Categories
Feature Articles

The blind spot on Australia’s PBS: review of anti-VEGF therapy for neovascular age-related macular degeneration

v6_i1_a24

Case scenario

A 72 year old male with a two-day history of sudden blurred vision in his left eye was referred to an ophthalmologist at a regional Australian setting. On best corrected visual acuity (BCVA) testing his left eye had reduced vision (6/12-1) with metamorphopsia. Fundoscopy showed an area of swelling around the left macula and optical coherence tomography and fundus fluorescein angiography later confirmed pigment epithelial detachment of his left macula and subfoveal choroidal  neovascularisation.  He  was  given  a  diagnosis  of  wet macular degeneration and was commenced on monthly ranibizumab (Lucentis®) injections – a drug that costs the Australian health care system approximately AUD $1430 per injection and will require lifelong treatment. Recent debate has risen regarding the optimum frequency of dosing and the necessity of this expensive drug, given the availability of a cheaper alternative.

Introduction

Age-related macular degeneration (AMD) is the leading cause of blindness in Australia. [1] It predominantly affects people aged over 50 years and impairs central vision.  In Australia the cumulative incidence of early AMD for those aged over 49 years is 14.1% and 3.7% for late AMD. [1] Macular degeneration occurs in two forms. Dry macular (non- neovascular) disease comprises 90% of AMD and has a slow progression characterised by drussen deposition underneath the retinal pigment epithelium. [2] Currently there is no agreed treatment of advanced dry AMD and is managed only by diet and lifestyle. [3,4] Late stages of dry macular degeneration can result in “geographic atrophy” causing progressive atrophy of the retinal pigment epithelium, choriocapillaries and photoreceptors. [2]

Wet (neovascular) macular degeneration is less common and affects 10% of AMD patients but causes rapid visual loss. [2] It is characterised by choroidal  neovascularisation  (CNV)  secondary  to  the  effects of vascular endothelial growth factor (VEGF) causing blood vessels to grow from the choroid towards the retina. Leakage of these vessels leads to retinal oedema, haemorrhage and fibrous scarring. When the central and paracentral areas are affected it can result in loss of central vision. [2,5] Untreated, this condition can result in one to three lines of visual acuity lost on the LogMAR chart at three months and three to four lines by one year. [6] Hence visual impairment from late AMD leads to significant loss of vision and quality of life.

Currently there are three main anti-VEGF drugs available for wet macular degeneration:   ranibizumab    (Lucentis®),   bevacizumab    (Avastin®) and aflibercept (Eylea®). This feature article attempts to summarise the development in treatments of wet macular degeneration and highlights the current controversies regarding the optimal drug and frequency of dosing in context of cost to the Australian Pharmaceutical Benefits Scheme (PBS).

Earlier treatments for wet AMD

Neovascular (wet) AMD was largely untreatable over a decade ago but the management has transformed over this period. [2] Initially laser  photocoagulation  was  used  in  the  treatment  of  wet  AMD with the aim of destroying the choroidal neovascular membrane by coagulation.  During the 1980s, the Macular Photocoagulation study reported favourable outcomes for direct laser photocoagulation in small classic extrafoveal and juxtafoveal choroidal neovascularisation (CNV). However the outcomes for subfoveal lesions were poor and laser photocoagulation was limited by lack of stabilisation of vision, high reoccurrence rates in 50%, risk of immediate moderate visual loss in 41% and laser induced permanent central scotomata in sub-foveal lesions. [2,7]

During the 1990s photodynamic therapy (PDT) with verteporfin was introduced. It involved a two stage process: an intravenous infusion of verteporfin that preferentially accumulated in the neovascular membranes, followed by activation with infrared light that generated free radicals promoting closure of blood vessels. The TAP trial reported that the visual acuity benefits of verteporfin therapy in predominantly classic CNV subfoveal lesions was safely sustained for five years. [8] However the mean visual change was still a 13-letter average loss for PDT compared with a 19-letter average loss for untreated controls. [2,9] In addition, photosensitivity, headaches, back pain, chorioretinal atrophy and acute visual loss were observed in 4% as adverse effects. [2]

Anti-VEGF therapies

A breakthrough in treatment came during the mid-2000s with the identification of VEGF as the pathophysiological mechanism in driving the choroidal neovascularisation and associated oedema. This led to the establishment of the first anti VEGF drug, pegatanib sodium, an RNA aptamer that specifically targeted VEGF-165. [10] The VISION trial,  involving 1186 patients with subfoveal AMD receiving pegatanib injections every six weeks, had 70% of patients with stabilised vision (less than three lines of vision loss) compared to 55% of sham controls; yet still only a minority of patients actually gained vision. [10]

A second anti-VEGF agent, bevacizumab (Avastin®) soon came into off- label use. Bevacizumab was initially developed by the pharmaceutical company Genetech® to inhibit the tumour angiogenesis in colorectal cancer but its mechanism of action as a full length antibody that binds to all VEGF isoforms proved to have multiple purposes. Despite a lack of clinical trials to support its use in wet AMD, anecdotal evidence led ophthalmologists to use it in an off-label fashion to inhibit angiogenesis associated with wet macular degeneration. [11,12]

In 2006, however, Genetech® gained Food and Drug Administration (FDA) approval for the anti-VEGF drug ranibizumab, a drug derived from the same bevacizumab molecule, as a fragment but with a smaller molecular size to theoretically aid retinal penetration. [13] Landmark clinical trials established that ranibizumab not only prevented vision loss but also led to a significant gain in vision in almost one-third of patients. [14,15] The ANCHOR trial, involving 423 patients,  compared ranibizumab dosed at 0.3 mg and 0.5 mg given monthly over two years with PDT and verteporfin given as required. This trial found 90% of ranibizumab treated patients achieved visual stabilisation with a loss of < 15 letters compared to 65.7% of PDT patients. Furthermore, up to 41% of the ranibizumab treated group actually gained >15 letters compared to 6.3% of the PDT group. [15]

Further trials including the MARINA, [14] PRONTO, [16] SUSTAIN, [17] and PIER [18] studies confirmed the effectiveness of ranibizumab. Despite these results and the purpose built nature of ranibizumab for the eye, in countries like the US and other countries around the world where patients and health insurance companies bear the cost burden of treatment, bevacizumab (Avastin®) is more frequently used, and constitutes nearly 60% of injections in the US. [19] This occurrence is explained by the large cost difference between ranibizumab (USD $1593) and bevacizumab (USD $42) in context of apparent similar efficacy. [19] The cost difference is due to the fact that one vial of bevacizumab can be fractioned by a compounding pharmacy into numerous unit doses for the eye. [20]

Given the popular off-label use of bevacizumab, the CATT trial was conducted by the US National Eye Institute to establish its efficacy. The CATT trial was a large US multicentre study involving 1208 patients randomised to receive either bevacizumab 1.25 mg or ranibizumab 0.5 mg (monthly or as needed). The CATT trial results demonstrated that monthly bevacizumab was equivalent to monthly ranibizumab (mean gain of 8.0 vs 8.5 letters on ETDRS visual acuity chart in one year). [21] The IVAN trial, a UK multi-centre randomised controlled trial (RCT) involving 628 patients, showed similar results to the CATT trial with a statistically insignificant mean difference in BCVA of 1.37 letters between the two drugs. [22]

Hence debate has mounted in regards to the substantial cost difference in the face of apparent efficacy. [23] On the backdrop of this costly dilemma are three major pharmaceutical companies: Genetech®, Roche®  and Novartis®  Although  bevacizumab  was  developed  in 2004 by the pharmaceutical company Genetech®, the company was taken over in 2009 by the Swiss pharmaceutical giant Roche®, which is one-third owned by another pharmaceutical company, Novartis®. [24] Given that both ranibizumab and bevacizamab are produced essentially by the same pharmaceutical companies (Genetech/Roche/ Novartis) there is no financial incentive for the company to seek FDA or Therapeutic Goods Administration (TGA) approval for the cheaper alternative, bevacizumab. [13,24]

Another  major  concern  that  is  emphasised  in  the  literature  is the potentially increased systemic adverse effects reported with bevacizumab. [22] The systemic half-life of bevacizumab is six days compared to ranibizumab at 0.5 days and in theory it is postulated that systemic inhibition of VEGF could cause higher systemic vascular events. [2] The CATT trial reported similar rates of adverse reactions (myocardial infarction, stroke and death) in both bevacizumab and ranibizumab groups. [21] However, a meta-analysis of the CATT and IVAN data showed that there was an increased risk of serious systemic side effects requiring hospitalisation in the bevacizumab group (24.9% vs 19.0%). Yet this statement is controversial as most events reported were not identified in the original cancer trials involving patients receiving intravenous doses of bevacizumab (500 times the intravitreal dose). [21,22] Hence it has been questioned whether this is more attributable to chance or imbalance in the baseline health status of participants. [2,22] An analysis of US Medicare claims demonstrated that  patients  treated  with  bevacizumab  had  significantly  higher stroke and mortality rates than ranibizumab. [25] However this data is inherently prone to confounding bias considering the elderly at risk of macular degeneration are likely to have risk factors for systemic vascular  disease.  When  corrected  for  comorbidities  there  were no  significant  differences  in  outcomes  between  ranibizumab  and bevacizumab. [23,25] It has been argued that trials to date have been underpowered to investigate adverse events in bevacizumab. Hence until further evidence is available, the risk of systemic adverse effects favouring the use of ranibizumab over bevacizumab is unclear. [22]

Adding to the debate regarding the optimum drug choice for AMD, is the newest anti-VEGF, aflibercept (Eylea®) which attained FDA approval in late 2011. Aflibercept was created by the pharmaceutical companies Regeneron/Bayer® and is a novel recombinant fusion protein designed to bind to all isoforms of VEGF-A, VEGF-B and placental growth factor. [20] Aflibercept has a dispensed price the same as ranibizumab at AUD $1430 per injection on the PBS. [26] The binding affinity of aflibercept to VEGF is greater than ranibizumab and bevacizumab which allows for longer duration of action and hence extended dosing intervals. [27]

The VIEW 1 study, a North American multicentre RCT with 1217 patients, and the VIEW 2 study, with 1240 patients enrolled across Europe, the Middle East, Asia-Pacific and Latin America, assigned patients into one of four groups: 1) 0.5 mg aflibercept given monthly, 2) 2 mg aflibercept given monthly, 3) 2 mg aflibercept at two-monthly intervals after an initial 2 mg aflibercept monthly for three months, or 4) ranibizumab 0.5 mg monthly. The VIEW 1 trial demonstrated that vision was maintained (defined as losing less than 15 ETDRS letters) in 96% of patients on 0.5 mg aflibercept monthly, 95% of patients receiving 2 mg monthly, 95% of patients on 2 mg every two months and 94% of patients on ranibizumab 0.5 mg monthly. [28] Safety profiles of the drugs in both the VIEW 1 and VIEW 2 trials showed no difference between aflibercept and ranibizumab in terms of severe systemic side effects. Hence aflibercept has been regarded as equivalent in efficacy to ranibizumab with potentially less frequent dosing.

Frequency of injections

In addition to the optimal drug of choice for AMD, the optimal frequency of injection has come into question. Given the treatment burden of regular intravitreal injections and risk of endophthalmitis with each injection, extending treatment using “as-required” dosing is often used in clinical practice. Evidence from the integrated analysis of VIEW trials is encouraging as it showed that aflibercept given every two months after an initial loading phase of monthly injections for three months was non-inferior to ranibizumab given monthly in stabilising visual outcomes [28] Although the cost is similar to ranibizumab, the reduced number of injections may represent significant cost savings.

A meta-analysis of the IVAN and CATT trials showed that continuous monthly treatment of ranibizumab and bevacizumab, gives better visual function than discontinuous treatment with a mean difference in BCVA at two years of -2.23 letters. [22] The pooled estimates of macular exudation as determined by optical coherence tomography (OCT) favoured a continuous monthly regimen. However, there was an increase in the risk of developing new geographic atrophy of the retinal pigment epithelium (RPE) with monthly treatment when compared to the as-needed therapy, therefore visual benefits from the monthly treatment may not be maintained long-term. [22] It is unclear whether the atrophy of the RPE represents a drug effect or the natural history of AMD. Interestingly, mortality at two years was lower with the continuous compared to the discontinuous group. In relation to systemic side effects, the pooled results slightly favoured continuous therapy although this was not statistically significant. This appears to contradict the normal dose response framework, however it is hypothesised that immunological sensitisation with intermittent dosing may account for this. [22]

Hence it appears that continuous therapy for bevacizumab and ranibizumab may be favourable in terms of visual outcome. However in clinical practice, given the treatment burden for patients and their carers, the risk of rare sight threatening endopthalmitis and possible sustained rise in intraocular pressure with each injection, [29] the frequency of injections is often individualised based on maintenance of visual acuity and anatomic parameters of macular thickness on OCT.

Currently the “inject and extend” model is recommended, whereby after three monthly injections treatment is extended to five or six weeks if the OCT shows no fluid. Depending on signs of exudation and BCVA, treatment may be reduced or extended by one or two weeks per visit to a maximum interval of ten weeks. Although there are no large prospective studies to support this, smaller studies have reported encouraging results which offers another cost saving strategy. [30] However, given the use of the more expensive ranibizumab, it is still a costly endeavour in Australia.

Current Australian situation

Other practical issues play a role in the choice of anti-VEGF therapy in Australia. For instance, the subsidised cost of ranibizumab to the patient is lower than the unsubsidised full cost of bevacizumab. [13] Patients must pay between AUD $80 and $159 out-of-pocket per injection for bevacizumab, whilst ranibizumab costs the government AUD $1430 and the maximum out of pocket cost for the patient is around AUD $36. [26] Among ophthalmologists there is favour towards the use of ranibizumab because of its purpose built status for the eye. [13] It seems the quantity and quality of evidence for ranibizumab compared to bevacizumab is greater. [29] As bevacizumab is used off- label, its use is not monitored, hence there is no surveillance. Lack of appropriate surveillance has been argued as a case to favour the use of the FDA approved ranibizumab. Essentially the dilemma faced by ophthalmologists is summarised in the statement: “I would personally be reluctant to say to my patients, ‘The best available evidence supports the use of this treatment which is funded, but are you interested in changing to an unapproved treatment [Avastin] for the sake of saving the community some money?” [31]

Another issue in Australia is the need for bevacizumab to be altered and divided by a compounding pharmacist into a product that is suitable and safe for ocular injection. A recent cluster of infectious endophthalmitis  resulting  in  vision  loss  occurred  in  the  US  from non-compliance to recognised standards. [32] The CATT and IVAN studies had stringent quality and safety control with the bevacizumab repackaged in glass vials using aseptic methods. In these trials, the risk of sight-threatening endophthalmitis was rare for both ranibizumab (0.04%) and bevacizumab injections (0.07%). [21] However, in clinical practice, it is argued that many of the compounding pharmacies may not be as regulated as that of the clinical trials to give comparable inferences about safety.

Conclusion

Prior to development of anti-VEGF therapies, patients with wet macular degeneration were faced with a progressive and permanent decline in vision. Today the available treatments not only stabilise vision but also lead to an improvement in vision in a significant portion of patients. Currently there are no published “head-to-head” trials comparing the three available drugs – bevacizumab, ranibizumab and aflibercept – together, which is warranted. In addition, further analyses of the safety concerns of bevacizumab are required. Current research is focusing on improving anti-VEGF protocols to reduce injection burden and combination therapies with photodynamic therapy or corticosteroids. [3] However, topical therapies such as pazopanib, a tyrosine kinase inhibitor that targets VEGF receptors, currently in the pipeline, may offer a possible non-invasive therapy in the future. [2,33]

At present, the evidence and expert opinion is not unanimous in allowing health policy makers to rationalise the substitution of bevacizumab over ranibizumab or aflibercept. Practical concerns in terms of FDA or TGA approval, surveillance, compounding pharmacy and safety are still major issues.  In 2013, ranibizumab was the third- highest costing drug on the PBS at AUD $286.9 million and aflibercept prescriptions cost the Australian government AUD $60.5 million per annum. [26] From a public health policy perspective, Australia has an ageing population and with eye health burden only to increase, there is need to prioritise resources. The cost–benefit analysis is not limited to AMD but applies to other indications of anti-VEGF therapy such as diabetic macular oedema and retinal vein occlusion. Substitution of first-line treatment with bevacizumab, which has occurred elsewhere in the world, has the potential to save the PBS billions of tax-payer dollars over a few years and its review should be considered a high priority in current health policy.

Acknowledgements

Thanks to Dr. Jack Tan (JCU adjunct lecturer (ophthalmology), MMED (OphthalSci)) for reviewing and editing this submission.

Conflict of interest

None declared.

Correspondence

M R Seneviratne: ridmee.seneviratne@my.jcu.edu.au

References

[1]  Wang  JJ,  Rochtchina  E,  Lee  AJ,  Chia  EM,  Smith  W,  Cumming  RG,  et  al.  Ten-year incidence and progression of age-related maculopathy: the blue Mountains Eye Study. Ophthalmology. 2007;114(1):92-8.

[2] Lim LS, Mitchell P, Seddon JM, Holz FG, Wong T. Age-related macular degeneration. Lancet. 2012;379(9827):1728-38.

[3] Cunnusamy K, Ufret-Vincenty R, Wang S. Next generation therapeutic solutions for age-related macular degeneration. Pharmaceutical patent analyst. 2012;1(2):193-206.

[4] Meleth AD, Wong WT, Chew EY. Treatment for atrophic macular degeneration. Current Opinion in Ophthalmology. 2011;22(3):190-3.

[5] Spilsbury K, Garrett KL, Shen WY, Constable IJ, Rakoczy PE. Overexpression of vascular endothelial growth factor (VEGF) in the retinal pigment epithelium leads to the development of choroidal neovascularization. The American Journal of Pathology. 2000;157(1):13544.

[6] Wong TY, Chakravarthy U, Klein R, Mitchell P, Zlateva G, Buggage R, et al. The natural history  and  prognosis  of  neovascular  age-related  macular  degeneration:  a  systematic review of the literature and meta-analysis. Ophthalmology. 2008;115(1):116-26.

[7]  Photocoagulation  Study  Group  .  Argon  laser  photocoagulation  for neovascular maculopathy.  Five-year  results  from  randomized  clinical trials.  Macular.  Archives  of Ophthalmology. 1991;109(8):1109-14.

[8] Kaiser PK. Verteporfin therapy of subfoveal choroidal neovascularization in age-related macular degeneration: 5-year results of two randomized clinical trials with an open-label extension: TAP report no. 8. Graefe’s archive for clinical and experimental ophthalmology. Albrecht   von   Graefes   Archiv   fur   klinische   und   experimentelle   Ophthalmologie. 2006;244(9):1132-42.

[9] Bressler NM. Photodynamic therapy of subfoveal choroidal neovascularization in age-related macular degeneration with verteporfin: two year results of 2 randomized clinical trials. Archives of Ophthalmology. 2001;119(2):198-207.

[10] Gragoudas ES, Adamis AP, Cunningham ET, Feinsod M, Guyer DR. Pegaptanib for neovascular age-related macular degeneration. The New England Journal of Medicine. 2004;351(27):2805-16.

[11] Madhusudhana KC, Hannan SR, Williams CP, Goverdhan SV, Rennie C, Lotery AJ, et al. Intravitreal bevacizumab (Avastin) for the treatment of choroidal neovascularization in  age-related  macular  degeneration: results  from  118  cases.  The  British  Journal  of Ophthalmology. 2007;91(12):1716-7.

[12] Rosenfeld PJ, Moshfeghi AA, Puliafito CA. Optical coherence tomography findings after  an  intravitreal  injection  of  bevacizumab (avastin)  for  neovascular  age-related macular degeneration. Ophthalmic surgery, lasers & imaging : The Official Journal of the International Society for Imaging in the Eye. 2005;36(4):331-5.

[13] Chen S. Lucentis vs Avastin: A local viewpoint, INSIGHT. 2011. Available from: http://www.visioneyeinstitute.com.au/wp-content/uploads/2013/05/Avastin-vs-Lucentis-Insight-article-Nov-2011.pdf.

[14] Rosenfeld PJ, Brown DM, Heier JS, Boyer DS, Kaiser PK, Chung CY, et al. Ranibizumab for neovascular age-related macular degeneration. The New England Journal of Medicine. 2006;355(14):1419-31.

[15] 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(1):57-65.

[16] Lalwani GA, Rosenfeld PJ, Fung AE, Dubovy SR, Michels S, Feuer W, et al. A variable-dosing  regimen  with  intravitreal  ranibizumab  for  neovascular  age-related  macular degeneration:  year  2  of  the  PrONTO  Study.  American  Journal  of  Ophthalmology. 2009;148(1):43-58.

[17] Holz FG, Amoaku W, Donate J, Guymer RH, Kellner U, Schlingemann RO, et al. Safety and efficacy of a flexible dosing regimen of ranibizumab in neovascular age-related macular degeneration: the SUSTAIN study. Ophthalmology. 2011;118(4):663-71.

[18]  Abraham  P,  Yue  H,  Wilson  L.  Randomized,  double-masked,  sham-controlled  trial of ranibizumab  for  neovascular age-related macular degeneration: PIER study year 2. American Journal of Ophthalmology. 2010;150(3):315-24.

[19] Brechner RJ, Rosenfeld PJ, Babish JD, Caplan S. Pharmacotherapy for neovascular age-related macular degeneration: an analysis of the 100% 2008 medicare fee-for-service part B claims file. American Journal of Ophthalmology. 2011;151(5):887-95.

[20] Ohr M, Kaiser PK. Aflibercept in wet age-related macular degeneration: a perspective review. Therapeutic Advances in Chronic Disease. 2012;3(4):153-61.

[21] Martin DF, Maguire MG, Ying GS, Grunwald JE, Fine SL, Jaffe GJ. Ranibizumab and bevacizumab for neovascular age-related macular degeneration. The New England Journal

[22] Chakravarthy U, Harding SP, Rogers CA, Downes SM, Lotery AJ, Culliford LA, et al.Alternative treatments to inhibit VEGF in age-related choroidal neovascularisation: 2-year findings of the IVAN randomised controlled trial. Lancet. 2013;382(9900):1258-67.

[23] Aujla JS. Replacing ranibizumab with bevacizumab on the Pharmaceutical Benefits Scheme: where does the current evidence leave us? Clinical & experimental optometry : Journal of the Australian Optometrical Association. 2012;95(5):538-40.

[24] Seccombe M. Australia’s Billion Dollar Blind Spot. The Global Mail. 2013 June 4:5.

[25] Curtis LH, Hammill BG, Schulman KA, Cousins SW. Risks of mortality, myocardial infarction,  bleeding,  and  stroke  associated  with  therapies  for  age-related  macular degeneration. Archives of Ophthalmology. 2010;128(10):1273-9.

[26]  PBS.  Expenditure  and  prescriptions  twelve  months  to  30  June  2013.  Canberra, Australia Pharmaceutical Policy Branch; 2012-2013 [4th of June 2014]. Available from: http://www.pbs.gov.au/statistics/2012-2013-files/expenditure-and-prescriptions-12-months-to-30-06-2013.pdf.

[27] Stewart MW, Rosenfeld PJ. Predicted biological activity of intravitreal VEGF Trap. The British Journal of Ophthalmology. 2008;92(5):667-8.

[28] Heier JS, Brown DM, Chong V, Korobelnik JF, Kaiser PK, Nguyen QD, et al. Intravitre-al aflibercept (VEGF trap-eye) in wet age-related macular degeneration. Ophthalmology. 2012;119(12):2537-48.

[29] Tseng JJ, Vance SK, Della Torre KE, Mendonca LS, Cooney MJ, Klancnik JM, et al. Sustained increased intraocular pressure related to intravitreal antivascular endothelial growth  factor  therapy  for  neovascular  age-related  macular  degeneration.  Journal  of Glaucoma. 2012;21(4):241-7.

[30] Engelbert M, Zweifel SA, Freund KB. Long-term follow-up for type 1 (subretinal pigment epithelium) neovascularization using a modified “treat and extend” dosing regimen of intravitreal antivascular endothelial growth factor therapy. Retina (Philadelphia, Pa). 2010;30(9):1368-75.

[31] McNamara S. Expensive AMD drug remains favourite. MJA InSight. 2011. Available from;                     https://www.mja.com.au/insight/2011/16/expensive-amd-drug-remains-favourite?0=ip_login_no_cache%3D22034246524ebb55d312462db14c89f0.

[32] Gonzalez S, Rosenfeld PJ, Stewart MW, Brown J, Murphy SP. Avastin doesn’t blind people, people blind people. American Journal of Ophthalmology. 2012;153(2):196-203.

[33] Danis R, McLaughlin MM, Tolentino M, Staurenghi G, Ye L, Xu CF, et al. Pazopanib eye drops: a randomised trial in neovascular age-related macular degeneration. The British Journal of Ophthalmology. 2014;98(2):172-8.