DIABETIC EYE DISEASE

by Dr Robert Gregory BA, DM, FRCP, Consultant Physician, Leicester General Hospital

Most patients with type 1 diabetes eventually develop background retinopathy, but it does not mean that patients will inevitably lose their sight. Adequate screening is the first step in ensuring that this complication of diabetes is prevented.

The complication of diabetes most feared by patients and their families is blindness, caused by diabetic retinopathy. However, appropriate screening, and medical and surgical management can prevent blindness as a complication of diabetes in most patients.

Prevalence of diabetic retinopathy

Diabetic retinopathy is thought to be the most common cause of blindness in people of working age in the UK. It is responsible for 12% of new registrations for blindness and 8% for partial sightedness.

One of the targets of the 1989 St Vincent Declaration, issued by the WHO and the International Diabetes Federation, was to reduce new cases of blindness caused by diabetes by at least one third within 5 years.(1) However, it is difficult to measure the baseline incidence and prevalence of diabetic blindness; blind registration is not mandatory and not all blindness associated with diabetes is necessarily caused by diabetes.

In a survey carried out in Wisconsin, USA, 1,000 patients with insulin-dependent (type 1) diabetes diagnosed before the age of 30 years were studied using standardised fundus photography. This showed retinopathy in 17% of patients within 5 years of diagnosis, rising to 97.5% in those diagnosed more than 15 years previously.(2) However, the prevalence of sight-threatening proliferative retinopathy was 4% at 10 years, 25% at 15 years and 67% at 35 years after diagnosis. This means that although nearly all patients with type 1 diabetes can be expected to develop background retinopathy, it is not inevitable that it will progress to the point where it poses a threat to vision.

The Wisconsin study also looked at people with diabetes diagnosed after 30 years of age, with noninsulin-dependent (type 2) diabetes. Over 20% had retinopathy within 2 years of diagnosis, and some presented with retinopathy; 4% of patients had proliferative retinopathy within 4 years of diagnosis.

Data from the UK Prospective Diabetes Study (UKPDS) show that at least 18% of newly diagnosed patients with type 2 diabetes have retinopathy at diagnosis.

Factors determining development of retinopathy
The landmark Diabetes Control and Complications Trial (DCCT) proved that for type 1 diabetes the main factor determining the development and progression of retinopathy is the control of blood glucose concentration over time.(3) Indeed, the study was terminated early on ethical grounds because the patients assigned to receive intensive management were 50% less likely either to develop retinopathy in the first place or for their pre-existing retinopathy to deteriorate than patients in the conventional management group. Similar findings are reported for type 2 diabetes in the UKPDS.(4)

The observation that not all patients with 'chronically poor' diabetes control developed clinically significant diabetic retinopathy led to attempts to find markers of susceptibility to, or protection from, complications. Unfortunately, no clinically useful markers have so far emerged. This means that clinicians must strive for blood glucose concentrations as near as practicably possible to normal for all their patients.

Hypertension increases the risk of developing retinopathy and the risk of haemorrhage from new vessels. Because retinopathy develops in association with other microvascular complications of diabetes, including nephropathy, it is not surprising that many patients with significant retinopathy also have proteinuria and hypertension. Hypertension should therefore be treated aggressively in parallel with improving diabetic control.

How does diabetic retinopathy develop?

Even before the earliest background changes are visible, significant changes to the structure of the retinal microcirculation occur. These result in increased retinal blood flow and increased capillary permeability. Eventually, extravascular deposits of plasma lipoproteins become visible as hard exudates and in some cases cause macular oedema. The increase in retinal blood flow imposes shearing forces on the endothelium which damage it further, resulting in microaneurysm formation and intraretinal haemorrhages (dots and blots). This process is exacerbated by systemic hypertension.

The damaged capillaries may become occluded with platelet or leucocyte plugs, producing retinal ischaemia, and even infarction. This stimulates attempts by the venules draining the ischaemic area to revascularise it. Preproliferative changes, with venous beading, looping, and intraretinal microvascular abnormalities (IRMAs) are seen, ultimately resulting in the formation of new vessels.

Retinal new vessels grow forwards from the retina, into the preretinal space and on to the posterior surface of the vitreous body. If the vitreous detaches itself from the retina, these fragile new vessels are torn, resulting in either a preretinal or vitreous haemorrhage.

New vessels are supported by a connective tissue stalk which can fibrose and contract, eventually causing a traction retinal detachment. New vessels can form at the optic disc or elsewhere in the retina. In the most severe cases of retinal ischaemia, new vessels can form on the iris (rubeosis iridis), which can obstruct the aqueous outflow causing acute neovascular glaucoma.

Screening for diabetic retinopathy

Serious and sight-threatening diabetic retinopathy is often asymptomatic. Indeed, often the first warning that a patient has retinopathy is when they have a vitreous haemorrhage. The occult nature of the condition, together with the natural history and epidemiological evidence, make an overwhelming case for screening all diabetic patients for retinopathy. A patient who loses his or her sight as a result of diabetic retinopathy could successfully sue their doctor for failing to screen them adequately.

The epidemiological differences between the retinopathy of type 1 and type 2 diabetes mean different screening policies should be used. There is probably no need to screen a young patient with type 1 diabetes until 5 years after diagnosis; however, it is essential to screen a patient with newly diagnosed type 2 diabetes as soon as possible. In reality, most doctors examine all their diabetic patients about once a year. The examination provides an opportunity to reinforce the educational message about the need for good diabetic control, particularly if a glycosylated haemoglobin result is available. Patients with no retinopathy or mild background changes, should be re-examined every 12-16 months. Patients with background changes should be encouraged to achieve the best possible diabetic control in line with the evidence from the DCCT and UKPDS. Patients with worsening background retinopathy should be examined every 6 months.

Who should screen?

There has been fierce controversy in recent years over who should perform screening for diabetic retinopathy. In the UK, there too few ophthalmologists to provide this service, and they are best employed treating retinopathy once it has been detected. Apart from those patients with type 2 diabetes who have retinopathy at diagnosis, most patients who lose vision as a result of diabetes have not had regular eye examinations.

It is undesirable to rely on a single form of screening. I encourage patients to visit their optometrist for an annual NHS sight check (including measurement of intraocular pressures and refraction) as well as their GP for fundal examination. Not all GPs who provide diabetes care are sufficiently confident in their fundoscopy skills to screen their own patients. PCGs should be guided by these principles in adopting a preferred screening scheme - either a retinal photography service or a service provided by specially accredited optometrists.

There is no single best option and a decision on which service to offer must take into account the resources available in a particular locality. For example, a rural population with poor public transport might be better served by a mobile retinal camera visiting the GP surgery annually, while an urban population might find a scheme based at the local hospital more accessible. Whichever scheme is chosen, it should be subject to regular audit and quality assurance.

The British Diabetic Association has produced guidelines to help those districts contemplating setting up a screening programme.(5,6) Leicester General Hospital offers a retinal photography service to local GPs. This provides 35 mm transparencies and a report by a consultant diabetologist which includes an estimate of the risk of development of, or deterioration of retinopathy based on the patient's glycosylated haemoglobin result at the time of screening. A copy of this report is also sent to the patient. The diabetologist refers patients to an ophthalmologist if necessary, sending a copy of the transparencies with the referral letter.

How to screen

The screening procedure is summarised in Box 2.(7) The patient should be warned to expect an eye examination when the appointment is made; it is a waste of everyone's time if the patient arrives unprepared to have his or her pupils dilated.

Symptoms: Although symptoms occur late in the natural history of diabetic retinopathy it is important to ask the patient whether he or she has experienced symptoms such as visual blurring, floaters or pain in the eye.

Visual acuity: Testing visual acuity is an essential part of the examination. A Snellen or EDTRS chart with adequate illumination should be used - an illuminated box is best. An E-chart is useful for patients who do not read English. The patient should read down the lines with each eye in turn, either unaided or using distance spectacles if they usually wear them. If the visual acuity is worse than 6/6 in either eye, repeat the examination with the patient looking through a pinhole which corrects for refraction. Improvement when using a pinhole suggests the possibility of a cataract, or vitreous opacity. Lack of improvement suggests retinal or macular disease. Always compare with the previously recorded visual acuity, because any deterioration may indicate serious retinal pathology.

Fundoscopy: This should always be performed through dilated pupils in a darkened room. In most patients, 0.5% or 1% tropicamide (Minims tropicamide, Mydriacyl) is the mydriatic of choice, because of its short duration of action (3 hours). However, patients with brown eyes may need the addition of 2.5% phenylephrine (Minims phenylephrine) to achieve adequate dilation.

Before instilling the drops it is important to ask about previous eye surgery. The earliest intraocular lens implants were fixed to the iris and could be dislocated by mydriasis, but intracapsular implants are not affected. Mydriatics also affect accommodation, and patients should be warned not to drive after the examination until their vision has returned to normal. The risk of precipitating acute angle closure glaucoma by pupil dilation is small, and measurement of intraocular pressure and the depth of the anterior chamber before instillation of drops is not necessary. This complication is even rarer now that dilation is no longer reversed with miotic drops after the examination. However, all patients should be advised to attend the nearest eye casualty department immediately if they develop an acutely painful eye; early treatment is curative.

The patient's eye should first be examined, using a direct ophthalmoscope, for a cataract. Look for the red reflex from a distance of 30 cm. Then, with a +10 dioptre lens, move close to the eye and examine the cornea and lens for opacities. Adjusting the lens selected towards '0' will bring the more posterior structures into focus, revealing any vitreous opacities and, ultimately, the retina. If there is a cataract of sufficient density to obscure the retina, the patient should be referred to an ophthalmologist.

A set pattern should be followed for viewing the retina. A limitation of the direct ophthalmoscope is its narrow field of view, so care must be taken not to miss any areas. I recommend starting with the optic disc and looking for new vessels there. Next, follow the retinal veins from the disc in the four quadrants. Start with the nasal retina which can be examined without crossing the visual axis. Follow the veins as far as possible, noting any signs of retinopathy and looking particularly for peripheral new vessels which tend to arise at the branch points of retinal veins. Ask the patient to look at the light in order to examine the macula. The temporal retina can be examined by asking the patient to abduct the eye: "Look over my shoulder". Direct ophthalmoscopy is an insensitive method for detecting macular oedema and retinal detachment, and although indirect ophthalmoscopy is more suited to this purpose it is not widely used outside ophthalmology departments. Patients with background retinopathy and deterioration in visual acuity should be referred for an ophthalmological assessment so that significant macular oedema is not missed.

Surgical management

The indications for referral are derived from the results of several large prospective studies which have defined features associated with a high risk of visual loss. These studies have also demonstrated the effectiveness of laser photocoagulation therapy in proliferative retinopathy.(8,9) Panretinal photocoagulation was found to reduce the risk of severe visual loss by more than 50% compared with no treatment; while panretinal photocoagulation of preproliferative retinopathy was more effective at preventing severe visual loss compared with waiting until proliferative retinopathy had developed.

Laser photocoagulation in proliferative and preproliferative retinopathy works by destroying ischaemic areas of the peripheral retina, thereby reducing the local stimulus to new vessel production. Existing new vessels usually regress; however, this treatment is not always curative. Moreover, panretinal photocoagulation can cause visual field restriction and night blindness, which may affect a patient's ability to drive.

Focal or grid photocoagulation treatment of focal or diffuse macular oedema, respectively, halves the risk of subsequent visual loss by sealing leaking vessels and facilitating resorption of exudated material.

Vitrectomy is an accepted treatment for chronic severe vitreous haemorrhage with visual loss and for early retinal detachment.

Cataract in diabetic patients

Cataracts are common in diabetic patients due to the nonenzymatic glycation of lens proteins which alters their refractive properties. There are two reasons for removing cataracts in diabetic patients: to improve vision and to permit visualisation of the fundus to assess retinopathy. Unfortunately, the postoperative complication rate is higher than in nondiabetic cataract extractions, particularly when there is severe retinopathy.

Conclusion

Visual loss caused by diabetic retinopathy is usually preventable if the retinopathy is detected early and medical measures to improve diabetic control, with laser treatment where indicated, are implemented.

GPs have a responsibility to organise the screening of patients under their care, either by using an approved method themselves, or through a local screening programme. Failure to do so is medicolegally indefensible.

Achieving the St Vincent Declaration target is only the first step in the complete prevention of new blindness due to diabetes in the UK.

Box 1. Protocol for screening for diabetic retinopathy.

• Make the appointment and warn the patient not to drive afterwards
• Ask about visual symptoms and check contraindications to mydriatic drops
• Check visual acuities in each eye (use pinhole if necessary and highlight significant deterioration)
• Dilate pupils (1% tropicamide). Caution about driving afterwards and advise about acute painful eye
• Fundoscopy in darkened room
  - Examine for cataract
  - Examine retina
• Explain your findings to the patient
  - Is diabetic control adequate?
• Make next appointment or refer to ophthalmologist if necessary

Box 2. Indications for referring a diabetic patient to an ophthalmologist

Immediate referral (24 hours)

• Vitreous haemorrhage
• Retinal detachment
• Urgent referral (1 week)
• Visual deterioration (2 lines or more)
• New vessels on the disc
• New vessels elsewhere
• Rubeosis iridis
• Advanced diabetic eye disease
• Neovascular glaucoma

Referral soon (few weeks)

• Haemorrhages and/or hard exudates within 1 disc diameter of the macula
• Macular oedema
• Preproliferative retinopathy

Routine referral (few months)

• Cataract, especially if adequate views of the retina cannot be obtained
• Retinal findings of uncertain significance

Practical points

• Diabetic retinopathy is thought to be the most common cause of blindness in people of working age in the UK.
• The main factor determining the development and progression of retinopathy is the control of blood glucose concentration over time.
• Serious and sight-threatening diabetic retinopathy is frequently asymptomatic; often the first warning that a patient has retinopathy is when they have a vitreous haemorrhage.
• part from those patients with type 2 diabetes who have retinopathy at diagnosis, most patients who lose vision as a result of diabetes have not had their eyes examined regularly.
• Screening for diabetic retinopathy can be carried out by the GP, by fundoscopy or, as in many districts, by a preferred screening scheme. This is either retinal photography or by specially accredited optometrists.

References

1. World Health Organization (Europe) and International Diabetes Federation (Europe). Diabetes care and research in Europe: the St Vincent Declaration. Diabetic Med 1990; 13(Suppl 4): 6-12.
2. Klein R, Klein BE, Moss SE. The Wisconsin epidemiological study of diabetic retinopathy: a review. Diab Metab Rev 1989; 5: 559-70.
3. Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993; 329: 977-86.
4. United Kingdom Prospective Diabetes Study Group. UK prospective diabetes study 33: intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes. Lancet 1998; 35: 837-53.
5. British Diabetic Association. Retinal photography screening for diabetic eye disease. London: BDA 1997.
6. British Diabetic Association. Optometry screening programme for diabetic eye disease. London: BDA 1997.
7. Retinopathy Working Party. A protocol for screening for diabetic retinopathy in Europe. Diab Med 1991; 8: 263-7.
8. Early Treatment Diabetic Retinopathy Study Research Group. Early photocoagulation for diabetic retinopathy. ETDRS Report No. 9. Ophthalmology 1991; 98: 766-85.
9. British Multicentre Study Group. Photocoagulation for proliferative diabetic retinopathy: a randomised clinical controlled trial using the xenon arc. Diabetologia 1984; 26: 109-15.