NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Waugh N, Loveman E, Colquitt J, et al. Treatments for dry age-related macular degeneration and Stargardt disease: a systematic review. Southampton (UK): NIHR Journals Library; 2018 May. (Health Technology Assessment, No. 22.27.)

Cover of Treatments for dry age-related macular degeneration and Stargardt disease: a systematic review

Treatments for dry age-related macular degeneration and Stargardt disease: a systematic review.

Show details

Chapter 1Introduction to age-related macular degeneration

Age-related macular degeneration (AMD) is a progressive degenerative disease of the retina in which the macula is most affected.1 It is the commonest cause of blindness in the UK and it affects mainly older people.

Age-related macular degeneration goes through various stages, called early, intermediate and advanced. The first signs are yellowish deposits in the retina called drusen. Then abnormalities in the colour of the retina develop: paler areas called hypopigmentation, and darker areas with hyperpigmentation. Advanced AMD takes two forms, wet and dry, both of which lead to visual loss. Advanced dry AMD is characterised by atrophy of the retina – it wastes away and patches of retina and vision are lost. Because the patches were thought to resemble countries on a map, it became called ‘geographic atrophy’ (GA). The central most detailed vision is lost, making it difficult to drive, read or recognise faces.

Wet AMD, also called exudative AMD, is characterised by the development of abnormal new vessels [choroidal neovascularisation (CNV) and retinal angiomatous proliferation (RAP)]. It is now treated with drugs that inhibit a compound called vascular endothelial growth factor (VEGF), so they are called ‘anti-VEGF drugs’. They include bevacizumab (Avastin, Roche), ranibizumab (Lucentis, Novartis, Basel, Switzerland) and aflibercept (Eylea, Bayer). The AMD sections of this report are concerned with treatments for only dry AMD, at all stages, from prevention of early changes progressing to advanced AMD, both dry and wet, and treatment of advanced dry AMD. As part of the background, we also look at some epidemiological studies of risk factors for AMD.


The prevalence of AMD increases with age.2 Owen and colleagues3 reported an overall prevalence of advanced AMD in 2007–9 of 2.4% in the over 50s rising to 12.2% in the over 80s. The estimated number of people with advanced AMD in the UK was 513,000, about 2.4% of the population aged ≥ 50 years, with just over half (1.3%) having dry AMD. In the UK, there are about 2.6 million people with early AMD.

The Bridlington Eye Assessment Project (BEAP) showed that 38% of those aged > 65 years have no sign of AMD, 54% have early AMD, 2.8% have intermediate AMD and 4.5% have advanced AMD. The prevalence of advanced AMD rises with age, from 2.1% in those aged 65–70 years, to 7.5% in those aged 80–85 years, and 16% in those aged > 85 years.4 Visual acuity (VA) is often maintained at 6/9 or better in most eyes before the development of GA. AMD is by far the commonest cause of blind and partial sight certifications in the UK, accounting for about 59%.5

We have an ageing population with more people living longer; therefore, more people will live to develop AMD. They may otherwise be fit with a good quality of life, and so visual loss may have a dramatic effect in their remaining years.

We need to distinguish rates and numbers. The most recent meta-analysis of the prevalence of AMD in Europe, by Colijn and colleagues6 from the EYE-RISK consortium and the European Eye Epidemiology (E3) consortium, concludes that the prevalence of advanced AMD is now declining, perhaps because of healthier lifestyles. However, the number of people with any AMD will almost double.


Age-related macular degeneration causes central visual loss leading to gaps on items on which the eye naturally focuses, such as words on pages, bus numbers, faces and television. Vision becomes distorted, colours can fade and adaptation to dark can be impaired. Driving may become impossible. Visual impairment increases the risk of falls and injuries and can lead to depression and social isolation. Getting out and about safely, for example to go shopping, may become difficult. Independent living may become impossible. Sight loss is a leading cause of suicide among older people.7

Age-related macular degeneration reduces quality of life. Brown and colleagues8 assessed the quality of life among patients with mild [VA of 20/20 to 20/40 in the better-seeing eye (BSE)], moderate (VA 20/50 to 20/100 in BSE), severe (≤ 20/200) and very severe AMD (≤ 20/800). They used the time trade-off method, which asks how much of remaining life would be given up in return for perfect vision.

Patients scored their quality of life as:

  • 0.83 with mild AMD (similar to having moderate angina)
  • 0.68 with moderate AMD (similar to life following a moderate stroke, or having AIDS)
  • 0.47 with severe AMD (similar to end-stage renal failure on dialysis)
  • 0.40 with very severe AMD – a 60% loss of quality of life (similar to being bedridden after a major stroke or advanced prostate cancer with intractable pain).


The causes of AMD are not known. Risk factors include age, genetic predisposition, exposure to light, race, smoking, overweight and obesity, and diet.911 High fat diets and obesity increase the risk, whereas antioxidant nutrients protect. In the Danish Inter99 study, Munch et al.12 found that among people aged 30–60 years, macular drusen of > 63 µm was associated with physical inactivity, higher waist measurements (in men) and higher serum triglycerides (in women).

Chakravarthy and colleagues13 carried out a systematic review of risk factors for AMD, drawing on 18 cohort and six case–control studies. They found that cigarette smoking and a family history of AMD showed strong associations, and that there were moderate but consistent associations with risk factors for cardiovascular disease such as higher BMI, hypertension and higher plasma fibrinogen.

Smoking greatly increases the risk of AMD. The European Eye Study14 reported that current smokers had 2.6 times the risk of wet AMD and 4.8 times the risk of advanced dry AMD (GA) as opposed to non-smokers. The Melbourne Collaborative Cohort Study15 looked at patterns of diet, and found that diets rich in fruits, vegetables, chicken and nuts and low in red meat were associated with a lower prevalence of advanced AMD. Interestingly, they divided foods by method of cooking and noted that steamed fish conferred a lower risk than fried fish, probably reflecting broader dietary patterns. An earlier paper from the same study16 had reported that high red meat and processed red meat intake increased the risk of AMD, but that higher chicken intake reduced it. A third paper17 reported that higher trans-unsaturated fat intake was associated with increased prevalence of late AMD. Higher olive oil intake (> 100 ml/week) was associated with an odds ratio (OR) of 0.48 [95% confidence interval (CI) 0.22 to 1.04] compared with an intake of < 1 ml/week.

In a recent review, Zhu and colleagues18 provide a high-quality review of fish consumption and the incidence of AMD, with a meta-analysis of eight prospective cohort studies from the USA (n = 4), Australia (n = 2), Ireland (n = 1) and the Netherlands (n = 1). Some of the studies adjusted for a wide range of confounding variables, others for only a few. The incidence was reduced by 24% overall (OR 0.76, 95% CI 0.65 to 0.90, I2 = 50%; but heterogeneity in effect size not direction). Fish consumption is not clearly defined in the review but a diagram of the dose–response relationship shows that the reduction in OR increases with frequency, with once a week consumption reducing the risk by only 11% (RR 0.89, 95% CI 0.83 to 0.96). However, after an increase to three times a week, the relative risk (RR) plateaus at the 0.76 level.

High intake of dietary salt has also been suggested as a contributory cause.19 This could be mediated through its effect on blood pressure. The Complications of Age-related Macular Degeneration Prevention Trial (CAPT) research group reported that, compared with people who had normal blood pressure, those with definite hypertension (defined as a systolic blood pressure of ≥ 160, diastolic blood pressure of ≥ 95, or on treatment) had 1.55 times the risk of wet AMD and 1.86 times the risk of GA.20

Low-dose aspirin (100 mg on alternate days) taken for 10 years had no significant effect compared with placebo, with the hazard ratio (HR) for developing new AMD of 0.82 (95% CI 0.64 to 1.06).21 Heavy alcohol consumption (more than three standard drinks per day) was reported by Chong et al.22 to increase the risk of early AMD.

The prevalence varies among ethnicities, with the frequency of late AMD highest in white people, and lowest in Africans.10 This is partly due to varying genetic susceptibilities.

The presence of some genes increases susceptibility, particularly the complement factor H (CFH) gene, which is linked to the complement pathway, part of the immune system, and the Age-Related Maculopathy Susceptibility Gene 2 (ARMS2). These two genes are involved in > 60% of cases of advanced AMD.10 Conversely, some genes such as some variants of the apolipoprotein E gene, which regulates lipid and cholesterol transport in the central nervous system, appear to be protective. The mechanism underlying the protection may be via better transport of cholesterol and other metabolites out of the cells in the retinal pigment epithelium (RPE).

The structure of the eye

The sclera

The outer layer of the eyeball is the sclera, which forms part of the supporting wall of the eye. It is the ‘white of the eye’ and it surrounds most of the eye. However at the front of the eye, it is replaced by the cornea, which is transparent and allows light through.

The choroid

Inside the sclera, the next layer at the back of the eye is the choroid, which is the vascular layer of the eye, composed of blood vessels and connective tissue. There are sublayers within the choroid, including the choriocapillaris and Bruch’s membrane. The choriocapillaris consists of the capillaries that provide oxygen and nutrients to the retina.

Bruch’s membrane

Bruch’s membrane is the innermost part of the choroid, in contact with the retina. The innermost part of Bruch’s membrane is formed by the basement membrane of the RPE, which transmits waste products of metabolism from the photoreceptors (PRs) in the retina into the blood vessels in the choroid. The RPE and the choroid provide nourishment to the retinal PR cells.

Bruch’s membrane gets thicker with age, and this slows the transport of metabolites. With ageing, lipids accumulate in Bruch’s membrane. The conduction of fluids (hydraulic conductivity – the ability to let fluids pass) through the membrane is reduced.23,24 It is thought that oxidative change in the lipids may trigger an inflammatory process, including activation of complement. Reduced transport of nutrients into the retina and reduced transport of waste products of metabolism out of it may trigger a release of VEGF in an attempt to provide more blood supply, and this may lead to the development of abnormal new blood vessels in wet AMD. The RPE has a symbiotic relationship with the choriocapillaris; if the RPE is lost, then the choriocapillaris closes down. This is believed to be the result of reduced production of VEGF by the RPE.

There is a large variation in thickening of Bruch’s membrane with age. Lommatzsch et al.23 suggest that half of the thickening is due to natural ageing and half is due to other factors, such as genetic susceptibility and environmental factors. In early AMD, there is thickening of Bruch’s membrane due to lipid and protein deposits (drusen).25

The retinal pigment epithelium

The RPE lies between Bruch’s membrane and the PRs. Boulton and Dayhaw-Barker26 provide a good review of its functions, which include transport of ions, fluid and metabolites; support for the visual cycle; clearance of debris; protection against light and free radicals; and production of growth factors. The RPE changes with age. In AMD, accumulation of lipofuscin in the RPE can damage it.

The retina

The retina contains the PR cells, rods and cones, and it is only 0.5 mm thick. Retinal pigmentation is partly due to the presence of melanin in the RPE. There is more melanin in the macula so it appears darker.26 Melanin is protective but the amount of melanin falls with age, and one effect is to reduce antioxidant potential.26

The macula

The macula is an oval area near the centre of the retina, only 5.5 mm across. It is the most sensitive part of the retina. At the centre of the macula is the fovea. The macula is responsible for high acuity and colour vision. The macula is yellowish in colour due to the macular pigments.

Macular pigments

These consist of lutein, zeaxanthin and meso-zeaxanthin, which are found in high concentration in the macula and are known as macular pigments. The first two are obtained from the diet, and meso-zeaxanthin is formed in the macula from lutein. The levels are measured as macular pigment optical density (MPOD). Their distributions in the retina vary, with meso-zeaxanthin dominating in the centre of the macula and lutein at the periphery. Carpentier et al.27 provide an overview, with points including:

  • Adipose tissue may compete with the retina for uptake of lutein and zeaxanthin so obesity may lower MPOD.
  • Macular pigments protect the retina from the effects of blue light.
  • In the USA, the combined intake of lutein and zeaxanthin is about 2 mg per day.
  • Higher intakes appear to reduce the risk of AMD.
  • Taking supplements increases MPOD when that is low and some studies report improvements in visual function.

Lutein and zeaxanthin are members of the xanthophyll family. Most lutein and zeaxanthin comes from vegetables with highest concentrations, dark leafy vegetables such as spinach and kale, and egg yolk and maize. These carotenoids have antioxidant effects, protecting the RPE from oxidative stress. Increasing dietary intake leads to an increase in macular levels. A Dutch trial28 showed increases in VA and improvement in dark adaptation.

However, Stevens et al.29 from Aston University reported that among 158 patients recruited via the Macular Society helpline, those with AMD consumed a daily average 3.3 mg of lutein and zeaxanthin (text – table says under 2 mg/day), which is well below the 10 mg recommended after the AREDS 2 study.30 Many patients were not eating vegetables such as spinach and kale, but a control group of people without AMD had better intake.


Drusen are small yellow or white accumulations of extracellular material that build up between Bruch’s membrane and the RPE. Drusen come in two main forms: hard and soft. Most people > 40 years have a few small hard drusen, but if they are more numerous or if they are larger, they may be the start of macular degeneration.

Small drusen (< 63 µm) are considered by Holz et al.11 not to be associated with progression to AMD, but to be a non-specific change due to ageing. However, drusen volume31 or size32 are strong predictors of progression to GA or wet AMD.

Reticular pseudodrusen

Reticular pseudodrusen (RPD) are a specific phenotype of early AMD first described by Mimoun et al.33 as a yellowish interlacing network in the outer macula of AMD patients, best visualised under blue light. Other terms that have been used for RPD are subretinal drusenoid deposits, reticular macular disease and reticular drusen.

Arnold et al.34 reported RPDs typical predominant location between the upper edge of the fovea and the supero-temporal arcade. The fundus autofluorescence (FAF) findings of RPD and its common association with RAP were first noted by McBain et al.35 and Lois et al.36 and subsequently supported by others.37

The prevalence of RPD was initially reported to be 0.7% in the Beaver Dam Eye study38 and 1.95% in the Blue Mountains Eye study.39 The 15-year incidences were 3% and 4%, respectively. Later studies utilising multimodal imaging have reported a higher prevalence of the condition [4.9% in the Rotterdam study40 and 13.4% in the Alienor (Antioxydants, Lipides Essentiels, Nutrition et Maladies OculaiRes) study41]. This could be attributed to the fact that the former studies only utilised colour fundus photography (CFP) to detect RPD but the latter used newer imaging technologies. Known risk factors for RPD include age and female gender.3841

Reticular pseudodrusen are associated with all stages of AMD as well as being more prevalent in late AMD.4246 The prevalence of RPD in early AMD was reported to be 8.4% in the AREDS study,47 36–54% in neovascular age-related macular degeneration (nAMD) and ranging from 29% to 92% in GA.42 The Beaver Dam Eye study reported that eyes with RPD are at a sixfold higher risk of progressing to late AMD within 5 years than eyes with indistinct soft drusen but no RPD.38 The Blue Mountains Eye study reported a fourfold increased risk.39 Gil and colleagues48 studied the fellow eyes of patients with unilateral wet AMD and found that 58% had RPD, and that RPD increased the risk of progression compared with patients without RPD.

Fellow eyes of patients with unilateral wet AMD are known to have a higher risk of progression to late AMD.47,49 Studies in this group have shown that presence of RPD is an independent and additional risk factor (when combined with drusen and pigmentary changes) for progression to late AMD in the fellow eye.46,48,50,51 Eyes with RPD tend to progress to GA50,51 but some studies have also reported a higher risk of wet AMD.34,44 In patients with established GA, Marsiglia et al.52 reported that eyes with RPD have a higher rate of progression than eyes without RPD.

Reticular pseudodrusen have been reported to cause significant deterioration in rod function,42 although central VA is preserved, as reported by Hogg et al.50 using the Smith-Kettlewell low luminance acuity test and by Steinberg et al.53 using microperimetry. Compared with areas with no pathologic morphology, areas with RPD demonstrated a large and sharp decrease of scotopic sensitivity while there was only a mild decrease in photopic sensitivity.53 Other studies have shown that RPD are associated with reduction in photopic sensitivity when compared with healthy controls or people with typical drusen.5456 Ooto et al.57 suggested that in order to truly reflect a patient’s visual function, other parameters, such as contrast sensitivity and mesopic sensitivity, should be measured along with VA, as RPD are associated with deterioration in both.

Corvi et al.58 compared MPOD in patients with RPD and people without AMD, and reported lower levels. They also reported reduced best corrected visual acuity (BCVA) and retinal sensitivity. After 3 months supplementation with lutein 10 mg/day and zeaxanthin 2 mg/day, the mean MPOD in the RPD group improved to the same levels as in the control group. However, no significant improvements were seen in BCVA or retinal sensitivity. This may be because changes in function take longer to accrue. In the CREST study, Nolan and colleagues59 found that MPOD increased by 3 months but that changes in contrast sensitivity took 12 months to reach statistical significance.


Melanin in the choroid is protective against oxidative damage, and a reduction in pigment in the eye may increase the risk of developing AMD.

Retinal hypopigmentation results in paler areas and is usually associated with loss of the RPE cells. Conversely, hyperpigmentation can occur in early AMD. Neither change is specific to AMD. Another pigment, lipofuscin, appears harmful. It is composed of lipids and protein. A major component is a retinoid A2E (N-retinyl-N-retinylidene ethanolamine) which is a by-product of the visual cycle.

The ‘visual cycle’

Light reaching the photoreceptors in the retina triggers the conversion of the light-sensitive retinoid 11-cis-retinal into a different form, 11-cis-retinol, thereby generating an electrical signal to the brain. The trans form is then converted back to the cis form in the RPE and then returns to the photoreceptors, completing the visual cycle. If the two molecules of the trans form combines with one of the lipids (phosphatidilethanolomine) in the RPE, A2E is formed, and this can impair RPE function. Because the edges of patches of GA are thought to have A2E accumulation [as reflected in increased autofluorescence (AF)], reducing that accumulation may be one target of drug treatment.

Oxidative stress

Oxidative stress is defined by Betteridge60 as a disturbance in the balance between the production of reactive oxygen species (free radicals) and antioxidant defences. The retina has a very high metabolic rate, reflected in high oxygen consumption, and has a high concentration of polyunsaturated fatty acids and exposure to light, which, if coupled with inadequate levels of antioxidants, can make it very susceptible to oxidative stress. Yehoshua and Rosenfeld61 report that the evidence for cumulative oxidative damage being the cause of AMD has been growing, but that a mechanism for it is not yet known.

Barnett and Handa62 suggest that oxidative stress can affect the immune system, turning it from a protective to a pathological response, and can also lead to chronic inflammation.

The immune system

Ambati and colleagues63 have reviewed the immunology of AMD. In brief, they consider that overactivity in the alternative pathway of the complement system is involved in the development of AMD. This is associated with the genetic susceptibility via a variant of the CFH gene, known as CFH (402His), which causes a greater than normal complement response to retinal injury. Ambati and colleagues63 suggest that in individuals with ‘a complement hyperinflammatory phenotype’ there is an over-reaction to cellular damage in the retina.

Anderson, Hageman and colleagues64 first described the role of inflammation in AMD, and put forward the hypothesis that drusen were the result of local immune-mediated processes and the junction of the RPE and Bruch’s membrane.

The pathological role of the complement system has led to trials of drugs to inhibit that system.

Age-related macular degeneration

Classification (Macular Research Classification Committee 2013):65

  1. Normal ageing – people with small drusen (< 63 µm), also termed drupelets, should be considered to have normal ageing changes with no clinically relevant increased risk of late AMD developing.
  2. Early AMD – medium drusen (≥ 63 to < 125 µm), but without pigmentary abnormalities thought to be related to AMD.
  3. Intermediate AMD – large drusen or with pigmentary abnormalities associated with at least medium drusen.
  4. Late AMD – neovascular (wet) AMD or GA (advanced dry AMD).

In this report, we use the term dry AMD to cover all stages from early AMD to GA.

Early and intermediate AMD is characterised by drusen, and/or by changes in pigmentation.25 However, most people with drusen will not progress to severe visual loss and drusen may cause only mild or no visual symptoms. Up to 80% of people > 60 years have some drusen. Hard drusen are well-defined yellowish deposits with little risk of progression.

Soft drusen are larger, not well demarcated and are associated with a high risk of progression to late AMD. They may become larger and merge over time and can lead to RPE detachments, called drusenoid RPE detachments. They may disappear, but this is usually associated with atrophy of the outer retina. Drusen are associated with thinning of the overlying RPE. Fleckenstein and colleagues66 consider that GA is the natural end stage of soft drusen. A key component of drusen is amyloid beta,67 which is a waste product.

The underlying processes include locally intensive metabolism, oxidative stress, chronic inflammation, a pathological immune response and lipofuscin accumulation.9 Lipofuscin is considered toxic.

In atrophic AMD, there may be a single patch of atrophy or several. Over time, the patches may get bigger and may merge. The foveal centre (the area responsible for central vision) is lost last as atrophy occurs around the centre of the macula first before expanding into the fovea, which is the very centre of the macula. This potentially gives time for treatment before the central vision is lost.

Vision is lost from atrophic patches and the gaps in vision are called scotomas.

The atrophy is due to loss of the RPE, outer layers of the retina and the underlying choriocapillaris.11,66 On optical coherence tomography (OCT), GA appears as a flat patch where the retinal has withered away. A total of 20% of people with legal blindness have lost central vision due to GA. It tends to be of similar extent in both eyes66 but patients can have GA in one eye and wet AMD in the other, and can also have both GA and wet AMD in the same eye, if late dry AMD turns to wet AMD.

Geographic atrophy is also seen in patients treated with anti-VEGF drugs for wet AMD.35,68 In both the Comparison of Age-Related Macular Degeneration Treatments Trials (CATT)69 and Inhibition of VEGF in Age-related choroidal Neovascularisation (IVAN)70 trials (of ranibizumab vs. bevacizumab in wet AMD) it was observed that about one-fifth of patients developed GA after 2 years of anti-VEGF treatment.71 This GA appeared to be clinically similar to the GA that is seen in dry AMD and may occur because VEGF is required for the maintenance of the choriocapillaris by the RPE.

Progression and natural history

Data on natural history studies are important because natural history may be the only comparator for some interventions reported in observational studies.

Wet AMD will develop in 10–15% of people with intermediate AMD.72 In the AREDS trial, the average time to atrophy was 5–6 years in people with large drusen and hyperpigmentation, but 2.5 years in those with hypopigmentation.

Most people with AMD are at the early stage,2 as shown in Table 1.



Prevalences of dry AMD by age and stage

The KORA (Cooperative Health Research in the Region of Augburg) study from South Germany reported features of AMD in people < 50 years.73

The Geographic Atrophy Study by Sunness et al.74,75 reported that GA enlarged at 2.6 mm2 per year over a median follow-up of 4.3 years, in 212 eyes in 131 patients, mean age 78 years. However, there was a very wide range of progression rates from none to almost 14 mm2 per year. They noted a high concordance in rates of enlargement between eyes.

The Geographic Atrophy Progression Study,76 in patients with a mean age of 77 years, found that the GA enlarged by an average of 1.85 mm2 over 12 months, based on AF, and this was slightly higher based on CFP.

The Beaver Dam Study77 found a progression rate of 1.3 mm2 per year in 53 eyes of 32 patients (mean age about 81 years) over 5 years.

The AREDS trial group reported progression of 1.7 mm2 per year in 251 eyes of 181 patients (mean age 70 years) over a median follow-up of 6 years.

The FAM (Fundus Autofluorescence in age-related Macular Degeneration) study78 reported a similar progression rate of 1.75 mm2 per year (mean) or 1.52 mm2 per year (median) in 195 eyes of 129 patients (mean age 74 years), but over a median follow-up of only 1.8 years. They also reported a wide range of progression rates. They used FAF to determine areas of GA.

Decision problem

The questions for this review include:

  1. Can treatment of early AMD prevent or slow progression to advanced forms (wet or dry)?
  2. Can any treatments improve, or slow deterioration in, GA?
  3. Can any treatments prevent GA progressing to wet AMD?

As our aim is to identify interventions that might have reached a stage where they could be assessed by the NIHR programmes [mainly Efficacy Mechanism and Evaluation (EME) and Health Technology Assessment (HTA)], we are not interested in –

  • rehabilitation methods such as external low visual aids
  • diagnostics
  • research still at basic science stage, such as in vitro, including cell work, or methods of carriage of gene therapies into cells using viral carriers
  • treatments with some evidence of efficacy in animal studies but not yet tested in humans. Such research might fall within the remit of the Medical Research Council (MRC) Translational Research Programme.

Potential treatments might be divided into the following groups:

  1. Treatments where proof of concept in humans has already been achieved but where research is needed to evaluate clinical efficacy, and which might be suitable for the EME programme.
  2. Treatments where there is evidence that shows they can be effective, but where further research is needed to establish the clinical effectiveness and cost-effectiveness for the NHS in comparison with the current best alternative. Such research falls within the remit of the HTA programme.
  3. Interventions where there is sufficient evidence of lack of benefit, so that no further research is justified.
  4. Interventions where there is no good evidence of any benefit and on which no money should be spent. Identifying these may help people who see unjustified claims or adverts.


The most important outcomes are those that matter to patients: VA, contrast sensitivity, adverse effects of treatment, reading speed, ability to drive, health-related quality of life, progression of disease and patient preference.

However, VA loss is a late manifestation of AMD and not a good primary outcome in most trials, especially when the treatment is aiming at prevention of visual loss before it occurs. Early AMD may cause minimal or no symptoms. VA depends only on the centre of the fovea but this tends to go last in atrophic AMD and many patients have large areas of atrophy and experience considerable problems before the fovea goes. Reading and seeing faces of people can be extremely difficult and the ability to drive may be lost.

Progression of dry AMD is slow, and so it could be years before a trial could show a decline in vision. Therefore, predictors or biomarkers of future decline can be accepted if there is good evidence that they are strong predictors of subsequent visual outcomes. These will include changes detectable by investigation, but not necessarily by people with AMD, including:

  • Rod function, which may not correlate with VA as central VA acuity (as measured using VA charts) depends on foveal function, and the fovea is cone rich. But rod function is one of the earliest abnormalities detected in people who will later develop GA in AMD.
  • Macular pigment density, because it appears to be protective.
  • Integrity of the RPE layer, as determined by FAF and OCT.
  • Drusen volume and number. Disappearance of drusen may be a sign of developing GA.
  • Macular sensitivity, which can be measured by microperimetry.
  • Dark adaptation.

Both photopic and scotopic vision need to be considered. Scotopic vision refers to low levels of light such as in near darkness.

One issue is the clinical significance of changes in VA. In past evaluations, for example of the anti-VEGF drugs, a clinically significant difference in VA has usually been considered as a change of ≥ 10 letters. Changes of < 5 letters are not regarded as clinically relevant as may indicate normal variability. Changes of 5–9 letters are not regarded as clinically useful but might be regarded a valuable outcome to investigate if seen in a short-term study, suggesting that a larger or longer trial is justified.

In dry AMD, no change (which could be lack of deterioration) could be regarded as clinically meaningful if observed over a long enough period.


Microperimetry can detect changes in macular sensitivity in patients with early AMD and normal VA.79

Macular sensitivity measured using microperimetry focuses on the central macula instead of the entire visual field.8083

Testing is performed either with a modified Humphrey Field Analyser or with a microperimeter.83

There is limited evidence of the reproducibility of microperimetry in patients with AMD, but current studies have suggested that it provides consistent and reproducible readings.8486

Early AMD patients have rod sensitivity loss87 and impaired rod-mediated parameters of dark adaptation,88 which worsen as AMD progresses. The association between early AMD changes and macular sensitivity was further established by the observation that a correlation existed between altered AF signal and reduced macular sensitivity.79,89 In GA, macular sensitivity was reduced in areas of increased fundus AF signal at the junctional zone of areas of atrophy.90 However, this observation has not yet been proven to be a predictor of GA enlargement over time.90

Current evidence suggests that macular sensitivity is a valuable biomarker for early AMD and microperimetry has proven to be an easy and reliable test to measure it. It is not widely used in clinical practice, but has been used in clinical trials to evaluate the effects of treatments on macular sensitivity.9196

Review methods

For reasons of space, we summarise methods here. Further details are provided in Appendix 1, including the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram.

Search strategies

MEDLINE, EMBASE, Web of Science and The Cochrane Library were searched from 2005 to 13 July 2017 for reviews, journal articles and meeting abstracts. Searches were limited to English language.

Initial searches of all databases were undertaken in June 2016 and updated searches were run in June 2017 to check for any articles added in the previous year. The Association for Research in Vision and Ophthalmology (ARVO) website was also searched for meeting abstracts.

References of reviews were checked for relevant studies and clinical experts were also consulted for any other relevant literature.

Studies were selected for inclusion through a two-stage process using predefined and explicit criteria. Titles and abstracts of 7948 articles from the full literature search results were screened independently by two reviewers to identify all citations that appeared likely to have met the inclusion criteria, and checked by a third. The full texts of 398 articles were obtained for further screening and checking of references and 112 articles were included in the final report., the WHO search portal and UK Clinical Trials gateway were searched for ongoing and recently completed clinical trials.

Full details of the search strategies are in Appendix 1.

Inclusion and exclusion criteria


People with a confirmed diagnosis of dry AMD or Stargardt disease (STGD).


Any interventions that aim to preserve or restore vision in dry AMD or STGD.


To avoid overlap, we excluded studies on some interventions being reviewed in the NICE guideline process (e.g. smoking cessation, diagnostic technologies, monitoring and review, and rehabilitation support).


These are as above.


We placed no restriction on study design so included randomised controlled trials (RCTs), controlled clinical trials (CCTs) with a concurrent control group, and observational studies. This was partly so that we could assess the evidence base for treatments that might be advocated without a sufficient evidence base.

Systematic reviews were assessed for quality and summarised if they met quality criteria. Reviews were also used as a source for identifying primary studies, and for identifying studies published before 2005 that seemed relevant, such as earlier studies of included treatments.

Study selection and data extraction

Studies published as abstracts or conference presentations were only data extracted and included if sufficient details were presented to allow an appraisal of the methodology and the assessment of results to be undertaken. If such details were not available, key points from abstracts were summarised in the text.

Data were extracted by one reviewer using a standard data extraction form and checked by a second reviewer. At each stage, any disagreements between reviewers were resolved by consensus or, if necessary, by arbitration by a third reviewer.

Quality assessment strategy

The methodological quality of primary research studies was assessed using criteria based on those recommended by the Cochrane Collaboration and National Institutes of Health (NIH), National Heart, Lung and Blood Institute (NHLBI) (for further details, see Appendix 1). Quality criteria were applied by one reviewer and checked by a second reviewer, with any differences in opinion resolved by consensus or by arbitration by a third reviewer.

The quality of systematic reviews was assessed using the Centre for Reviews and Dissemination (CRD) checklist, with reviews assessed as good if four or more criteria were met.

Method of data synthesis

Studies were synthesised through a narrative review with tabulation of results of included studies. Formal synthesis through meta-analysis was not possible because studies were not of sufficient quality and were heterogeneous in terms of participant characteristics, outcomes and study design.


Trials and other studies listed by National Clinical Trial (NCT) numbers are available on the website ( by searching using NCT number. This website is a service of the US NIH.

Changes to the protocol

An outline protocol was registered on PROSPERO at an early stage. (This is mandatory for reviews commissioned by the HTA programme). However, during the systematic review, the protocol evolved over time, as agreed by the funder. The main change was to include additional outcomes, or predictors of outcomes, because of the awareness that many studies were relying on VA, which is a late outcome.

Quantity of evidence

We included 108 primary studies reported in 112 articles (see Figure 2). Of 104 dry AMD studies, there were 26 of pharmacological treatments, 30 in physical therapies, 3 of cell transplants and 45 of nutritional supplements. There were four studies in Stargardt’s, two of physical therapies and two of nutritional supplements. Two studies had subgroups of people with dry AMD and Stargardt’s97,98 making a total of six studies in STGD.

An overview of the study characteristics can be seen in Report Supplementary Material 6. There was a range of study designs, with 60 RCTs and CCTs, 24 cohort studies and cross-sectional studies, 13 single-arm before-and-after studies, 5 case–control studies, and 6 case series. Many studies had small sample sizes, the durations of intervention and follow-up were often short, and there were differences in the outcomes reported. We reported all outcomes of relevance if they were reported by the authors of each study. Further details are provided in Chapters 26. Baseline characteristics of participants are summarised in Report Supplementary Material 6. There was generally poor reporting of baseline characteristics across the studies. The risks of bias of RCTs and CCTs and quality of non-randomised studies are summarised in Report Supplementary Material 6. The overall quality of each study is reported within the results chapters of this report.

Details of methods and quality assessments are in Appendix 1. Report Supplementary Material 1–5 contain data extraction and quality assessment tables and can be downloaded as separate files from the HTA programme website (URL: Report Supplementary Material 6 has a list of excluded studies, most of which were excluded because they were on wet AMD or basic science, or were superseded by later studies.

Image 15-09-10-fig2
Copyright © Queen’s Printer and Controller of HMSO 2018. This work was produced by Waugh et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.
Bookshelf ID: NBK500483


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.1M)

Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...