Diabetic macular oedema (DMO) is an increasingly common visionthreatening disease that results from retinal vascular dysfunction and low-grade inflammation, developing into diabetic retinopathy (DR) and then to DMO over 10 or more years following the onset of diabetes.1 Areas of retinal tissue lose capillary vasculature and become ischaemic, stimulating secretion of vascular endothelial growth factor (VEGF) and other cytokines.2–4 Changes in paracellular and transcellular transport across the capillary endothelium and altered hydrostatic and osmotic pressure gradients result in fluid movement into retinal tissues, leading to consequent oedema and retinal damage.
Effective prevention and management of DR and DMO require intensive treatment of diabetes in terms of controlling glycaemia, blood pressure and lipid levels.5 Hyperglycaemia drives vascular dysfunction and DR, and it is important that patients understand the critical importance of controlling blood glucose levels. Treatment of DMO during the past decades has been almost entirely limited to laser photocoagulation but, recently, anti-VEGF agents have emerged as first-line agents in a subset of patients.6 Among these, two medications have been approved for this indication: aflibercept (EYLEA® ▼ ) and ranibizumab, (Lucentis®) and in large clinical trials, these have shown greater efficacy in central DMO than laser treatment or placebo, respectively.7–9
This article reports the proceedings of a symposium that reviewed the pathophysiology of DMO and the possible approaches to its management, particularly laser, anti-VEGF agents and corticosteroids. It also discussed the results of two large ongoing phase III clinical trials that are evaluating treatment of a large population of DMO patients with the anti-VEGF agent, aflibercept. These novel trials involve two regimens of aflibercept, in a head-to-head comparison with laser therapy. They are providing much-needed data on the comparativeefficacy of these treatments in DMO in terms of visual acuity (VA) and retinal pathology and are also providing useful data on their relative safety and tolerability.
Note from Publisher: This article was temporarily removed from the website for a short period of time. This was not related to the content of the article in any way, but was a technical requirement following a recent update to the prescribing information for aflibercept. The online article carried a version of the prescribing information which was no longer valid, owing to addition of a new licensed indication for aflibercept.
Professor Reinier Schlingemann
University of Amsterdam, The Netherlands
DMO is an increasing threat to vision worldwide and is the leading cause of blindness in young adults in developed countries.10 This rising prevalence is driven by the burgeoning numbers of people with type I and especially, type II diabetes. In Europe in 2013 it was estimated that 56 million people had diabetes, and this is expected to rise to 69 million by 2035 (22 % increase).11 Globally, 382 million people were estimated to have diabetes in 2013 with a projected rise to 592 million by 2035 (55 % increase).11 Among people with diabetes, approximately one-third have DR and approximately one-third of those have DMO. Therefore, 6.2 million people (11 % of people with diabetes) in Europe currently have DMO and 0.6–1.7 million have clinically significant MO.11,12 Given this burden, effective treatments for DMO are a critical need worldwide.
Main risk factors for the development of DR and DMO include: poorly controlled diabetes, chronic hyperglycaemia, dyslipidaemia, hypertension, high body mass index, low levels of physical activity and insulin resistance.13,14 Less strongly associated potential risk factors include: sleep apnoea; nonalcoholic fatty liver disease; levels of serum prolactin, serum adiponectin and serum homocysteine; age; renal disease; and pregnancy.13,15
The effects of diabetes on the retina are slow to manifest: signs of DR take approximately 10 years to appear after disease onset. The criteria for clinically significant DMO that will benefit from laser therapy were defined by the Early Treatment Diabetic Retinopathy Study Group during the 1980s.16 These include thickening of the retina (≤500 μm of the foveal centre), possibly with hard exudates and changes in the vasculature. Early damage is seen to both vascular and neural cells. Within retinal capillaries, pericytes and endothelial cells are lost leading to the appearance of ‘ghost’ capillaries. The relationship between neuropathy and vasculopathy in DMO, however, is unclear.
In DMO, there appear to be two pathological processes: a primary one causing vascular loss in small areas and a secondary one causing these areas to enlarge. The areas of vascular loss become ischaemic, stimulating growth factor secretion, which attracts inflammatory cells, and microaneurysms may occur. This creates a vicious circle of increasing vascular activation and inflammation in which areas of capillary nonperfusion tend to enlarge to sight-threatening DR (see Figure 1).2–4
Possibly the most important factor stimulated in DMO is vascular endothelial growth factor-A (VEGF-A). VEGF is secreted by all hypoxic cells and has vital roles in maintaining normal tissue function and in disease. These functions include: cell survival, permeability, angiogenesis, inflammation, mitogenicity, chemotaxis and neuroprotection.17,18 VEGF causes blood vessels to grow but also to leak and is consequently a target for several treatments of DMO.
Evidence that VEGF is involved is DMO first came from a South American tudy in which 88 patients with the disease were given intravitreal treatment with at least one injection of the anti-VEGF agent, bevacizumab (1.25 mg or 2.5 mg).19 In just 1 month, both VA in terms of best corrected VA (BCVA) and retinal thickness were significantly improved and this was sustained during 6 months of follow-up (p<0.0001 for both parameters). Anti-VEGF agents are now widely used to successfully treat DMO in the clinic as discussed in the next section.20
Further evidence of the role of inflammation in DMO has come from analysis of vitreous fluids from patients with DMO. Two studies conducted in Japan (n=92 and n=53) revealed significantly raised inflammatory cytokines in patients with DMO versus patients without diabetes (p<0.05 for all).21,22 In particular, levels of interleukin-6 (IL-6) IL-8, monocyte chemoattractant protein-1 (MCP-1) and intracellular adhesion molecule (ICAM-1) were elevated indicating an inflammatory state. In addition, corticosteroids such as triamcinolone acetonide have demonstrated efficacy in the treatment of DMO, reducing retinal thickness and improving VA.23,24 Corticosteroids have multifactorial actions against various inflammatory cytokines but they also have a direct effect in restoring the blood–retina barrier (BRB) in retinal endothelium, independently of inflammation. Their efficacy in DMO therefore not necessarily supports inflammation as a driver of DMO, a notion widely advocated.
The accumulation of fluids in retinal tissues in DMO follows the wellestablished principles of the Starling equation.25,26 This states that the flow of liquids between capillaries and surrounding tissues is the result of both hydrostatic pressure and osmotic pressure gradients resulting from vascular and tissue solute concentrations. In normal tissue, these forces are balanced and there is no net change in fluid volume but in DMO, they become unbalanced leading to disrupted fluid transport in and out of the tissue and fluid accumulation. The changes in hydrostatic and osmotic pressure and inflammation in DMO are mediated by multiple cellular and protein factors leading to localised breakdown of the BRB and consequent leakage into retinal tissues.27 The passage of plasma solutes out of retinal capillaries occurs through either a paracellular pathway via tight junctions between epithelial cells or via a transcellular pathway through the cells involving caveolae. Studies on rat retinal cells exposed to VEGF showed a transient down-regulation of some proteins such as occludin and claudin-5 that control tight junctions.28 In addition, long-term upregulation of the vesicular transport-related genes encoding caveolin-1 and plasmalemma vesicle protein-1 (PV-1) was observed. This indicates that in DMO there is a transient induction of paracellular transport but a more important sustained activation of transcellular transport that results in BRB breakdown. These findings were supported by a study of monkey eyes in which exposure to VEGF resulted in intense retinal microvascular leakage.29 Electron microscopy of leaky blood vessels in these tissues showed significantly increased pinocytotic vesicles (caveolae) that had moved to a luminal position indicating greatly increased transcellular transport. This increased transcellular transport of fluid in DMO is likely to decrease osmotic pressure in tissues, and as vascular hydrostatic pressure is also increased in DR, these forces result in net fluid movement to the tissues and consequent oedema.
Mr Ian Pearce
Royal Liverpool University Hospital, UK
The primary objective of diabetes management is to regain control of blood glucose levels and to optimise blood pressure and blood lipid levels.5 Despite these interventions, over time patients develop multiple diabetesrelated complications, of which they are often unaware of until they have progressed significantly.11 Patients with diabetes are 25x more likely to be blind than patients without diabetes30 and many patients fear blindness more than premature death. Ophthalmologists therefore have an important role in educating patients in the need to control hyperglycaemia, the major driver of retinopathy. Ophthalmologists are also likely to be the first healthcare professionals to observe end-organ damage in these patients and are responsible for reporting it to their treating physicians.
If diabetes is well managed, the risk of macro- and micro-vascular complications can be reduced.5,31 This has been demonstrated in several large-scale clinical studies including the Diabetes Control and Complications Trial (DCCT),32 the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study33 and the UK Prospective Diabetes Study (UKPDS).34 The DCCT showed that rates of DR rise steeply in patients whose glycated haemoglobin (HbA1c) levels are not maintained at approximately 7 % (see Figure 2).32 For every 1 % reduction in HbA1c there is a 30 % reduction in the progression of DR. However, too rapid a reduction in HbA1c can also result in DR and so no more than a 1 % reduction in 6 months is advised.
Clinical studies also show that controlling blood pressure in patients with diabetes is vital. The UKPDS showed that treating to target a blood pressure of <150/85 mmHg (preferably 140/80) substantially reduces the risk of death and complications due to diabetes.34 In another large study, DMO was reduced over a 4-year period when patients with diabetes received blood pressure-lowering treatment.35 Lipidaemia is also influential in DR but not as important as glycaemia and blood pressure. An analysis of data from the DCCT showed that higher serum lipids (total lipids, low-density lipoprotein [LDL] and highdensity lipoprotein [LDL] cholesterol, total-to-HDL cholesterol ratio and triglycerides) are associated with an increased risk of clinically significant MO (CSMO) and retinal hard exudate.36 In addition, other clinical studies have shown reductions in the progression of retinopathy following treatment with atorvostatin37 or fibrates.38
The successful management of diabetes and avoidance of DR requires a systematic approach in which all of the risk parameters are adequately controlled. An analysis of data from the National Health and Nutrition Examination Surveys (NHANES) in the US found that only 19 % of patients with diabetes met all their targets for HbA1c, blood pressure and LDL cholesterol, indicating a substantial need for improvement.39 To counter this problem, multifactorial treatments are needed to provide intensive control of all risk factors. This approach was evaluated in a Danish study (Steno-2) in which patients with diabetes received tight glucose regulation, renin-angiotensin system blockers, aspirin and lipid-lowering agents.40 This intense treatment regimen produced a risk reduction of 43 % for the progression of DR compared with controls given standard treatment (p=0.01).40
Until recently, laser photocoagulation was the ‘gold standard’ therapy in DMO and had been for 3 decades. Focal laser photocoagulation targets specific areas whereas macular grid photocoagulation applies a localised pattern to treat areas of diffuse leakage. Despite the emergence of medications, laser therapy still has a role in the treatment of DMO. The classic Early Treatment in Diabetic Retinophy Study (ETDRS) that commenced during the 1980s demonstrated that macular (focal/macular grid) laser photocoagulation was more effective when administered early rather than deferred.41 In this study, laser treatment reduced the risk of moderate vision loss by approximately 50 %, and was effective at preserving vision/preventing further vision loss. 41
More recently, intravitreal corticosteroid and anti-VEGF agents have become available for the treatment of DMO and these can be used as alternatives to or in combination with laser therapy.6,42,43 Extended-release implants of dexamethasone or fluocinolone acetonide are effective and more convenient than injections.44,45 In some countries such as the UK, however, their use is currently restricted to pseudophakic patients who are unresponsive to anti-VEGF agents due to their side effects.46
Since VEGF is a major factor in the inflammation that drives DMO, several anti-VEGF medications have been used effectively to treat DMO.20 These are used alone or in combination with or before laser therapy or with other treatments. Ranibizumab (Lucentis®) is a humanised fragment, a mouse monoclonal antibody and also binds VEGF-A.47 Aflibercept (Eylea®) is a fully human, recombinant fusion protein that tightly and stably binds VEGF-A and also binds placenta growth factor (PlGF).48 Bevacizumab (Avastin®) is a humanised, recombinant monoclonal antibody with high affinity for VEGF-A. It is widely approved for systemic treatments of various cancers but is also used extensively off-label for ocular diseases including DMO.49,50 Pegatanib (Macugen®) is a 28-base RNA aptamer covalently linked to two branched 20 kDa polyethylene glycol moieties that binds potently to VEGF. It is approved for the treatment of wet AMD but is also used for off-label treatment of DMO.51
Both aflibercept and ranibizumab have shown significant efficacy against DMO in clinical trials. The efficacy of aflibercept is discussed in the next section. In the ranibizumab for macular oedema (RIDE and RISE) phase III studies, patients with DMO were randomized to ranibizumab 0.3 mg/ day, 0.5 mg/day or sham treatment.7 In both studies there was a gradual improvement in VA in terms of BCVA scores in both ranibizumab-treated groups. At 24 months there was a substantially greater improvement in VA for the ranibizumab-treated groups compared with the controls (RIDE: 12.0, 10.9 and 2.3 letters; RISE: 12.5, 11.9 and 2.6 letters). At this stage, patients receiving sham injections were switched to ranibizumab 0.5 mg/day. Over the following 12 months, however, these patients showed slight gains but did not achieve the improvements in VA as those who had been treated with ranibizumab from the outset. This suggests that patients with DMO should be treated early to avoid vision losses.
Professor Jean-François Korobelnik
University Hospital, Bordeaux, France
The VIVID and VISTA trials are two similarly designed phase III head-tohead comparisons of the efficacy and safety of intravitreal aflibercept and laser therapy. These are double-masked, randomized, activecontrolled trials conducted at multiple centres in Europe, Asia, Australia, (VIVID) and the US (VISTA).8 A total of 872 patients with type 1 or 2 diabetes who presented with DMO with central involvement and ETDRS BCVA 20/40 to 20/320 were recruited. Patients were then randomised to receive either intravitreal aflibercept injection, 2 mg every 4 weeks (2q4), 2 mg every 8 weeks after five initial monthly doses (2q8), or macular laser photocoagulation.
Baseline characteristics were similar between groups within each study; central retinal thickness (CRT) was slightly greater in VIVID but the proportion who had received prior anti-VEGF therapy was markedly greater in VISTA (in the US) as was the duration of disease (see Table 1). The primary endpoint results show that VA rapidly improved during the first 4 weeks of treatment with aflibercept and then improved more steadily up to 52 weeks in both the VIVID and VISTA trials and this was substantially greater than improvements seen with laser treatment. BCVA letter improvement for aflibercept 2q8, 2q4 and laser treatment were 10.7, 10.5 and 1.2, respectively (p<0.0001), in the VIVID trial and 10.7, 12.5 and 0.2 (p<0.0001) in the VISTA trial (see Figure 3).8 This shows that less-frequent dosing of aflibercept at 8-week intervals rather than 4-week intervals has little effect on visual outcomes and both are markedly more effective than laser treatment. Early reports of data from extended follow-up in these trials indicate that the benefits in VA in aflibercept-treated groups subsequently stabilised and the similarities between dosing regimens and improvements over laser treatment were maintained to 100 weeks.9
For secondary endpoints, the aflibercept treatment regimens showed greater improvements in retinal pathology compared with laser treatment. Mean reductions in CRT at 52 weeks, as determined by optical coherence tomography (OCT), for aflibercept 2q8, 2q4 versus laser were: 192.4, 195.0 and 66.2 μm, respectively, (p<0.0001) in the VIVID study and 183.1, 185.9 and 73.3 μm, respectively, (p<0.0001) in the VISTA study (see Figure 4).
The aflibercept treatments were well tolerated and there were no newsafety signals compared with laser treatment. The frequency of ocular and non-ocular adverse events were similar in the three treatment groups in both studies. Treatment emergent, Anti-Platelet Trialists’ Collaborationdefined arterial thromboembolic events occurred at similar frequencies in the aflibercept 2q8, 2q4 and laser groups (3.5 %, 3.1 % and 2.8 %). These events were non-fatal myocardial infarction, non-fatal stroke and vascular death. Serious adverse events (SAEs) occurred in 1.7 % of aflibercept 2q8-treated patients, 1.7 % of aflibercept 2q4-treated patients and 4.2 % of laser-treated patients. SAEs included mainly cataract, DR and vitreous haemorrhage.8 Early reports indicate that no new safety concerns were identified in the additional period to week 100.9
In DMO, hyperglycaemia drives vascular dysfunction and low-grade inflammation that is mediated by numerous cytokines among which VEGF is central. These processes result in disrupted fluid transport, a breakdown in the BRB and leakage into retinal tissues causing fluid accumulation and macular cysts. Inhibition of VEGF action therefore is a key approach to managing DMO and preventing retinal damage and eyesight loss.
In patients with diabetes, prevention of DR progression to DMO requires a systematic approach. Management should involve good glycaemic control as well as control of blood pressure and lipid levels (especially LDL-cholesterol). Maintaining these parameters below set limits markedly reduces the risk of DR and DMO progression but this requires intensive monitoring and treatment. Such individual attention may not be available to all patients for durations of approximately 10 years that DMO takes to develop after diabetes onset.
For patients who have developed DMO, until recently, laser photocoagulation was the only treatment available.52 However, anti- VEGF agents, especially aflibercept and ranibizumab, are now becoming established as the ‘standard of care’ as first-line treatments for a subset of patients with DMO. In the RIDE and RISE trials, ranibizumab showed significant improvements in VA and retinal thickness versus sham treatment and was well tolerated in such patients. This efficacy was markedly better than can be achieved with corticosteroid treatments. The VIVID and VISTA trials, in which aflibercept was directly compared with laser-only treatment, showed significant improvements in efficacy over laser treatment. It is notable that the lower frequency (2q8) regimen provided similar efficacy to the higher frequency (2q4) and there were no safety differences between these two regimens. Aflibercept was very well tolerated and no retinal atrophy was seen in either of the VIVID or VISTA trials. In cases in which DMO occurs in both eyes of a patient, at some clinics, intravitreal treatment of the second eye is not given at the same time but may be delayed by hours or days. This ensures that there are no adverse events following the first injection; there is little evidence that consecutive injections of an anti-VEGF agent into each eye could cause clinically relevant blood levels of the medication. Use of the 2q8 regimen may be more suitable in cases of bilateral administration. It should be emphasised that anti- VEGF medications have not replaced laser therapy but can be used prior to, or in combination with, laser treatment especially in cases of focal leakage.
The disease stage at which different intravitreal medications should be used and how long they should be administered in DMO is not universally agreed. In the past, it was normal practice to wait until HbA1c levels and blood pressure were under control before initiating treatment. This strategy, however, costs time and may lead to retinal damage so now treatments increasingly commence earlier before these targets are reached. Anti-VEGF agents are indicated for first-line use in DMO when the fovea is involved and when vision is impaired. Corticosteroid injections and implants are mainly used second-line and are restricted to patients who fail to respond to anti-VEGF agents. Patients with a CRT of approximately 420 μm should be given anti-VEGF agents but those with a CRT of 370 μm may be given laser, corticosteroid or bevacizumab treatment. Treatment decisions are, however, challenging in DMO and therapy is often stopped too early. A 6-month course of anti-VEGF treatment is the absolute minimum but further treatment should be based on individual requirements.53,54 It is important when treating DMO that the clinician takes an interest in the patients’ diabetes history and adjusts the treatment strategy appropriately.
Aflibercept and ranibizumab are both approved for use in DMO. They have proved effective in clinical trials and as experience of their regular clinical use increases it is likely that they will be used in a wider population of patients as a first-line treatment. Such strategies are likely to help control this increasingly prevalent threat to vision worldwide.