeks. Eyes with longer induction phases reactivated earlier than the eyes with shorter induction phases. This study provides evidence that better outcomes may be achieved if extension of intervals between VEGF inhibitors for nAMD is capped at three monthly.

PERSISTENTLY ACTIVE LESIONS

The activity of a neovascular lesion is an important consideration when practitioners plan nAMD treatment with VEGF inhibitors. A study from the registry evaluated outcomes of eyes with nAMD that never became inactive despite ongoing treatment with VEGF inhibitors.²⁸ Lesions were graded ‘active’ at each visit if there was ‘subretinal or intraretinal fluid attributable to leak from the choroidal neovascular lesion or fresh hemorrhage’. Eyes were divided into four groups (‘persistent’, ‘high’, ‘moderate’, and ‘low’) according to the proportion of visits at which each eye was graded active during the first 12 months of treatment. The mean VA change was similar in all the four groups after 12 months of treatment. The mean number of intravitreal injections administered over 12 months increased only modestly with increasing lesion activity (7.7, 8.0, 8.5 and 8.3, respectively). This study showed that persistent activity, in eyes with nAMD on VEGF inhibitor treatments for 12 months, was not necessarily associated with worse visual outcomes. This surprised us at the time, but subsequently it has become apparent from the CATT study and our own data that a degree of lesion activity may reduce the risk of developing macular atrophy.²⁹

EFFECT OF TREATMENT SWITCH

Clinicians may switch treatment from one VEGF inhibitor to another hoping for a more favourable effect. The effect of switching treatment was studied in 384 eyes with nAMD on ranibizumab treatment (mean duration 39.8 months) that had switched to aflibercept.³⁰ The mean VA did not change 12 months after switching treatment. The number of treatments decreased by one injection per year in the 12 months after the switch with a modest increase in the treatment interval and a decrease in the proportion of eyes that was graded active after switching.

RISK OF ENDOPHTHALMITIS

The major risk of treatment with VEGF inhibitors – infectious endophthalmitis – is rare but potentially devastating.³¹ A study from the Registry reported the incidence and outcomes of endophthalmitis from 88,150 VEGF inhibitor injections in 4,564 nAMD patients.³² Some of these patients (8%) had completed 10 years from the start of treatment and had received up to 60 intravitreal injections. The incidence of infectious endophthalmitis (18 cases [0.02%]) and non-infectious endophthalmitis (11 cases [0.01%]) was low. There was no increase in the risk of endophthalmitis (both infectious and non-infectious) with each successive injection. The mean VA at 12 months after endophthalmitis was within 10 letters of the pre-endophthalmitis level in half of the cases with infectious endophthalmitis and three-quarters of cases with noninfectious endophthalmitis. The incidence of infectious endophthalmitis observed in this study was lower than reported in the MARINA clinical trial (0.05%). The evidence on the low risk of endophthalmitis with VEGF inhibitor treatments in real world practice was reassuring.

UPCOMING PUBLICATIONS

Data from the FRB! Project have shown that the Australian health system has much to be proud of in its management of nAMD. The project will continue to provide Australian practitioners with software free of charge to track their patients’ outcomes, to compare these with their peers and to contribute to the national outcomes data, which provide a solid political foundation to support intravitreal therapy. We are looking forward to seeing whether the new VEGF inhibitor due to be released next year, brolucizumab, actually does last longer than the other agents in real world clinical practice in Australia.

To earn your CPD points from this article, answer the assessment available at mieducation.com/ AMD-Real-World-Outcomes-FRB 

Dr Sanjeeb Bhandari, MBBS, MD (Ophthalmology) is an ophthalmic surgeon from Kathmandu, Nepal. He is undertaking research for his PhD under the supervision of Professor Mark Gillies in the Macula Research Group at the Save Sight Institute, The University of Sydney. His research uses data from the international Fight Retinal Blindness! Registry of real-world treatment outcomes for macular diseases. His research interests include treatment outcomes of retinal diseases and uveitis. 

Professor Mark Gillies, MBBS, PhD, FRANZCO, a University of Sydney Medical School Foundation Fellow and a clinician-scientist, is the Director of Research, Save Sight Institute and of the Macula Research Group of the University of Sydney. He specialises in medical retina, emerging treatments for macular degeneration and diabetic eye disease. He is the chief investigator of the Save Sight Registries that includes Fight Retinal Blindness! Registry. He has published over 230 articles in high-ranked peer-reviewed journals and book chapters. He has given over 85 podium presentations as an invited speaker at international and national meetings in the last 10 years. 

References 

1. Eye Health in Australia – A background paper to the National Framework for Action to Promote Eye Health and Prevent Avoidable Blindness and Vision Loss. Commonwealth of Australia 2005. (Accessed August 08, 2019, at www1.health.gov.au.) 
2. Rosenfeld PJ, Brown DM, Heier JS, et al. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med 2006;355:1419-31. 
3. Brown DM, Michels M, Kaiser PK, Heier JS, Sy JP, Ianchulev T. Ranibizumab versus verteporfin photodynamic therapy for neovascular age-related macular degeneration: Two-year results of the ANCHOR study. Ophthalmology 2009;116:57-65.e5. 
4. The CATT Research Group. Ranibizumab and Bevacizumab for Neovascular Age-Related Macular Degeneration. N Engl J Med 2011;364:1897-908. 
5. Berg K, Pedersen TR, Sandvik L, Bragadottir R. Comparison of ranibizumab and bevacizumab for neovascular age-related macular degeneration according to LUCAS treat-and-extend protocol. Ophthalmology 2015;122:146-52. 
6. Flay BR. Efficacy and effectiveness trials (and other phases of research) in the development of health promotion programs. Prev Med 1986;15:451-74. 
7. Furberg CD, Pitt B. Withdrawal of cerivastatin from the world market. Curr Control Trials Cardiovasc Med 2001;2:205-7. 
8. Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices. U.S. Food and Drug Administration, 2017 (Accessed 28 December, 2017, at www.fda.gov/downloads/MedicalDevices/ DeviceRegulationandGuidance/GuidanceDocuments/ UCM513027.pdf.) 
9. Sherman RE, Anderson SA, Dal Pan GJ, et al. Real- World Evidence – What Is It and What Can It Tell Us? N Engl J Med 2016;375:2293-7. 
10. Lis Y, Roberts MH, Kamble S, J JG, Raisch DW. Comparisons of Food and Drug Administration and European Medicines Agency risk management implementation for recent pharmaceutical approvals: report of the International Society for Pharmacoeconomics and outcomes research risk benefit management working group. Value Health 2012;15:1108-18. 
11. Gliklich RE, Dreyer NA, Leavy MB. AHRQ Methods for Effective Health Care. Registries for Evaluating Patient Outcomes: A User’s Guide. 3rd ed. Rockville (MD): Agency for Healthcare Research and Quality (US); 2014. 
12. Gillies MC, Walton R, Liong J, et al. Efficient capture of high-quality data on outcomes of treatment for macular diseases: the fight retinal blindness! Project. Retina 2014;34:188-95. 
13. Gillies MC, Walton R, Simpson JM, et al. Prospective audit of exudative age-related macular degeneration: 12-month outcomes in treatment-naive eyes. Invest Ophthalmol Vis Sci 2013;54:5754-60. 
14. Kaiser PK, Blodi BA, Shapiro H, Acharya NR. Angiographic and optical coherence tomographic results of the MARINA study of ranibizumab in neovascular age-related macular degeneration. Ophthalmology 2007;114:1868-75. 
15. Hjelmqvist L, Lindberg C, Kanulf P, Dahlgren H, Johansson I, Siewert A. One-Year Outcomes Using Ranibizumab for Neovascular Age-Related Macular Degeneration: Results of a Prospective and Retrospective Observational Multicentre Study. J Ophthalmol 2011;2011:405724. 
16. Finger RP, Wiedemann P, Blumhagen F, Pohl K, Holz FG. Treatment patterns, visual acuity and quality-of-life outcomes of the WAVE study – a noninterventional study of ranibizumab treatment for neovascular age-related macular degeneration in Germany. Acta Ophthalmol 2013;91:540-6. 
17. Heimes B, Lommatzsch A, Zeimer M, et al. Long-term visual course after anti-VEGF therapy for exudative AMD in clinical practice evaluation of the German reinjection scheme. Graefes Arch Clin Exp Ophthalmol 2011;249:639-44. 
18. Gillies MC, Walton RJ, Arnold JJ, et al. Comparison of outcomes from a phase 3 study of age-related macular degeneration with a matched, observational cohort. Ophthalmology 2014;121:676-81. 
19. Gillies MC, Campain A, Barthelmes D, et al. Long-Term Outcomes of Treatment of Neovascular Age-Related Macular Degeneration: Data from an Observational Study. Ophthalmology 2015;122:1837-45. 
20. Vaze A, Fraser-Bell S, Gillies M. Reasons for discontinuation of intravitreal vascular endothelial growth factor inhibitors in neovascular age-related macular degeneration. Retina 2014;34:1774-8. 
21. Kruger Falk M, Kemp H, Sorensen TL. Fouryear treatment results of neovascular age-related macular degeneration with ranibizumab and causes for discontinuation of treatment. Am J Ophthalmol 2013;155:89-95.e3. 
22. Martin DF, Maguire MG, Fine SL, et al. Ranibizumab and bevacizumab for treatment of neovascular age-related macular degeneration: two-year results. Ophthalmology 2012;119:1388-98. 
23. Arnold JJ, Campain A, Barthelmes D, et al. Two-year outcomes of “treat and extend” intravitreal therapy for neovascular age-related macular degeneration. Ophthalmology 2015;122:1212-9. 
24. Heier JS, Brown DM, Chong V, et al. Intravitreal aflibercept (VEGF trap-eye) in wet age-related macular degeneration. Ophthalmology 2012;119:2537-48. 
25. Schmidt-Erfurth U, Kaiser PK, Korobelnik JF, et al. Intravitreal aflibercept injection for neovascular age-related macular degeneration: ninety-six-week results of the VIEW studies. Ophthalmology 2014;121:193-201. 
26. Gillies MC, Nguyen V, Daien V, Arnold JJ, Morlet N, Barthelmes D. Twelve-month outcomes of ranibizumab vs. aflibercept for neovascular age-related macular degeneration: data from an observational study. Ophthalmology 2016;123:2545-53. 
27. Essex RW, Nguyen V, Walton R, et al. Treatment patterns and visual outcomes during the maintenance phase of treat-and-extend therapy for age-related macular degeneration. Ophthalmology 2016;123:2393-400. 
28. Barthelmes D, Walton R, Campain AE, et al. Outcomes of persistently active neovascular age-related macular degeneration treated with VEGF inhibitors: observational study data. Br J Ophthalmol 2015;99:359-64. 
29. Grunwald JE, Pistilli M, Daniel E, et al. Incidence and growth of geographic atrophy during 5 years of comparison of age-related macular degeneration treatments trials. Ophthalmology 2017;124:97-104. 
30. Barthelmes D, Campain A, Nguyen P, et al. Effects of switching from ranibizumab to aflibercept in eyes with exudative age-related macular degeneration. Br J Ophthalmol 2016;100:1640-5. 
31. Jager RD, Aiello LP, Patel SC, Cunningham ET, Jr. Risks of intravitreous injection: a comprehensive review. Retina 2004;24:676-98. 
32. Daien V, Nguyen V, Essex RW, Morlet N, Barthelmes D, Gillies MC. Incidence and outcomes of infectious and noninfectious endophthalmitis after intravitreal injections for age-related macular degeneration. Ophthalmology 2018;125:66-74.