Managing diabetic retinopathy
BMJ 2010; 341 doi: https://doi.org/10.1136/bmj.c5400 (Published 25 October 2010) Cite this as: BMJ 2010;341:c5400
- Correspondence to: Z Ockrim zoe_ockrim{at}hotmail.com
- Accepted 8 September 2010
Summary points
Prevalence of blindness caused by diabetic retinopathy is rising, despite effective treatments
Duration of diabetes is the most important risk factor; hypertension and hyperglycaemia are other risk factors
Fundus photography is the best way to detect retinopathy early—when still treatable and before vision is irreversibly lost
Annual retinal examination is recommended for patients with diabetes, although compliance is low
Laser photocoagulation and intravitreal injection of anti-vascular endothelial growth factor agents are effective treatments for diabetic macular oedema
Laser photocoagulation and vitrectomy are effective treatments for proliferative retinopathy
Diabetic retinopathy is the most common cause of blindness in working age people in England, Wales, and Scotland. Despite the availability of effective treatment, the prevalence of blindness as a result of diabetic retinopathy is increasing. Between February 1999 and March 2000, the number of patients on the blind register as a result of diabetic retinopathy in this age group was 2.05 per 100 000 population compared with 1.26 per 100 000 population in 1990-1.1
The global prevalence of diabetes in adults will rise from 171 million in 2000 to 366 million in 2030.2 Increased visual impairment as a result of diabetic retinopathy will impose a human and economic cost on entire communities, particularly in middle income countries.
Diabetic retinopathy is a chronic yet treatable condition. Primary prevention by improved medical management of diabetes, early detection, and timely treatment all reduce the risk of visual loss. In addition, the emergence of new treatments will change the management of retinopathy.
Sources and selection criteria
We searched Medline and Embase with the keywords “diabetic macular oedema”, “diabetic maculopathy”, “diabetic retinopathy”, and “controlled clinical trials”. We used evidence from published abstracts from major international scientific meetings and from the UK National Institute for Health and Clinical Excellence. We gave priority to evidence from well conducted systematic reviews and large well designed randomised controlled trials.
Who gets diabetic retinopathy?
The duration of diabetes is the major risk factor for the development of diabetic retinopathy. Large longitudinal studies of patients with diabetes in Wisconsin found that retinopathy develops within five years of diagnosis of diabetes in about 25% of people with type 1 diabetes,3 40% of people with type 2 diabetes who are taking insulin,3 and 24% of people with type 2 diabetes who are not taking insulin.3 The 25 year cumulative rate of progression to diabetic retinopathy, diabetic macular oedema, and clinically relevant macular oedema was 83%, 29%, and 17%, respectively.3 In addition, patients with poor glycaemic control and uncontrolled hypertension are at greater risk than those with good control of these factors.4 5 6
What changes occur in diabetic retinopathy?
Diabetic retinopathy is a consequence of microvascular changes. The first changes are pericyte death and basement membrane thickening, which lead to altered blood flow in the retinal capillaries.
These changes lead to capillary leakage and capillary vessel occlusion, with subsequent formation of new vessels. There are two different subtypes, which may coexist (for a summary of normal anatomy see fig 1⇓).
Fig 1 Normal retinal anatomy. The optic nerve carries sensory information and allows the retinal vasculature to enter and exit the retina. The macular area, in particular the fovea, is responsible for most detailed vision
Non-proliferative and proliferative diabetic retinopathy
Non-proliferative diabetic retinopathy is characterised by microaneurysms, haemorrhages, venous beading, capillary loss, and intraretinal microvascular abnormalities (fig 2⇓). The rate of progression to proliferative diabetic retinopathy after 10 years is 6.6%.7
Fig 2 Direct visualisation of the retina can illustrate the microvascular changes that are associated with diabetic retinopathy such as microaneurysms (arrow; A), retinal haemorrhages (arrowhead; A) and venous beading (B)
Proliferative diabetic retinopathy develops secondary to capillary closure—which upregulates growth factors, such as vascular endothelial growth factor—and is defined by the growth of abnormal new vessels from the retina or optic disc on to the posterior surface of the vitreous or the iris (fig 3⇓). These blood vessels may rupture, causing vitreous haemorrhage, or may form a sheet of fibrovascular tissue that exerts traction on the retina, leading to retinal detachment.
Fig 3 Proliferative diabetic retinopathy is characterised by new blood vessels on the retina (arrow; A), the optic nerve (arrowhead; A), or the iris (B)
Diabetic maculopathy
Increased capillary permeability occurs early in diabetic retinopathy.8 Hyperglycaemia leads to an increase in retinal blood flow as a result of changes in perfusion pressure and a lack of compensatory autoregulation in the diabetes.8 This, in turn, leads to increased exudation in the retina. The increased perfusion pressure may also contribute to the production of microaneurysms by exerting pressure on the vessel wall, which is weakened by the loss of pericytes.9
Vascular leakage may also occur in response to capillary closure. Increased basement membrane thickening reduces the bore of affected vessels. Other factors that affect retinal ischaemia include adhesion of leucocytes to the vascular endothelium as a result of upregulation of adhesion molecules, such as intercellular adhesion molecule 1.
Diabetic maculopathy occurs when retinopathy affects the macula and central visual acuity is threatened. Patients with type 2 diabetes have a higher prevalence of maculopathy than those with type 1 diabetes (53% v 42%).10 The condition can be further classified as diabetic macular oedema and macular ischaemia, and the two often coexist. Macular ischaemia is characterised by capillary loss in the macular area; it is demonstrated using fluorescein angiography (fig 4⇓) and currently is untreatable. Diabetic macular oedema is caused by increased permeability of the inner blood-retina barrier and decreased efflux of fluid across the retinal pigment epithelium. This leads to the accumulation of intraretinal fluid, which causes retinal swelling and reduced central vision. The Early Treatment of Diabetic Retinopathy Study defined clinically relevant macular oedema as diabetic macular oedema threatening central vision (box).
Definition of clinically significant macular oedema
The Early Treatment of Diabetic Retinopathy Study defined clinically relevant macular oedema as any one of the following:
Thickening of the retina at or within 500 µm of the centre of the macula
Hard exudates at or within 500 µm of the centre of the macula if associated with thickening of adjacent retina
Zone(s) of retinal thickening one disc diameter or larger (1500 µm), any part of which is within one disc diameter of the centre of the macula11
Fig 4 Macular ischaemia is characterised on fundus fluorescein angiography by capillary loss and an increase in the size of the foveal avascular zone (arrow)
Diabetic macular oedema may be focal or diffuse. Focal oedema is mainly caused by leakage from microaneuryms, dilated retinal capillaries, or intraretinal microvascular abnormalities. Clusters of microaneuryms may be seen in areas of circinate exudate, and fundus fluorescein angiography confirms that the microaneurysms are leaking. Diffuse oedema is caused by generalised leakage from dilated capillaries throughout the posterior pole and is generally associated with capillary loss.
How is diabetic retinopathy diagnosed?
The main symptom of diabetic retinopathy is reduced vision, but this occurs only when the condition is advanced and may be irreversible. Early changes in diabetic retinopathy are generally asymptomatic, and treatment may be needed long before patients are aware of any loss of vision. All patients over the age of 11 years with type 1 diabetes must have a retinal examination annually or more frequently if clinically indicated.12 Type 2 diabetes may have been present for years before diagnosis, so all patients with type 2 diabetes should have a retinal examination at least every 12 months, starting as soon as the diagnosis is made.
In 2000, a review of available cohort studies concluded that the best diagnostic test was dilated retinal photography, with ophthalmoscopy if photographs do not allow for grading of changes.13 The photographs are graded according to the presence of the vascular changes described above. However, retinal thickening and macular oedema can be seen only with stereoscopic viewing of the retina. Fortunately, other easily visible signs such as microaneurysms, haemorrhages, and exudates in the macular area act as markers for macular oedema (fig 5⇓).
Fig 5 Surrogate markers of diabetic macular oedema include microaneurysms (arrowhead; A), exudates (arrow; A), and intraretinal blot haemorrhages (arrow; B)
How is diabetic retinopathy managed?
Evidence from large randomised trials and long term follow-up studies show that primary prevention, early detection, and effective treatment reduce the risk of visual loss. Because the disease is generally asymptomatic until the late stages, patient education is crucial. All patients should understand why they need to attend for retinopathy screening and treatment and that good control improves their chances of retaining good vision
Optimal control of diabetes involves aiming for a glycated haemoglobin (HbA1c) of 6.5% (intensively lowering blood glucose to a target 6.0% has been associated with increased mortality14), blood pressure of 140/80 mm Hg or lower, and total cholesterol less than 4 mmol/l or low density lipoprotein cholesterol less than 2.0 mmol/l.15 However, even patients with “good” control develop complications 20 years after the onset of diabetes.
Primary prevention
Optimal glycaemic control
Improved glycaemic control can reduce but not abolish the risk of retinopathy. The Diabetes Control and Complications trial (DCCT) found that intensive treatment reduced the risk of development and progression of diabetic macular oedema by 26% compared with usual management over nine years of follow-up.16 Similarly, the UK Prospective Diabetes Study (UKPDS) published in 1997 examined the role of tight control in patients with type 2 diabetes. It found a 17% reduction in the risk of progression of retinopathy, a 29% reduction in the need for laser treatment, and a 16% reduction in the risk of legal blindness in the intensively treated group compared with those randomised to usual management at 10 years.5 In practice, however, perfect glycaemic control is unattainable in type 1 diabetes because of unpredictable hypoglycaemia during ultra-tight control. Perfect control is also unattainable in most people with type 2 diabetes, and control tends to deteriorate with time.5
Control of blood pressure
UKPDS randomised patients with hypertension either to tight control of blood pressure (<150/85 mm Hg) with a β blocker or an angiotensin converting enzyme inhibitor (and other agents if needed) or to less tight control (<180/105 mm Hg) without the use of these agents. After seven years’ follow-up, the progression of diabetic retinopathy was reduced by 35% in the tight control group compared with the less tight group. At nine years, the risk of moderate visual loss and the need for laser treatment were reduced by 47% and 35%, respectively, in the tight control group.6 No benefit of the angiotensin converting enzyme inhibitor (captopril) over the β blocker (atenolol) was seen.6
Several large interventional studies have looked at the effect of individual angiotensin converting enzyme inhibitors in patients with diabetes.17 18 However, the effect on diabetic retinopathy has been, at best, a secondary outcome measure, and we have no clear evidence that one method of lowering blood pressure is superior to another.
The DIabetic REtinopathy Candesartan Trial (DIRECT) was a large randomised trial designed to assess whether blockade of the renin-angiotensin system reduced the incidence or progression of diabetic retinopathy in normoalbuninuric normotensive patients. Candesartan had no effect on the incidence of new diabetic retinopathy in patients with type 1 diabetes and no effect on the progression of established retinopathy in patients with type 1 and type 2 diabetes.
Lipid lowering agents
Observational data from the Early Treatment of Diabetic Retinopathy Study suggest that lipid lowering agents may decrease the risk of vision loss in patients with diabetic retinopathy.19 The mechanisms by which such agents improve exudative diabetic retinopathy are unclear, although laboratory evidence suggests that oxidised low density lipoprotein cholesterol is toxic to retinal endothelial cells.20
More recently the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) study was a large randomised study looking at the effects of long term fenofibrate on cardiovascular events in patients with type 2 diabetes. A tertiary end point showed that fewer patients who received fenofibrate needed laser treatment than in the control group (3.4% v 4.9%; P=0.0002; number needed to treat 67, 95% confidence interval 43 to 143).21 The protective effects seemed to be independent of blood glucose, blood pressure, and baseline lipid values. Further research is needed to substantiate these findings.
Protein kinase C inhibitors
Protein kinase C has a role in the pathogenesis of diabetic macular oedema.22 Ruboxistaurin was designed as an orally active protein kinase C-β inhibitor, but in a large randomised controlled trial it failed to reduce the progression of maculopathy in patients with early diabetic retinopathy.23 However, a post hoc analysis showed that it did reduce the incidence of moderate visual loss. More effective agents to prevent retinopathy may become available in the future.
Glitazones
Glitazones are effective at lowering glycosylated haemoglobin, but their use is limited by systemic side effects, such as increased peripheral pedal oedema (pioglitazone), cardiac failure, increased incidence of myocardial infarction (rosiglitazone), and fractures.24 In addition, a 2.6-fold increase in diabetic macular oedema was noted in one large case series.25 Although this seems to be reversible,26 it has made ophthalmologists wary of using glitazones in the presence of diabetic macular oedema.
Regular screening and early detection
Diabetic retinopathy is an appropriate target for screening: it is an important health problem with a recognisable presymptomatic state, the screening procedure is considered acceptable, an appropriate treatment is available, and screening is cost effective.27
In 2001, the National Service Framework for Diabetes Standards was published for England and Wales.15 It stated that by 2007 all patients with diabetes should be screened annually for diabetic retinopathy, preferably by digital fundus photography.
Current screening techniques rely on visual inspection of photographic images and are therefore labour intensive. Digitally acquired images may be screened at several levels. The most basic level separates eyes that have any diabetic retinopathy from those that have none. Higher levels assign grades of retinopathy and divide the disease into referable or observable retinopathy. Computerised image analysis may be a cost effective means of grading these images. A recent study of an automated grading algorithm showed a sensitivity for detection of observable or referable retinopathy of 96.6% (95% confidence interval 95.4 to 97.4) compared with 62-82% for manual grading.28
Optical coherence tomography uses near infrared light to image cross sections of the retina (fig 6⇓). Optical coherence tomography is more sensitive than clinical examination or stereoscopic fundus photography for the detection of retinal thickening and clinically relevant macular oedema.29 However, it is uncertain whether treatment of subclinical diabetic macular oedema would improve the prognosis of patients with diabetic retinopathy.29 Therefore, although optical coherence tomography is useful for measuring the response to treatment of diabetic macular oedema, its role in the screening process has yet to be determined.
Fig 6 Optical coherence tomography (OCT) is an interferometric technique using near infrared light, which allows cross sectional images of the retina to be visualised. (A) OCT scan (right) showing the normal macular contour, with the central depression of the foveal avascular zone (the scanned area is shown in the photograph on the left). (B) OCT scan showing the thickened retina associated with diabetic macular oedema and the hyporeflective fluid filled cysts that it produces. (C) Vitreomacular traction can also be identified with OCT, and in this image the posterior vitreous face is attached at only one point. In some patients gross distortion of the foveal anatomy may be seen (D)
Although patients with diabetes are recommended to have a retinal examination at least every 12 months, many are either unable or unwilling to be examined. Non-attendance is associated with worse outcomes,30 but relatively little is known about the factors that influence patient compliance with screening recommendations. These factors will vary in different healthcare systems. For example, in a study of African-Americans with diabetes in New Orleans, cost was a major deterrent to attending an eye clinic,31 whereas in the United Kingdom patients and care providers seemed to underestimate the severity of retinopathy.32 In view of the major investment in screening and treatment programmes, identifying interventions to reduce non-attendance should be a research priority.
Current treatments for established retinopathy
Most patients with retinopathy do not need treatment and can be monitored safely by annual retinal examination with referral to an ophthalmologist only if signs of maculopathy or severe non-proliferative or proliferative diabetic retinopathy develop.
Laser photocoagulation
Laser photocoagulation is a well established treatment for diabetic retinopathy and has undergone only small modifications in the past 25 years. The main aim of this technique is to induce regression of the new blood vessels and to reduce central macular thickening and thus prevent visual loss from proliferative diabetic retinopathy and diabetic maculopathy, respectively.
Each laser burn causes the thermal destruction of retinal pigment epithelial cells and the adjacent photoreceptors. This reduces the production of vascular endothelial growth factor and lessens the stimulus for neovascularisation.33
This technique has been shown to be effective in large multicentre randomised trials. The Diabetic Retinopathy Study, a trial of laser treatment for proliferative diabetic retinopathy, found a 50% reduction in severe visual loss after scatter laser photocoagulation for new vessels at the optic disc.
The Early Treatment of Diabetic Retinopathy Study assessed the effect of laser photocoagulation versus observation for diabetic macular oedema. At three years the risk of moderate visual loss was reduced by 50% (from 24% to 12%) in the laser group (number needed to treat 8.3).34 Visual acuity improved in only 3% of patients. Despite current interventions, diabetic macular oedema still causes visual loss, and in some patients even the most aggressive treatment cannot prevent loss of vision.35 More treatment options are needed.
Intravitreal steroids
Intravitreal injection of corticosteroids has been used to treat inflammatory eye disease for many years, although the mechanism of action is not fully understood. A large randomised trial comparing intravitreal injection of 4 mg triamcinolone with standard laser photocoagulation found that steroids were initially more effective than laser but by two years eyes treated with laser had better visual acuity and less macular oedema. In addition, intravitreal steroid increased the risk of cataract formation and raised intraocular pressure.36
Anti-vascular endothelial growth factor treatments
Concentrations of vascular endothelial growth factor are raised in the vitreous of eyes with diabetic macular oedema.37 Anti-vascular endothelial growth factor drugs including pegaptanib (Macugen), bevacizumab (Avastin), and ranibizumab (Lucentis) may therefore be useful in the management of diabetic macular oedema.
In a recent large randomised trial of intravitreal ranibizumab versus standard laser treatment,38 patients treated with ranibizumab were more likely to gain at least 10 letters of visual acuity (48.8% v 27.6%; P<0.001) and less likely to lose 10 or more letters (3.2% v 13.3%; P<0.001) after one year of treatment. However, patients treated with ranibizumab received an average of eight to nine injections in the first year. Each injection costs £760 (€885; $1200) for the drug alone. In addition patients were seen every four weeks, which is unlikely to be feasible in clinical practice.
Bevacizumab has a similar mechanism of action to ranibizumab but is much cheaper. A smaller randomised trial (150 eyes; 129 patients) conducted in Tehran compared intravitreal bevacizumab injection alone, bevacizumab together with intravitreal triamcinolone, and macular laser alone in the treatment of diabetic macular oedema. Retreatment was performed at 12 week intervals when indicated. A two line improvement in visual acuity at 36 weeks was detected in 37%, 25%, and 14.8% of patients respectively.39
Vitrectomy
Surgical removal of the vitreous (vitrectomy) is important in the treatment of proliferative diabetic retinopathy. It aims to improve vision by removing any blood in or behind the vitreous, reattaching detached areas of retina, and reducing the stimulus for neovascularisation by complete pan-retinal laser photocoagulation. A recent large case series showed that sight threatening complications are rare, and in 90% of patients vision is improved or stabilised.40 Recurrent vitreous cavity haemorrhage is one of the most common complications; a small randomised trial found that pretreatment with intravitreal bevacizumab one week before vitrectomy makes surgery easier and reduces the risk of postoperative haemorrhage.41
Theoretically, vitrectomy should be beneficial in diabetic macular oedema that does not respond to laser treatment, but randomised trials have shown little effect.42 Vitreous traction (demonstrated by optical coherence tomography) may be a contributing factor in a minority of patients with macular oedema. In this situation vitrectomy may be beneficial.
The table[t1] provides a summary of randomised controlled trials of clinically relevant treatments for diabetic macular oedema.
Randomised controlled trials of clinically relevant interventions for diabetic macular oedema
Conclusions
Blindness from diabetic retinopathy is usually avoidable. The incidence of retinopathy can be reduced by good medical management. If retinopathy develops, sight loss should be preventable by early detection and laser photocoagulation. If vision is lost, at least some sight may be regained by vitrectomy or treatment with anti-vascular endothelial growth factor agents. Despite the availability of effective evidence based treatment, diabetic retinopathy still causes blindness in rich countries and increasingly in the developing world. Although we need to develop new treatments, our immediate priority should be to improve the delivery of interventions that have already been shown to be effective.
Tips for non-specialists
Diabetic retinopathy is still the leading cause of blindness in people of working age in the UK
The risk of diabetic retinopathy can be reduced by good control of hyperglycaemia and hypertension
Diabetic retinopathy is asymptomatic until it reaches an advanced and often untreatable stage but is easily detected by retinal examination
All patients with diabetes over the age of 11 should have a retinal examination at least every 12 months
Laser photocoagulation can prevent visual loss from diabetic macular oedema and proliferative retinopathy but is unlikely to restore vision that has been lost
New treatments are more likely to restore vision, but currently they are costly and difficult to deliver
Additional educational resources
Resources for patients
Royal National Institute of Blind People (www.rnib.org.uk)—Promotes awareness of the ocular complications of diabetes. The website is produced in conjunction with the Royal College of Ophthalmologists and offers a discussion forum for patients
Diabetes UK (www.diabetes.org.uk)—Charity that offers advice on all aspects of diabetes care including diabetic retinopathy
English National Screening Programme for Diabetic Retinopathy (www.retinalscreening.nhs.uk/); Scottish National Diabetic Retinopathy Screening Programme (www.ndrs.scot.nhs.uk/)—Information about screening protocols and what to expect
Resources for healthcare professionals
Royal College of Ophthalmologists (http://www.rcophth.ac.uk/about/publications/)—Website of the professionals body of eye doctors that includes advice about care for diabetic retinopathy
National Institute for Health and Clinical Excellence. Type 2 diabetes: national clinical guideline for the management of type 2 diabetes (update). 2008. www.nice.org.uk/nicemedia/live/11983/40803/40803.pdf
National Institute for Health and Clinical Excellence. Type 1 diabetes in adults: national clinical guideline for diagnosis and management in primary and secondary care. 2004. www.nice.org.uk/nicemedia/live/10944/29393/29393.pdf
Questions for future research
Can specific treatments prevent diabetic retinopathy?
Are there simpler and cheaper alternatives to monthly injections with ranibizumab for reversing visual loss in diabetic maculopathy?
Can the outcomes of vitrectomy for diabetic retinopathy be improved by pretreatment with anti-vascular endothelial growth factor?
What can be done to reduce non-attendance at diabetic retinopathy screening and treatment clinics, and how can patient education be improved?
Is optical coherence tomography a useful screening tool for macular oedema and, if so, what are the criteria for treatment or referral?
Notes
Cite this as: BMJ 2010;341:c5400
Footnotes
Competing interests: All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.
Provenance and peer review: Not commissioned; externally peer reviewed.