Blindness is one of the most feared complications of diabetes with an incidence of 50-65 per 100,000 diabetic population per year in Europe.^{245, 246, 247} However, with good care, visual impairment due to diabetes can be avoided for the vast majority of patients.

6.1 Risk identification and prevention

6.1.1 RISK FACTORS FOR DIABETIC RETINAL DISEASE

The following risk factors have been shown to determine the development and progression of diabetic retinal disease:

• poor glycaemic control^{18, 248, 249}
• raised blood pressure ²³²
• increasing number of microaneurysms^{250, 251}
• duration of diabetes^{252, 253}
• microalbuminuria and proteinuria^{254, 255}
• raised triglycerides and lowered haematocrit²⁵⁶
• pregnancy.²⁵⁷
  Evidence level 1++, 2+

The evidence with regard to smoking as a risk factor for diabetic retinal disease is conflicting. Tha available evidence suggests that smoking may be a risk factor for retinopathy in type 1 diabetes;^{258, 259} however, in type 2 diabetes, the evidence is controversial, and smoking may protect against the progression of retinopathy in certain patients.^{260, 261} Smoking is an independent risk factor for cardiovascular disease in all patients with diabetes and should therefore be discouraged (see sections 3.4 and 4.2).

Patients with multiple risk factors should be considered at high risk of developing diabetic retinal disease.

See section 8 for specific guidance on assessment and referral during pregnancy.

Diabetic retinal disease is the commonest cause of visual impairment in type 1 diabetes, but not in type 2 diabetes.²⁶² Patients with diabetes have approximately a twofold increased risk of cataract^{263, 264} and the risk is increased with poor glycaemic control.²⁶⁵ One study has indicated that intensive glycaemic control reduced the incidence of cataract extraction in type 2 diabetes. Evidence level 2++

6.1.2 RISK FACTOR MODIFICATION

The evidence that modifying risk factors has a beneficial outcome in diabetic retinal disease exists for only some of the risk facyors identified above. Evidence level 1++

Tight control of blood glucose reduces the risk of onset and progression of diabetic eye disease in type 1 and 2 diabetes.^{231, 248, 266} Evidence level 1++

Reducing HbA1c by 1.5% and, if possible, to 7% in type 1 and 2 diabetes^{231, 248} and reducing blood pressure to 144/82 mm Hg in type 2 diabetes reduces the incidence and progression of sight-threatening diabetic eye disease²³² and this is likely also to be the case for type 1 diabetes. Evidence level 1++

Reducing blood pressure and HbA1c below these targets is likely to reduce the risk of eye disease further.^{171, 172} Microvascular endpoints (including retinopathy) are decreased by 37% with each 1% reduction in HbA1c, and by 13% for each 10 mm Hg reduction in systolic blood pressure^{172, 174} indicating that any improvement in these parameters is beneficial. Evidence level 2++

Good glycaemic control (HbA1c ideally around 7%) and blood pressure control (< 140/80 mm Hg) should be maintained to prevent onset and progression of diabetic eye disease.

Rapid improvement of glycaemic control can result in short term worsening of diabetic retinal disease although the long term outcomes remain beneficial.²⁴⁸ Evidence level 2++

Sight-threatening retinal disease, if present, should be stabilized before rapid clinical improvements in glycaemic control are achieved.

There is an absence of good evidence for any additional benefit of ACE inhibitors in diabetic eye disease. One recent multicentered RCT which addressed this issue was methodologically flawed,²⁶⁷ however there are a number of ongoing trials involving ACE inhibitor therapy.

6.2 Screening

6.2.1 WHO SHOULD BE SCREENED ?

Up to 39% of patients with type 2 diabetes have retinopathy at diagnosis, with 4-8% being sight-threatening.^{231, 268} Evidence level 1++

Screening for diabetic retinal disease is effective at detecting unrecognised sight-threatening retinopathy^{269, 270} Evidence level 2++, 4

In type 1 diabetes, pre-proliferative retinopathy has been identified 3.5 years after diagnosis in post-puberty patients²⁷¹ and within two months of onset of puberty. Evidence level 2+

For patients with no retinopathy at baseline, the chance of developing sight-threatening retinopathy within two years is less than 1% in both type 1 and type 2 diabetes on preliminary data.^{272, 273} Patients with existing diabetic retinal disease may require more frequent retinal examination. Evidence level 2++

Systematic annual screening for diabetic retinal disease should be provided for all people with diabetes.

Patients with type 2 diabetes should be screened from diagnosis.

Patients with type 1 diabetes should be screened from age 12 years. If onset of type 1 diabetes is post-puberty, screening should start three years after diagnosis.

6.2.2 HOW SHOULD SCREENING BE PERFORMED?

Diabetes UK propose that an effective system of screening should achieve a sensitivity of 80% and specificity of 95% with a technical failure rate of less than 5%.²⁷⁴ Visual acuity measurements help in interpretation of maculopathy.²⁷⁵ Evidence level 4

Retinal photography can frequently achieve a sensitivity of 80% and is a more effective screening method than direct ophthalmoscopy, which only rarely achieves 80% sensitivity even when carried out by properly trained operators.²⁷⁰ Evidence level 2++

Between 3% and 14% of retinal photographs are ungradeable^{269, 276} although this rate may be improved by digital imaging. Slit lamp biomicroscopes with dilated indirect ophthalmoscopy used by properly trained individuals can achieve sensitivities similar to,²⁷⁰ or greater than,²⁷⁷ retinal photography, with a lower technical failure rate. However, slit lamp biomicroscopy has only limited validation as a screening tool.²⁷⁸ Evidence level 2++

Patients prefer screening to be performed at a site convenient to them.^{279, 280} Evidence level 4

Retinal photography or slit lamp biomicroscopy used by trained individuals should be used in a programme of systematic screening for diabetic retinopathy.

Dilated direct ophthalmoscopy should only be used for opportunistic screening.

Screening modalities should aim to detect sight threatening retinal disease with a sensitivity >= 80% and specificity >= 95%.

Patients with ungradeable retinal photographs should receive slit lamp and indirect ophthalmoscopy examination where possible.

Where possible and practical, screening should be performed at a site convenient to patients.

6.2.3 GRADING AND QUALITY ASSURANCE

When grading retinal appearances, digital imaging is more sensitive than polaroid prints and probably similar to 35 mm film.²⁸¹ Initial data indicates that high-resolution automated techniques can identify the absence of microaneurysms on digital images with a sensitivity of 85%, although further research is required in this area to validate the technique.²⁸² Evidence level 2+

All screening modalities should undergo quality assurance checks. For retinal photography it has been suggested that this should happen in 1% of photographs.²⁶⁹ Evidence level 4

Retinal photographs should be graded using digital images or 35 mm film by an appropriately trained grader.

At least 1% of all screening events (photography or slit lamp) should be reviewed.

The Health Technology Board for Scotland (HTBS) is carrying out a health technology assessment to determine the most efficient, effective and comprehensive national screening programme for diabetic retinopathy in Scotland.

6.3 Treatment

6.3.1 LASER PHOTOCOAGULATION

Laser photocoagulation for high risk retinopathy and clinically significant macular oedema (CSMO) is of proven benefit.^{283, 284} Older patients with diabetes and those with type 2 diabetes in particular benefit from photocoagulation before high risk features develop.²⁸⁵ Laser treatment for CSMO helps to stabilise vision when used for focal or diffuse maculopathy, but was not shown to be helpful for ischaemic maculopathy.^{284, 286, 287} It has also been shown to be effective for iris neovascularisation (rubeosis) due to microvascular disease.²⁸⁸ Evidence level 1++, 3

There are no clinical trial data assessing the strategy of whether treatment should be deferred in diffuse maculopathy until visual acuity is affected. There is no evidence for the use of laser in ischaemic maculopathy.

All patients with sight-threatening retinopathy (moderate proliferative diabetic retinopathy or worse) should receive laser photocoagulation.

Patients with severe pre-proliferative or mild proliferative diabetic retinopathy should receive close follow up or laser photocoagulation.

Focal or modified grid laser photocoagulation should be used for patients with focal CSMO but not for patients with ischaemic maculopathy.

Diffuse maculopathy should be treated if there is a concern that the disease is progressing.

6.3.2 VITRECTOMY

Early vitrectomy is of proven value in patients with type 1 diabetes and persistent vitreous haemorrhage for improving long term vision. Its value in type 2 diabetes is less certain. Patients with type 1 or type 2 diabetes who have severe fibrovascular proliferation with or without retinal detachment threatening the macula also have better visual acuity after vitrectomy.²⁸⁹ Evidence level 1+

Vitrectomy for diffuse diabetic macular oedema has been shown to result in resolution of oedema and improvement in visual acuity.^{290, 291, 292} Evidence level 3

Patients with type 1 diabetes and persistent vitreous haemorrhage should be referred for early vitrectomy.

Vitrectomy should be performed for tractional retinal detachment threatening the macula and should be considered for severe fibrovascular proliferation.

Vitrectomy should be considered in patients with diffuse diabetic macular oedema.

Patients with type 2 diabetes and vitreous haemorrhage which is too severe to allow photocoagulation should be referred for consideration of a vitrectomy.

6.3.3 REFERRAL INTERVALS

Delay in treatment of greater than two years from diagnosis of sight-threatening diabetic retinopathy is associated with poor outcome and severe visual loss.²⁸³ When vitrectomy is required, a delay of over one year is associated with poorer visual outcome.²⁹³ Evidence level 1++

The Royal College of Ophthalmologists²⁷⁵ recommend the following referral intervals:

Assess by ophthalmologist within 4 weeks if:

• there is unexplained drop in visual acuity
• there are hard exudates within 1 disc diameter of the fovea
• macular oedema is present
• there are unexplained retinal findings
• pre-proliferative or more advanced (severe) retinopathy is present.
  Evidence level 4

Assess by ophthalmologist within 1 week if:

• there is new vessel formation
• there is evidence of pre-retinal and/or vitreous haemorrhage
• rubeosis iridis is present.
  Evidence level 4

Assess by ophthalmologist within 1 day if:

• there is sudden loss of vision
• there is evidence of retinal detachment.
  Evidence level 4

6.3.4 CATARACT EXTRACTION IN PATIENTS WITH DIABETES

Visual outcome following cataract surgery in patients with diabetes is closely linked to age and severity of retinopathy present before surgery.^{294, 295} Whilst postoperative progression of pre-existing proliferative diabetic retinopathy and CSMO has been documented, the balance of evidence does not show an increase in long-term incidence of CSMO following cataract extraction.^{294, 295, 296} Evidence level 1++, 2++

Cataract extraction should not be delayed in patients with diabetes.

Cataract extraction is advised when sight-threatening retinopathy cannot be excluded.

When cataract extraction is planned in the context of advanced disease which is not stabilised prior to surgery, the risk of progression and the need for close postoperative review should be fully discussed with the patient.

6.3.5 METHOD OF ASSESSING RETINOPATHY

Slit lamp biomicroscopy carried out by an appropriately experienced ophthalmologist is as good as the gold standard of 7-field stereoscopic photography for the assessment of CSMO.^{277, 284, 286} Evidence level 1+

Either good quality 7-field stereo photography or slit lamp biomicroscopy (both dilated) carried out by an appropriately experienced ophthalmologist should be used to investigate:

CSMO

proliferative diabetic retinopathy and severe non-proliferative diabetic retinopathy.

6.4 Rehabilitation

There is very little evidence relating to programmes of rehabilitation for patients with diabetic eye disease. Awareness of low vision aids is poor, but once available, patients benefit from being instructed in their use.

Low vision aid clinics²⁹⁷and community self-help groups^{298, 299} as part of a low vision service^{300, 301} can improve the quality of life and functional ability for patients with visual impairment. Evidence level 3, 4

Community support, low vision aids and training in their use should be provided to people with diabetes and visual impairment.

Patients with visual impairment should be assisted to register as blind / partially sighted at an early stage.