Diabetic retinopathy: management and monitoring

Prognostic factors for progression of non-proliferative diabetic retinopathy to proliferative diabetic retinopathy or diabetic macular oedema

Progression of non-proliferative diabetic retinopathy to diabetic macular ischaemia

There was no evidence on factors that can be used to predict how likely it is that non-proliferative diabetic retinopathy could progress to diabetic macular ischaemia. Therefore, the committee could not make recommendations on this and included progression to macular ischaemia in a recommendation for research on prognostic factors.

Non-proliferative diabetic retinopathy

Evidence was insufficient to determine which treatment strategies are the most effective to prevent progression to the sight-threatening complications of diabetic retinopathy (see evidence review E). Monitoring is generally used in current practice. The committee was therefore unable to make recommendations for this group. Instead, they made a recommendation for research on treatment strategies for severe non-proliferative diabetic retinopathy and another recommendation for research aimed at identifying the thresholds or criteria that should be used for starting treatment for non-proliferative diabetic retinopathy.

Proliferative diabetic retinopathy

A network meta-analysis (see evidence review E) showed that some anti-VEGF treatments resulted in slight improvements in visual acuity in comparison to panretinal photocoagulation. However, the committee noted that these differences were not clinically meaningful. Some individual studies showed that anti-VEGFs reduced the incidence of diabetic macular oedema, but there was no clear difference between anti-VEGFs and panretinal photocoagulation for any of the other outcomes. It was not possible to tell whether the effectiveness of different treatments changed depending on severity of retinopathy at baseline, because this was not clearly reported in the studies. In addition, many of the studies were low-quality and had small sample sizes, making it difficult to be certain of the effectiveness of each treatment option.

Given that panretinal photocoagulation and anti-VEGFs are of similar effectiveness, particularly for visual acuity, the committee used their clinical experience to recommend that panretinal photocoagulation should be used as first-line treatment when possible. This is because panretinal photocoagulation does not carry some of the risks that are associated with anti-VEGFs, such as endophthalmitis. It also requires fewer hospital visits and reduces the risk of progression that might otherwise be seen if a person is offered anti-VEGF treatment but cannot attend regular appointments. They agreed that panretinal photocoagulation is the standard of care for proliferative diabetic retinopathy.

The committee thought that panretinal photocoagulation was particularly effective for people with high-risk proliferative diabetic retinopathy. They agreed it can also benefit people with early proliferative retinopathy because, for these people, the alternative option is frequent monitoring. They also agreed that the risks associated with progression if people do not attend follow-up appointments are greater than the risk of adverse events from panretinal photocoagulation, particularly with modern panretinal photocoagulation. For this reason, they recommended that all people with proliferative diabetic retinopathy are offered panretinal photocoagulation when they are first diagnosed.

Evidence from 2 studies showed that early panretinal photocoagulation reduced the number of people who progressed or developed severe visual loss at 2 years. This supported the committee's experience that panretinal photocoagulation brings additional benefits if provided early.

How the recommendations might affect practice

The recommendations for people with proliferative diabetic retinopathy are in line with current practice and should not increase the number of people who are given panretinal photocoagulation. The recommendation to use anti-VEGFs if eyes do not respond to panretinal photocoagulation reinforces what should happen in current practice.

The recommendation for temporary anti-VEGF use for people who need vitrectomy or cataract surgery will reduce complications for the person as well as reducing the time and costs associated with additional treatment if their vitrectomy or cataract surgery is delayed.

The recommendations on when panretinal photocoagulation should be offered, and when treatment should start, are not expected to have a major impact on practice.

Return to recommendations 1.5.1 to 1.5.6

Return to recommendation 1.5.15

Effectiveness of different thresholds or criteria for starting treatment

The committee discussed the effectiveness of early macular laser compared with deferred macular laser for people with diabetic macular oedema. The evidence for this population was from 1 large study that showed that early laser slowed worsening of best-corrected visual acuity at 2- and 3-year follow-ups (see evidence review B). Eyes receiving early macular laser were also less likely to develop clinically significant macula oedema compared with eyes that received deferred treatment. The committee thought these improved outcomes were important and matched their clinical experience. They therefore used this information, combined with evidence of cost effectiveness from the treatment strategies review (see evidence review G), to recommend that all people with clinically significant diabetic macular oedema are offered treatment.

Good vision

Some people with centre-involving diabetic macular oedema have good vision. These people may have fewer benefits from anti-VEGF, steroids or macular laser treatment than people who have visual impairment, but still experience the adverse effects associated with treatment. However, the committee highlighted that:

Although the analysis for the whole population with diabetic macular oedema suggested that macular laser was not the most clinically effective treatment option (see evidence review G), it still showed benefits for improving visual acuity and was the most cost-effective option in comparison to anti-VEGFs or steroids. The committee thought it was important to highlight that macular laser can have benefits for people who still have good vision. They noted that macular laser is not always offered to this group of people even though it can delay progression to the point where a person needs anti-VEGF treatment, thereby benefitting the person and reducing treatment costs.

However, the committee was aware that macular laser may not be the only option for this group. Evidence from the review on thresholds for starting treatment (see evidence review B) showed that outcomes may be similar for some people whether they are initially offered observation, anti-VEGF or macular laser. The committee interpreted this to mean that it is safe to consider initially observing some people with centre-involving diabetic macular oedema and good vision. Therefore, as the benefits of treatment are likely to be smaller for people with good vision, and there is currently limited evidence comparing macular laser to observation (delayed treatment), the committee recommended clinicians should consider both options. They noted that the most appropriate option needs to be carefully considered for each person to reduce their risk of progression. Therefore, they recommended that the decision should include a discussion with the person about the benefits and risks of each option to make a shared decision over which to choose.

Impaired vision

Evidence from the network meta-analysis (see evidence review G) showed that in people with centre-involving diabetic macular oedema, a number of treatments, including anti-VEGFs, some steroids and some combinations of treatments, are more effective at improving visual acuity at 12 months than standard threshold laser alone. They are also more effective at reducing central retinal thickness at 12 and 24 months than standard threshold laser.

Subgroup analysis showed similar results for people with a central retinal thickness of 400 micrometres or more and some evidence of the benefits of anti-VEGFs in people with a central retinal thickness of less than 400 micrometres. However, the smaller evidence base for the latter group meant there was more uncertainty around the effects of different treatments.

Improvements in visual acuity and central retinal thickness, even at 12 months, are considered important by people with diabetic macular oedema. Although there was more limited data on effectiveness of treatments on visual acuity at 24 months, the committee was confident that the short-term results were enough to make recommendations on the most effective treatments for people with centre-involving macular oedema.

For people with non-proliferative and proliferative diabetic retinopathy

There was no evidence for people with non-proliferative or proliferative diabetic retinopathy and so the committee did not think they could make recommendations for this group.

For people with diabetic macular oedema

Evidence was available from 2 studies. Each study used different clinical indicators to determine if treatment should be switched, as well as using different types of treatment. Neither study showed a clear effect of switching treatments based on their switching criteria, so there was insufficient evidence to determine which clinical features best indicate the need to switch treatments for people with diabetic macular oedema.

There was no evidence of which clinical features might indicate the need to stop treatment, so the committee did not make recommendations on this.

The committee discussed how, ideally, there would be a list of biomarkers that can be used to define responsiveness to anti-VEGF therapy to help determine whether to continue, switch or stop treatment. Therefore, they made a recommendation for research on the effectiveness of clinical features or factors that suggest treatment for diabetic macular oedema should be switched or stopped.

Although the committee did not think they could recommend a specific switching criteria, they thought it important to highlight when a decision about switching or changing treatments should be made. If the decision to switch is made too soon, there may not be sufficient time for the treatment to show an effect. This may have been reflected in the evidence, where one of the studies used a 3-month loading phase.

The committee recommended the use of anti-VEGF treatment. However, they could not recommend a specific amount of time for the loading phase before assessing a response because different anti-VEGFs have different recommended loading phases. Instead, they advised that this should first be done after the loading phase of anti-VEGF treatment, and then 12 months after the start of treatment to assess for a delayed response.

Non-proliferative diabetic retinopathy

Evidence for monitoring frequencies for people with non-proliferative retinopathy showed that risk of progression between monitoring visits is higher for people with severe or very severe retinopathy compared with those with moderate retinopathy.

Based on this evidence and their clinical experience, the committee recommended different monitoring frequencies for people who are not currently having treatment and have not previously had treatment, depending on the severity of their disease.

People with moderate non-proliferative diabetic retinopathy have a relatively slow rate of progression and so monitoring every 6 to 12 months was considered appropriate. For people with severe or very severe non-proliferative diabetic retinopathy, whose disease progresses more quickly, monitoring every 3 months was considered beneficial. This will reduce the risk of progression to proliferative diabetic retinopathy or diabetic macular oedema remaining unnoticed for too long. This is important because, once the disease has progressed to proliferative diabetic retinopathy or diabetic macular oedema, it needs to be treated as soon as possible to avoid vision loss. However, the committee discussed how people with diabetic retinopathy often have to attend multiple appointments for other diabetes-related complications, so attending additional appointments every 3 months might not always be achievable. The committee therefore recommended that monitoring should take place every 3 to 6 months for this group. These recommendations reflect current practice.

Because there was limited evidence on the most effective monitoring frequencies for people with non-proliferative retinopathy who have not started treatment, the committee made a recommendation for research on monitoring frequencies for people with non-proliferative diabetic retinopathy.

Proliferative diabetic retinopathy or diabetic macular oedema

There was no evidence on monitoring frequencies for people with proliferative diabetic retinopathy or diabetic macular oedema who are receiving treatment or who have previously received treatment. The committee therefore made recommendations in this area based on their clinical experience, and in line with current practice.

They noted that monitoring during treatment with intravitreal therapies would be determined by the treatment protocol and so did not make recommendations for this area. However, they agreed that some guidance on monitoring frequency after treatment completion is required to improve consistency across the country. Therefore, they agreed that disease regression in people who have received treatment for proliferative diabetic retinopathy should be assessed at 2 to 3 months after treatment has ended. This should be an appropriate time so that any progression following the end of treatment can be identified before it leads to more serious consequences. They also recommended that, for the first 12 months after the end of treatment, monitoring frequency should be individualised depending on the treatment given and response to treatment. After 12 months, people who are eligible for the screening programme should be discharged to this service.

People who are discharged to the screening programme are expected to attend both eye screening appointments and regular appointments with primary care optometrists. This will help identify if further treatment is needed in the future and identify other eye diseases that are not covered by the eye screening programme.

Because there was limited evidence on the most effective monitoring frequencies for people with proliferative diabetic retinopathy or diabetic macular oedema who have had treatment previously, the committee made a recommendation for research on monitoring frequencies for people with proliferative diabetic retinopathy or diabetic macular oedema.

Under 18 or pregnant people with non-proliferative diabetic retinopathy, proliferative diabetic retinopathy or diabetic macular oedema

None of the evidence reported separate results for people under 18 or pregnant people. However, the committee agreed that the same recommendations should apply to under 18s as to adults. Although the risk of developing diabetic retinopathy is lower in under 18s, if it is identified, it should be monitored in the same way. The committee was aware of existing recommendations on monitoring diabetic retinopathy and the timing of retinal assessments in pregnancy in NICE's guideline on diabetes in pregnancy, so it agreed to refer to this guideline.

Ultrawide-field fundus imaging to detect proliferative diabetic retinopathy in people with non-proliferative diabetic retinopathy

There was no evidence on the diagnostic accuracy of ultrawide-field imaging for detecting proliferative diabetic retinopathy in people with non-proliferative diabetic retinopathy. A range of tests can be used in clinical practice and the committee did not think they could tell which is the most effective without evidence. Therefore, they made a recommendation for research on diagnostic test accuracy for monitoring disease progression to provide evidence on this in future.

Ultrawide-field fundus imaging to detect proliferative diabetic retinopathy in people with previously treated diabetic retinopathy

Evidence was available from a single study which assessed the diagnostic accuracy of ultrawide-field fundus imaging for people who had previously had treatment for proliferative diabetic retinopathy. The committee thought that the sensitivity of ultrawide-field imaging was sufficient to consider it as an additional test alongside other tests used to diagnose proliferative diabetic retinopathy. They also highlighted that it is quicker than other imaging tools.

The committee discussed whether ultrawide-field imaging could be used as the sole diagnostic test for diabetic retinopathy. However, they were concerned about the potential for this form of imaging to miss some important indications such as rubeosis. These other indications can be picked up by current standard techniques, such as slit-lamp biomicroscopy. For this reason, they decided to recommend that ultrawide-field imaging should be used alongside other forms of clinical examination to detect proliferative diabetic retinopathy. This could be during face-to-face appointments with a clinician, or at virtual clinics, where imaging takes place and is reviewed later by the clinician. Using ultrawide-field imaging at a virtual clinic would not result in it being a stand-alone test, as anyone with eyes showing signs of progression would then see an ophthalmologist for further assessment and to make a decision about whether treatment is needed.

The committee noted that using more than 1 technique was beneficial not only for diagnosing proliferative retinopathy, but also in other ways. While ultrawide-field imaging can be efficient, it is often carried out in diagnostic testing centres. This means that people miss out on the interaction with healthcare professionals who can answer questions and reduce any anxiety that people may have about their test results. This supported the committee's decision to recommend ultrawide-field imaging alongside other techniques.

Optical coherence tomography for the detection of diabetic macular oedema

Evidence was available from a high-quality systematic review that compared the diagnostic accuracy of optical coherence tomography (OCT) to that of fundus examination or photography. This showed OCT was effective to diagnose diabetic macular oedema development or progression.

Although the review showed that OCT can result in some false positives, the committee thought this was a result of the ability of OCT to detect subclinical macular oedema. OCT is therefore a useful test to identify people whose disease needs to be monitored until it reaches a threshold where treatment may be needed, as well as identifying people who already have diabetic macular oedema. The committee also discussed how OCT scans play an important role in monitoring treatment response to anti-VEGF treatment. Therefore, the committee decided that OCT should be recommended as the primary diagnostic method for diabetic macular oedema. They highlighted that this reflects current practice.

Vitrectomy in combination with other treatment strategies

The committee reviewed evidence on the effectiveness of vitrectomy alone or in combination with other treatments for proliferative diabetic retinopathy and macular oedema.

Evidence for people with proliferative retinopathy or macular oedema did not clearly show that any of the adjuvant treatment regimens to a vitrectomy can improve outcomes following treatment (see evidence review F). However, the trials that were reviewed were small and the inclusion criteria varied, which was not helpful in decision making. With no clear evidence, the committee did not make any recommendations on treatment combinations.

Proliferative diabetic retinopathy

The evidence did not show that vitrectomy was more beneficial than other interventions. However, the committee thought this was because of limitations in the evidence base, such as mixed populations. This made it difficult to draw conclusions about the benefits of vitrectomy for groups of people with different complications.

Diabetic macular oedema

The committee agreed that there was no evidence to support the use of vitrectomy to treat diabetic macular oedema (see evidence review F).

However, for people with diabetic macular oedema that does not respond to anti-VEGF treatment and evidence of vitreoretinal traction or epiretinal membrane, vitrectomy should be considered (see evidence review B). The committee highlighted that, without vitrectomy, these people are at risk of permanent damage to the eye. With no evidence on timing of vitrectomy for this group, the committee did not think they could specify when this should be done. However, they said this should be done early enough after a person's condition shows no response to anti-VEGF treatment to ensure that the eye does not incur any permanent damage. Although there is limited evidence for people with vitreoretinal traction or epiretinal membrane secondary to diabetic macular oedema, the committee did not make a recommendation for research on this topic because the small number of people in the group has made it hard to meet targets for trial recruitment in the past (see evidence review F).