Introduction

Some of the most frequently occurring ocular diseases that cause certified visual loss are associated with pathological retinal neovascularisation and oedema. Of these diseases, wet (neovascular) age-related macular degeneration (AMD), diabetic retinopathy (proliferative diabetic retinopathy (PDR) and diabetic macular oedema (DMO)), and retinal vein occlusion (RVO) are of particular epidemiological importance as leading causes of blindness. In the United Kingdom, AMD is the leading cause of severe sight impairment (legal blindness) and partial sight certifications for all ages, accounting for over half of all visual impairment certifications.1 Diabetic retinopathy is the third most common cause of blindness and partial sight certifications for all age groups in the United Kingdom (6.3 and 7.6%, respectively);1 however, in people of working age (aged 16–64 years), diabetic retinopathy is the leading cause of severe sight impairment certification (17.7%).2 RVO is the second most common type of retinal vascular disease (after diabetic retinopathy) and includes branch RVO (BRVO) and central RVO (CRVO), accounting for <2% of all severe sight impairment and partial sight certifications in the United Kingdom.1

Vascular endothelial growth factor A (VEGF-A), a key regulator of angiogenesis and vascular permeability (reviewed by Ferrara et al3) has been implicated in the pathogenesis of retinal diseases associated with neovascularisation and oedema, including wet AMD,4, 5 diabetic retinopathy (particularly, PDR and DMO),6, 7 and RVO,6, 8, 9 as well as other ocular diseases such as retinopathy of prematurity.6 Although angiogenesis is the result of a highly complex molecular process involving several different receptors and ligands, VEGF-A seems to be a requirement for blood vessel growth in both normal and pathological angiogenesis. Consequently, several anti-VEGF agents have been developed for the treatment of these diseases.

Individualised treatment for retinal vascular diseases

In the years since the Human Genome Project10 was completed, huge progress has been made in unravelling the genetic basis of disease and understanding what drives diseases at a molecular level. By comparing patterns and frequencies of single-nucleotide polymorphisms (SNPs) in patients and controls, we have become able to identify which SNPs are associated with which diseases,11 helping drive the concept and clinical application of personalised medicine. Nowhere has this been better exemplified than in the field of oncology, where for example, only those patients who are most likely to respond are given a targeted treatment based on their genetic profile (eg, HER2/ErbB2 in breast cancer). Such is the potential for personalised medicine that the UK government's Technology Strategy Board has joined forces with Cancer Research UK and other bodies to fund the Stratified Medicines Programme,12 which they see as a significant step in making targeted therapies available for people with cancer in the United Kingdom. The benefit to patients of such an approach is clear. Individualised treatment allows the identification of patients who are most likely to benefit from the treatment. Tailoring treatment to the individual patient in this way should increase the chance of treatment success, while sparing patients from unnecessary drug exposure and risk of adverse events. Furthermore, avoiding unnecessary treatment also has the potential to improve the cost-effectiveness of treatment.

While treatment decisions for patients with retinal vascular diseases are not currently based on genetics, gene association work has already identified multiple genes that may be associated with AMD,13 and understanding how these are implicated in the pathogenesis of the disease opens up new research strategies based on specific pathways and molecules. In particular, dysregulation of the complement system has been shown to have a major part in the pathogenesis of wet and dry AMD,14 and a number of AMD-associated genetic loci have been identified.13 Numerous companies are currently developing genetically based and complement-targeted therapies with the goal of reducing complement-related AMD disease processes.14, 15

While in ophthalmology there is still some way to go before individualised treatment approaches that are based on genetics are available (as they are in certain cancers), it is already possible to begin to consider a similar patient-centred approach based on an individual's disease characteristics. It may be possible to use vision loss, visual acuity (VA) instability, or other signs of an active disease state as markers for requiring treatment, rather than using fixed dosing schedules. This type of approach should reduce the risks associated with over-treatment and under-treatment, thereby optimising the risk/benefit profile of the treatment and the efficient use of NHS resource. There is evidence to support this principle for the treatment of wet AMD, and visual impairment due to either DMO or macular oedema secondary to RVO with ranibizumab, as I will discuss later in this article.

Current management and recent therapeutic developments

Diagnosis

Wet AMD. Early diagnosis and treatment are vital for vision preservation in retinal vascular diseases, particularly wet AMD, because of the rapidly progressive nature of the disease. Patients with wet AMD typically present (to a general practitioner, optometrist, or local eye unit/eye casualty)16 with distortion, blurring, or loss of vision with a rapid onset. Some patients with unilateral wet AMD may be asymptomatic or report mild vision distortion and only be detected in a routine assessment.17 The Royal College of Ophthalmologists (RCOphth) recommends that suspected cases of wet AMD should be referred directly to the nearest AMD centre, eye casualty, or eye clinic, due to the aggressive nature of the disease within 1 week of initial presentation, with no more than 1 week between evaluation and treatment.

DMO. DMO can arise as early as the mild non-proliferative or as late as in the severe proliferative stages of diabetic retinopathy.18 DMO was defined by the Early Treatment of Diabetic Retinopathy Study (ETDRS) group as being clinically significant macular oedema when there is retinal thickening and/or hard exudates within 500 μm of the fovea or when there is a zone of oedema of at least 1 disc diameter in width and part of which is within 1 disc diameter from the fovea.19 DMO appears as retinal thickening on binocular stereoscopic slit-lamp examination and can be confirmed with retinal imaging techniques such as optical coherence tomography (OCT).20 Macular oedema may also be present in wet AMD and RVO, as well as other ocular diseases; however, the natural history of DMO distinguishes it from the other instances of macular oedema. In an attempt to reduce diabetes-related visual impairment in England, the English National Screening Programme for Diabetic Retinopathy (ENSPDR) was set up and provides annual photographic screening for every diabetic patient (over the age of 12 years) in England. All patients identified by screening as having sight-threatening diabetic retinopathy are referred to ophthalmology clinics.21, 22

RVO. Patients with RVO (including BRVO and CRVO) typically present with painless loss of vision.23, 24 BRVO (located in one of the branches of the central vein) is more common than CRVO (located in the central vein and affecting most of the retina) and usually occurs at sites where arterioles cross over veins.25, 26 Retinal imaging with fluorescein angiography is crucial for diagnosis and prognosis, allowing the identification of the specific type of RVO (eg, perfused vs non-perfused and BRVO vs CRVO), the identification of macular oedema (if present), its extent, persistence, regression, and degree of ischaemia.27 OCT provides additional information such as quantitative and qualitative assessment of retinal thickness and the exact location of the accumulated fluid (within the retinal layer vs the subretinal space).27 Clinical features that may be apparent at presentation—such as haemorrhage, cotton wool spots, and macular oedema27—overlap with those of other retinal vascular diseases, including diabetic retinopathy, hypertensive retinopathy, and retinopathy related to blood dyscrasias; thus, differential diagnosis is important.28

Management

Wet AMD. Treatment modalities for wet AMD have improved dramatically over the past decade, prior to which laser photocoagulation was the only available treatment option.29 Photodynamic therapy with verteporfin was licensed for wet AMD with predominantly classic subfoveal choroidal neovascularisation (CNV) in 2000,30 and recommended by the National Institute for Health and Clinical Excellence (NICE) in 2003 for patients with a confirmed diagnosis of classic subfoveal CNV with no sign of occult lesions.31 However, it was only with the emergence of anti-VEGFs that an effective treatment became available for patients with wet AMD, regardless of lesion type. Licensed anti-VEGFs for wet AMD include pegaptanib (Macugen®), which was authorised in the EU in 2006,32 and ranibizumab (Lucentis®▾), authorised in 2007.33 Ranibizumab is now considered the standard of care for wet AMD. This is because, in addition to demonstrating efficacy in preventing visual loss in large, randomised, controlled, clinical trials in the majority of patients,34, 35, 36, 37 ranibizumab has also been shown, on average, to provide significant gains in VA.34, 35, 36

The approved posology of ranibizumab for wet AMD consists of intravitreal injection (0.5 mg) given monthly and continued until maximum VA is achieved (defined as stable VA for 3 consecutive monthly assessments while on treatment). Subsequently, patients are monitored monthly for VA, and upon detection of reduced VA due to wet AMD, treatment is resumed until stable VA is reached.38 Notably, the current approved posology for ranibizumab represents an evolution in the treatment dosing paradigm for wet AMD, towards an individualised treatment approach, whereby injections are only administered at times of VA instability, during which patients are most likely to benefit; hence, minimising the chances of both under-treating or over-treating.

The initial marketing authorisation for ranibizumab was based on the pivotal trials MARINA36 and ANCHOR,34 which investigated a monthly dosing regimen, and a study which investigated a quarterly regimen, PIER.37 Almost all of the patients receiving monthly 0.5-mg ranibizumab injections in MARINA and ANCHOR maintained their VA at 1 year (94.6 and 96.4%, vs 62.2 and 64.3% of controls, respectively; P<0.001 for both), at least a third of patients (33.8% MARINA, 40.3% ANCHOR) gained 15 or more letters of VA (vs 5.0 and 5.6% of controls, respectively; P<0.001 for both comparisons) and, on average in both trials, there was an improvement in VA (+7.2 letters and +11.3 letters in MARINA and ANCHOR, respectively; Figure 1).34, 37 These outcomes were also maintained to 24 months.35, 36 However, in the PIER study, on average, patients who received fixed quarterly injections of 0.5 mg ranibizumab after a loading phase of 3 monthly injections, did not maintain the initial gain in VA seen at 3 months (mean of +4.3 letters vs baseline) at the 12-month time point (mean change from baseline −0.2 letters), although the difference compared with sham remained statistically significant at year 1 (Figure 2).37 Although, on average, quarterly injections were not frequent enough to maintain the initial gains in VA, an exploratory analysis of the ranibizumab group in the PIER study showed that patients could be stratified depending on their initial increase in VA and whether they were able to maintain this initial gain. Of the 40 patients (66%) who showed an initial increase in VA, 16 patients (40%) were ‘sustained responders’ (ie, had initial VA gains during the 3 monthly ranibizumab loading phase injections that were sustained to 1 year with subsequent quarterly injections), whereas the remainder of the patients showed a gradual decline in VA from month 4 (with or without initial VA gain), and may have benefitted if treated more frequently (Figure 3).39 This subanalysis showed that different patients needed different frequency of treatment to maintain initial gain.39 Therefore, the originally approved posology for ranibizumab included a loading phase of 3 monthly doses followed by an individualised pro re nata (PRN) maintenance phase with a VA-based retreatment criteria (treatment was resumed upon a VA loss of >5 letters), because on average, VA appeared to plateau after 3 consecutive monthly injections in MARINA, ANCHOR and, PIER.40

Figure 1
figure 1

Mean changes from baseline in VA with monthly 0.5 mg ranibizumab in patients with wet AMD with minimally classic/occult CNV, in the MARINA study. Only data from the licensed dose of ranibizumab (0.5 mg) are shown. P-value shown is for difference in mean change in VA vs sham at 24 months. ETDRS, Early Treatment Diabetic Retinopathy Study. Adapted from Rosenfeld et al.36

Full size image
Figure 2
figure 2

Mean changes from baseline in VA in patients with wet AMD with quarterly 0.5 mg ranibizumab following a loading phase of 3 monthly injections, in the PIER study. Only data from the licensed dose of ranibizumab (0.5 mg) are shown. Arrows indicate injection time point. P-value shown is for difference in mean change VA vs sham at 12 months. ETDRS, Early Treatment Diabetic Retinopathy Study. Adapted from Regillo et al,37 Randomized, double-masked, sham-controlled trial of ranibizumab for neovascular age-related macular degeneration: PIER study year 1. Am J Ophthalmol 2008; 145: 239–248, with permission from Elsevier © 2008.

Full size image
Figure 3
figure 3

Mean changes from baseline in VA in wet AMD in three subgroups of patients receiving 0.5 mg ranibizumab in the PIER study. Adapted from Holz et al,39 The effects of a flexible visual acuity-driven ranibizumab treatment regimen in age-related macular degeneration: outcomes of a drug and disease model. Invest Ophthalmol Vis Sci 2010; 51(1): 405–412, with permission from Association for Research in Vision and Ophthalmology © 2010.

Full size image

Since these initial trials, further evidence has emerged which suggested that an individualised patient-centred approach would be more suitable. In a retrospective analysis of data from studies evaluating monthly ranibizumab injections including the MARINA and ANCHOR trials as well as from the ranibizumab monotherapy arm in the DENALI study (a Phase IIIb study conducted in the United States and Canada),41 the time course of vision stability was evaluated.42 This retrospective analysis demonstrated that <20% of wet AMD patients reached VA stability by month 3; however, nearly 80% of patients reached stability within the first year of ranibizumab treatment. Furthermore, the analysis showed that once VA stability is achieved, the incremental VA benefit of continued monthly injections was minimal (the mean absolute change between the visit in which visual stability was achieved and the subsequent visit was ⩽0.3 letters in the MARINA, ANCHOR, and DENALI studies).42 The impact of treatment interruptions in ranibizumab monotherapy arms in trials that used PRN or quarterly dosing regimens including EXCITE,43 SUSTAIN,44 and MONT BLANC was also explored.45 The analysis demonstrated that once VA stability is achieved, the majority (60–80%) of patients have stable VA 2 months after last injection.42 Furthermore, the analysis demonstrated a consistent trend for better VA outcomes when treatment is resumed when VA is unstable (mean change after treatment re-initiation following a visit in which unstable VA was identified was +2.7, +4.3, and +3.0 letters in the EXCITE, SUSTAIN, and MONT BLANC studies, respectively).42

On the basis of these data, a revised ranibizumab wet AMD posology was approved by the EMA in September 2011.40 Ranibizumab can now be administered using a three-step individualised PRN treatment regimen: (1) treatment is initiated by monthly injections until maximum VA is achieved (defined as stable VA for 3 consecutive assessments while on ranibizumab); (2) treatment is then interrupted and patients are monitored monthly; (3) treatment is resumed if monitoring indicates loss of VA associated with active wet AMD and is continued monthly until VA is stable. This approach allows patients the opportunity to receive enough injections to gain maximum VA (at treatment initiation), prevents unnecessary injections (by treatment interruption once maximum VA has been achieved), and allows patients to receive injections when they are most likely to benefit from them (at treatment re-initiation). With the original posology, there was no opportunity for further injections unless >5 letter VA loss was observed. Waiting for a >5 letter VA drop may allow vision to be lost that cannot be regained.

Bevacizumab, which is unlicensed for any ocular use, is also sometimes used for treating wet AMD in clinical practice. Bevacizumab is a humanised full-length anti-VEGF monoclonal antibody that differs from ranibizumab in a number of its properties, including molecular structure, size and design, systemic pharmacokinetics, and formulation. It is licensed only for intravenous administration for the treatment of colorectal cancer and certain cases of breast, renal, lung, ovarian, fallopian tube, and peritoneal cancer.46 Some differences between the ocular and systemic safety profiles of bevacizumab and ranibizumab have been suggested, but despite the widespread use of bevacizumab for wet AMD in clinical practice, evidence from large-scale randomised trials were lacking until recently.

The head-to-head study of ranibizumab and bevacizumab, CATT, is a 2-year, randomised, prospective single-blind, non-inferiority trial to evaluate the comparative safety and efficacy of the two agents. In the 1-year primary end point analysis, while the non-inferiority limit was met for monthly bevacizumab compared with monthly ranibizumab, the non-inferiority limit (5 letters difference) was not met for PRN bevacizumab (n=300) compared with either monthly bevacizumab (n=286) or monthly ranibizumab (n=298). From an anatomical perspective, 4 weeks after their first injection, no fluid was seen on the OCT in 17.3% of bevacizumab-treated patients, while 27.5% of ranibizumab-treated patients were dry on OCT (P<0.001).47 The mean number of injections required in the bevacizumab PRN arm was significantly higher than the number of PRN ranibizumab injections required (7.7±3.5 and 6.9±3.0, respectively; P<0.003).47 Consistent with the move towards individualised treatment with ranibizumab, PRN ranibizumab was non-inferior to monthly ranibizumab dosing.47 Importantly, the observed mean VA gain from baseline to 1 year in the ranibizumab PRN arm of the CATT study (6.8 letters)47 represents the best outcome in randomised controlled trials for 1-year VA gain, with less than monthly ranibizumab dosing in comparison with the relevant arms of PIER (mean gain of 4.3 letters in the 0.5-mg quarterly group),37 EXCITE (mean gain of 3.8 letters in the 0.5-mg quarterly group),43 SUSTAIN (mean gain of 3.6 letters using a VA/OCT-based PRN regimen),44 and SAILOR (a phase IIIb study in which cohort 1 evaluated 0.3 and 0.5 mg ranibizumab using VA/OCT-based PRN regimen; mean gain of 2.3 letters in the 0.5-mg group).48 Thus, the findings of the CATT study supports the currently approved posology for ranibizumab, by suggesting that a more refined individualised regimen may optimise clinical outcomes while reducing the number of injections.

While not powered to identify rare but serious adverse events, safety differences were also identified in the CATT study, despite the population of the study being relatively fit due to patients being excluded from entering the study if they had significant concomitant medical conditions. No significant differences were observed between ranibizumab and bevacizumab in rates of death (1.3 and 1.4% for the ranibizumab and bevacizumab monthly groups, respectively; 1.7 and 3.7% for the respective PRN groups; P=0.18 for comparing all groups, P=0.22 for between drug comparison), nonfatal myocardial infarction (0.7% for both the ranibizumab and bevacizumab monthly groups; 1.0 and 0.3% for the respective PRN groups; P=0.78 for comparing all groups, P=0.73 for between drug comparison), or nonfatal stroke (1.0 and 0.7% for the ranibizumab and bevacizumab monthly groups, respectively; 0.3 and 0.7% for the respective PRN groups; P=0.88 for comparing all groups, P=1.0 for between drug comparison).47 However, comparing rates of serious systemic adverse events associated with hospitalisation between ranibizumab- and bevacizumab-treated patients (combining dosing-regimen groups) demonstrated a statistically significant higher rate with bevacizumab (24.1 (141) vs 19.0% (114); RR 1.29; 95% CI 1.01–1.66; P=0.04).47 Intravitreal bevacizumab also led to a significantly higher incidence of gastrointestinal disorders (including haemorrhage) compared with ranibizumab (2.6 vs 0.8%; P=0.02). These observed adverse events are consistent with those noted in the Summary of Product Characteristics for bevacizumab and are known to be potential risks related to systemic exposure of anti-VEGFs.46

DMO. The standard of care in DMO consisted of focal/grid laser photocoagulation (laser treatment) since the trial of the ETDRS in 1985, which demonstrated that this treatment substantially reduced the risk of visual loss.49 Some recent trials showed that laser treatment may also improve vision. For example, the Diabetic Retinopathy Clinical Research Network (DRCR.net) randomised study compared focal/grid photocoagulation to intravitreal administration of the corticosteroid triamcinolone acetonide (IVTA). The study showed that from baseline to year 3, laser treatment was associated with a mean gain of 5 letters and with an improved VA by ⩾10 letters in 44% of patients (vs worsening VA by ⩾10 letters in 12% of patients).50

Pharmacotherapeutic options in DMO were very limited until recently. Unlicensed and contraindicated use of IVTA is widespread and its rationale is based on the anti-angiogenic properties of corticosteroids (possibly due to downregulation of VEGF).51, 52 Notably, in the IVTA vs laser DRCR.net study, at 4 months, VA in the 4-mg IVTA group was superior to that in the 1-mg IVTA group (mean difference between the groups adjusted for baseline VA and prior macular photocoagulation, 3.6 letters; P=0.001) and in the laser-treated group (mean difference between the groups adjusted for baseline VA and prior macular photocoagulation, 3.8 letters; P<0.001); however, the VA differences between the groups disappeared by the end of year 1. By year 3, laser treatment conferred better VA outcomes compared with IVTA treatment (mean difference between laser and 1-mg IVTA groups adjusted for baseline VA and prior macular photocoagulation, 5.6 letters; 95% CI, 0.8–10.4; respective difference between laser and 4-mg IVTA groups, 4.7 letters; 95% CI, 0.0–9.5).50, 53 Furthermore, IVTA may increase the risk for secondary glaucoma and secondary cataracts.50

Ranibizumab was authorised in the EU for the treatment of visual impairment due to DMO in 2011,40 thereby significantly expanding the treatment armamentarium for DMO. Phase II studies such as READ-2 and RESOLVE demonstrated efficacy and tolerability of ranibizumab in DMO.54, 55 Phase III studies of ranibizumab in DMO further supported its utility in this indication. The independent DRCR.net Protocol I study was a 4-arm trial (854 eyes) evaluating ranibizumab plus prompt laser to ranibizumab plus deferred laser (⩾24 weeks), IVTA plus prompt laser, and sham injection plus prompt laser.56 Ranibizumab (or sham) injections were administered monthly for the first 3 months (totalling four injections) followed by PRN dosing based on VA/OCT criteria.56 The Protocol I study showed that at 1 year, ranibizumab plus prompt or deferred laser was superior to sham plus laser (mean increase of 9 letters for both ranibizumab groups vs 3 letters for the sham plus laser group; P<0.001 for comparisons vs sham), whereas IVTA plus laser was comparable to sham plus laser (mean increase of 4 letters for the IVTA plus laser group; P=0.31 vs sham plus laser).56 Two-year VA outcomes were similar to 1-year outcomes (Figure 4).56 The RESTORE phase III study (N=345) compared ranibizumab monotherapy (plus sham laser) to ranibizumab plus laser to laser alone (plus sham injections).57 Ranibizumab (0.5 mg) regimen included a loading phase of 3 monthly injections followed by PRN dosing (based on VA stability criteria).57 The mean number of ranibizumab injections in the ranibizumab and ranibizumab/laser groups was 7.0±2.8 and 6.8±3.0, respectively.57 The mean average change in best corrected visual acuity (BCVA) letter score from baseline to month 1 through month 12 was significantly superior with ranibizumab and ranibizumab plus laser vs laser alone (6.1, 5.9, and 0.8 letters, respectively; P<0.0001 for both comparisons vs laser alone); the difference between the two ranibizumab groups was not significant (P=0.61).57 The safety profile of ranibizumab in DMO trials was consistent with that of ranibizumab in wet AMD.56, 57 In addition to the DRCR.net and RESTORE studies, data from two other Phase III studies (RISE, RIDE) have strengthened the evidence for ranibizumab in DMO.58, 59

Figure 4
figure 4

Mean changes from baseline in VA in patients with DMO in the DRCR.net Protocol I study. No equivalent to the unpreserved triamcinolone formulation used in this study is currently available in the United Kingdom. P-values shown are for difference in mean change in VA vs sham plus prompt laser at 52 weeks. Adapted from Elman et al,56 Randomized trial evaluating ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema. Ophthalmology 2010; 117(6): 1064–1077, e1035, with permission from Elsevier © 2010.

Full size image

RVO. The standard of care for macular oedema secondary to BRVO consisted of grid laser photocoagulation since the Branch Vein Occlusion Study demonstrated the efficacy of this treatment approach (over no treatment).60 In contrast, as the Central Vein Occlusion Study showed no VA benefit for grid laser photocoagulation (over no treatment) in patients with macular oedema secondary to CRVO at any follow-up point,61 the standard of care for CRVO was observation until the recent development of pharmacotherapeutic options.

The rationale for using corticosteroids for macular oedema secondary to RVO is based on their anti-angiogenic and anti-inflammatory properties.51, 52, 62 The SCORE-BRVO study—a randomised trial (N=411) comparing IVTA (1 and 4 mg) to standard-of-care (grid laser photocoagulation that may be prompt or deferred depending on the absence/presence of dense macular haemorrhage) in patients with macular oedema secondary to BRVO—failed to demonstrate an advantage for IVTA over prompt/deferred laser.63 In this study, the mean letter gain from baseline to 1 year was not statistically different across the treatment groups (4.2, 5.7, and 4.0 letters for the standard care, 1-mg IVTA, and 4-mg IVTA groups, respectively; P=0.70), and neither was the difference in the proportion of patients who gained ⩾15 letters from baseline to 1 year (28.9, 25.6, and 27.2% for the standard care, 1-mg IVTA, and 4-mg IVTA groups, respectively; P=0.89 for all comparisons).63 Rates of elevated intraocular pressure (IOP) and cataracts were comparable in the standard-of-care and 1-mg IVTA groups and higher in the 4-mg IVTA group. There was, however, a dose-dependent higher frequency initiation of IOP-lowering medications and an increase in lens capacity onset/progression in the IVTA groups compared with the standard care group.63 In contrast to the SCORE-BRVO study, the complementary SCORE-CRVO trial—a randomised trial (N=271) comparing IVTA (1 and 4 mg) to observation in patients with macular oedema secondary to CRVO—demonstrated a superiority for IVTA over observations in these patients.64 In SCORE-CRVO study, from baseline to 1 year, IVTA-treated patients lost, on average, fewer letters compared with the observation group (loss of 12.1, 1.2, and 1.2 letters for the observation, 1-mg IVTA, and 4-mg IVTA groups, respectively; P=0.004); similarly, a higher proportion of patients in the IVTA groups gained ⩾15 letters from baseline to 1 year (6.8, 26.5, and 25.6% for the observation, 1-mg IVTA, and 4-mg IVTA groups, respectively; P=0.001 for both comparisons vs observation).64 Similar to the SCORE-CRVO findings, rates of IOP and cataracts were comparable in the observation and 1-mg IVTA groups and higher in the 4-mg IVTA group, and there was a dose-dependent higher frequency initiation of IOP-lowering medications in the IVTA groups compared with the observation group.64

Another corticosteroid that has been investigated and is now authorised for the treatment of RVO is dexamethasone intravitreal implant (DEX implant; Ozurdex®). The GENEVA studies (N=1267) were randomised trials comparing DEX implant (0.35 or 0.7 mg) with sham treatment in patients with macular oedema secondary to BRVO or CRVO at 6 months, followed by an open-label 6-month extension phase in which patients could receive a second DEX implant (0.7 mg) based on BCVA and retinal thickness.65, 66 The studies demonstrated that from baseline to 6 months, mean VA improvement was better in the DEX groups than in the sham group (P⩽0.006); the greatest between-group difference was at day 60 (∼7 letters).65 From day 30 to 90 (but not later), the proportion of patients with ⩾15 letter gain was significantly greater in the DEX groups (P<0.001); at day 60, 29% of patients in both DEX groups gained ⩾15 letters compared with 11% of the sham group (P<0.001),65 and this was maintained at 12 months in patients who received two 0.7-mg DEX implants (30 and 32%, 60 days after the first and second implant, respectively).66 Furthermore, patients in the DEX groups achieved the 15-letter gain faster than those in the sham group (P<0.001 vs sham).65 Rates of elevated IOP were overall higher in the DEX groups than the sham group (P⩽0.002) and the percentage of eyes receiving IOP-lowering medication increased in the DEX implant treatment groups from ∼6% at the beginning of the study to ∼24% by day 180, whereas there was no change in the sham group. By day 180, there was no difference in the rates of elevated IOP between the DEX groups and sham.65 Rates of cataracts were not significantly different between the DEX (7.3% in 0.7-mg group, 4.1% in 0.35-mg group) and sham (4.5%) groups at 6 months.65 However, at 12 months, patients who received two 0.7-mg DEX implants had a higher rate of cataract progression compared with sham (29.8% of phakic eyes vs 5.7% of phakic eyes, respectively).66

Ranibizumab was authorised in the EU for the treatment of visual impairment due to macular oedema secondary to RVO in 2011,40 thereby expanding the treatment options for RVO using an anti-VEGF approach. Pilot studies of ranibizumab in RVO provided preliminary proof for the potential utility of ranibizumab in this disease67, 68, 69 and provided the rationale for the two Phase III 12-month studies, BRAVO (in patients with macular oedema secondary to BRVO; N=397)25, 70 and CRUISE (in patients with macular oedema secondary to CRVO; N=392).71, 72 In both studies participants were randomised to receive monthly intravitreal ranibizumab (0.3 or 0.5 mg) or sham injections from day 0 to month 5 (thereafter, all patients with study eye BCVA of ⩽20/40 or central retinal thickness of ⩾250 μm received ranibizumab PRN).25, 70, 71, 72 In the BRAVO study, patients could receive rescue laser treatment once during the treatment period and once during the observation period if criteria were met.70 In both studies, the mean letter gain from baseline to month 6 was superior with ranibizumab compared with sham group. In BRAVO, patients gained 18.3 and 7.3 letters in the 0.5-mg group and sham group, respectively (P<0.0001 for ranibizumab vs sham),25 and in CRUISE, the respective values were 14.9 and 0.8 letters (P<0.0001 for ranibizumab vs sham).72 The treatment benefits were maintained through month 12 on a PRN regimen (Figure 5). Also, in both studies, the proportion of patients with ⩾15 letter gain from baseline was significantly higher in the ranibizumab groups (vs sham) at 6 months (BRAVO: 61.1 and 28.8% of patients in the 0.5 mg and sham groups; P<0.0001 for ranibizumab vs sham;25 CRUISE: 47.7 and 16.9% for the respective groups; P<0.0001).72 These proportions were maintained in both studies from month 6 through 12 when ranibizumab was given PRN. A total number of 2.7 (BRAVO) and 3.3 (CRUISE) 0.5-mg ranibizumab injections were needed to maintain patients visual stability from months 6 to 12.70, 71 Safety profile of ranibizumab in these trials was consistent with that found in other studies.70, 71

Figure 5
figure 5

Mean changes from baseline in VA in patients with macular oedema secondary to RVO. Mean changes from baseline in VA in patients with (a) BRVO in the BRAVO study (adapted from Brown et al,70 Sustained benefits from ranibizumab for macular edema following branch retinal vein occlusion: 12-month outcomes of a phase III study. Ophthalmology 2011; 118(8): 1594–1602, with permission from Elsevier © 2011) and (b) CRVO in the CRUISE study (adapted from Campochiaro et al,71 Sustained benefits from ranibizumab for macular edema following central retinal vein occlusion: twelve-month outcomes of a phase III study. Ophthalmology 2011; 118(10): 2041–2049, with permission from Elsevier © 2011). Only data from the licensed dose of ranibizumab (0.5 mg) are shown. †Sham patients received 0.5 mg ranibizumab PRN treatment from months 6 to 11. P-values shown are for difference in mean change VA vs sham/ranibizumab. ETDRS, Early Treatment Diabetic Retinopathy Study.

Full size image

The viability of a stability-based individualised approach for ranibizumab in RVO was evaluated in a retrospective analysis conducted using data from the ranibizumab arms (pooled doses) in the Phase III BRAVO and CRUISE trials (in which patients received monthly injections from day 0 to month 5 and were dosed PRN thereafter (months 6–11)). The analysis demonstrated that 59 and 53% of ranibizumab-treated patients in the BRAVO and CRUISE trials, respectively, reached VA stability (⩽3 letters in 3 consecutive monthly visits with treatment at the first two visits) up to month 6.73 The mean VA change 1 month after ranibizumab treatment at a VA stability visit was small (BRAVO, 0.8 letters; CRUISE, 1.5 letters). During the PRN period, 1 month after ranibizumab re-initiation, mean VA gains were clinically relevant (BRAVO, 7.1 letters; CRUISE, 9.3 letters).73 Thus, this retrospective analysis is consistent with that conducted for the wet AMD trials and supports the currently approved posology.

Notably, the approved posology for all the ranibizumab indications is similar (the exception is that for visual impairment due to DMO or macula oedema secondary to RVO, further treatment is not recommended if there is no response after the first three injections; this is not the case for wet AMD).38 This harmonisation facilitates patient flow as it allows all patients with wet AMD, DMO, or macular oedema secondary to RVO to be treated using the same approach.

The impact of anti-VEGF treatments for DMO and RVO on clinic capacity

Following the introduction of anti-VEGF treatment for wet AMD, there was a consequent rise in the number of wet AMD patients, potentially suitable for treatment as well as the number (frequency) of follow-up appointments. The associated increase in clinical workload has been substantial and there is concern that the introduction of anti-VEGF treatments for DMO and RVO could further exacerbate pressure on clinic capacity in the hospital eye service.

The prevalence of diabetes (particularly type II) is increasing in the Western world with the ageing population. In the United Kingdom (in 2009), it was estimated that 2.6 million people were diagnosed with diabetes and that ∼500 000 more people had undiagnosed diabetes; the prevalence of diabetes in the adult population ranged from 3.9 (Scotland) to 5.1% (England).74 In an epidemiologic study of patients with type I diabetes (in the United States), the 14-year rates of progression to proliferative retinopathy and incidence of macular oedema were 37 and 26%, respectively.75 In a cohort study evaluating patients with type II diabetes in the United Kingdom, the cumulative 5-year incidence of developing sight-threatening diabetic retinopathy in patients without retinopathy at baseline was 3.9%.76 In contrast to diabetes and its associated retinal morbidities, RVO is relatively infrequent. In an epidemiologic study in the United States, the prevalence of BRVO was 0.6% and the prevalence of CRVO was 0.1%.77 In a recent analysis of pooled data from population studies worldwide, the overall RVO prevalence was 0.52% (0.44% BRVO, 0.08% CRVO), translating to ∼16 million individuals worldwide affected by RVO.78

The authorisation of anti-VEGF therapy for DMO and macular oedema secondary to RVO represented an increase in the number of patients eligible for anti-VEGF therapy, an important factor with respect to clinic capacity pressures. The RCOphth issued a preferred practice guidelines addressing the diabetic retinopathy screening and ophthalmology clinic set up in England as well as a guidance for management of RVO addressing patient pathways.20, 21 Ideally, referrals of patients with diabetic retinopathy should come through ENSPDR; whereas referrals of patients with RVO are likely to come from an optometrist, general practitioner, or other health workers.20, 21 In both cases, referral pathways should be streamlined to ensure timely treatments. Intravitreal injection facilities may be integrated into the patient pathways (similar to the way laser clinics are integrated in the guidance for diabetic retinopathy patients), so that patients could benefit from effective monitoring and care.21

Intravitreal injection facilities may be shared with wet AMD services. Importantly, the unified ranibizumab posology for wet AMD, DMO, and macular oedema secondary to RVO should streamline treatment of all these indications, thereby maximising clinic capacity. Conversely, utilising different treatments for each of these diseases may impact negatively on the capacity burden in specialist retinal clinics.

Summary

Treatment for retinal vascular diseases including wet AMD, DMO, and RVO has improved dramatically in recent years, due primarily, to the development and authorisation of anti-VEGF therapy. Treatment regimens have evolved through experience gained in clinical trials and clinical practice. The current treatment regimen for ranibizumab across these indications reflects an individualised treatment approach designed to treat patients when they could benefit the most while minimising the number of unnecessary intravitreal injections, and hence the risk of adverse events. To maximise patient outcomes, care should be taken to integrate intravitreal injection facilities into the patient pathways for DMO and RVO (similar to the approach taken for wet AMD); the new unified posology for ranibizumab may streamline the treatment of these patients, thereby minimising impact on clinic capacity and optimising patient outcomes.