Treatment of diabetic retinopathy by pan-retinal laser photocoagulation (PRP) is usually effective in preserving visual acuity, but is not usually applied till the high-risk stage of proliferative diabetic retinopathy (PDR) has been reached. A review of a selection of clinical guidelines showed that treatment at the non-proliferative DR (NPDR) stage is currently either not recommended, or recommended only in certain circumstances. The UK Health Technology Assessment Programme has however recently published a systematic review and economic evaluation of pan-retinal photocoagulation at the severe NPDR stage.
This review examines the clinical effectiveness of PRP for severe NPDR, but also looks at two other aspects. First, there have been developments in methods of laser photocoagulation. So, if the evidence supported PRP at the NPDR stage, one question would be which form of laser treatment would be used. Second, there have been advances in drug treatment of diabetic macular oedema, with the arrival of the anti-VEGF drugs such as ranibizumab, bevacizumab and aflibercept, which are being used in combination with laser treatment. The review is free to download to download from: http://dx.doi.org/10.3310/hta19510 (Royle et al 2015), but is written in the context of the UK health system.
The questions addressed in this review were;
- Would it be worthwhile to intervene with pan-retinal laser photocoagulation earlier in diabetic retinopathy, at the severe NPDR stage rather than wait till the high-risk PDR stage? Treating at early PDR stage would be another option
- If so, what form of laser treatment should be used?
- Are drug-laser combinations clinically effective and cost-effective?
The review did not consider the effectiveness of laser treatment of diabetic macular oedema which is performed with focal or grid laser. Note that PRP is sometimes called “scatter photocoagulation”.
Evidence on the timing of PRP came almost entirely from the Early Treatment Diabetic Retinopathy Study (ETDRS). This was a large high quality study that recruited patients with moderate to severe NPDR or early PDR, with or without macular oedema, in the years 1980 to 1985. Patients were randomised to immediate PRP (“early photocoagulation”) or to observation and PRP at the high-risk PDR stage (“deferred photocoagulation”). Those with no macular oedema were further randomised to different intensities of PRP, known as full or mild scatter. Those with macular oedema randomised to early photocoagulation were further randomised to either full or mild scatter, and to early or delayed focal laser treatment for the DMO.
There were three groups of eyes in ETDRS;
- Category 1 moderate to severe NPDR or early PDR but no macular oedema
- Category 2 mild to moderate NPDR (“less severe retinopathy”) and macular oedema
- Category 3 severe NPDR or early PDR (“more severe retinopathy”) and macular oedema
The primary endpoint of the ETDRS was the development of severe visual loss (SVL). The absolute risks of SVL in the trial were low – 2.6% with early laser and 3.7% with deferred PRP. The 5 year relative risk of SVL for eyes assigned to early photocoagulation compared with deferral was 0.77 (99% CI 0.56 to 1.06). So early photocoagulation reduces the risk of SVL by about 23%, though the 99% CI interval overlapped with no difference. (The ETDRS trial used more rigorous 99% CIs rather than the usual 95%).
The relative risks for the three categories differed;
- Category 1 = 1.37 (99% CI 0.67 to 2.77)
- Category 2 = 0.59 (99% CI 0.32 to 1.09)
- Category 3 = 0.70 (99% CI 0.44 to 1.11)
Compared with deferral of photocoagulation, early photocoagulation reduced progression to high-risk PDR in two of the baseline categories though this was not statistically significant using 99% CIs. Strategies that used full scatter reduced progression to high-risk PDR by 50% and mild scatter by 25% compared to the deferred groups.
By 5 years, 3.9% in the deferred group and 2.2% in the early group had had vitrectomy. Most patients undergoing vitrectomy had type 1 diabetes. The indications for vitrectomy were either vitreous haemorrhage (53.9%) or retinal detachment with or without vitreous haemorrhage (46.1%).
One harm associated with early scatter photocoagulation was early moderate visual loss, shown more frequently at 6 weeks and 4 months compared to eyes assigned to deferral; however, there was no difference at 3-year follow-up.
The conclusions of the authors of the study were cautious, leaving some uncertainty regarding PRP at the severe NPDR stage;
“Provided careful follow-up can be maintained, scatter photocoagulation is not recommended for eyes with mild or moderate non-proliferative retinopathy. When retinopathy is more severe, scatter photocoagulation should be considered and usually should not be delayed if the eye has reached the high-risk proliferative stage.”
The evidence from ETDRS suggests that treatment of severe NPDR and early PDR was more effective – though confidence intervals were wide - in reducing future visual loss than waiting to treat at high-risk PDR stage, but ETDRS did not provide results separately for severe NPDR and early PDR. The primary endpoint, severe visual loss (defined as visual acuity <5/200 at two consecutive follow-up visits 4 months apart), was very severe. The reduction in high risk PDR might have been expected to lead to further reductions in visual loss with longer follow-up.
Types of laser photocoagulation
Only studies published since 2000 were included, in order to reflect current practice. Studies at any stage of retinopathy were included because of a dearth of laser studies at NPDR stage. For effectiveness in terms of visual state, a minimum duration of six months was preferred, but trials with shorter follow-up were included, because regression of neovascularisation can be seen 2-3 months after PRP. Observational studies of shorter duration provided data on adverse effects.
There were 12 RCTs, generally of good quality, but often with small numbers of patients. The majority of the patients had PDR, with a few with very severe NPDR.
The types of laser, and method of use, varied considerably amongst studies. Newer lasers can do a number of burns at the same time, known as pattern or multi-spot, which reduces the time required for PRP. However, other variations include the type of laser and wave length used (for instance argon vs. diode; 810-nm vs. 532-nm wave length; whether micropulse technology is used), and the parameters than can be changed when actually applying the laser (power, which can be decreased to ‘sub-threshold’ levels or increased to achieve ‘light’ or more ‘marked’ burns; spot size; duration of the laser burn).
There were three trials of multi-spot or pattern photocoagulation against single-spot argon PRP, with a total of 280 eyes treated. Multi-spot photocoagulation appeared to be as effective but with less pain and faster, so less onerous for patients.
Other studies examined different ways of giving standard PRP, some suggesting that lighter burns PRP with conventional lasers gave similar effectiveness but fewer adverse effects than more intense burns. None of the studies showed a significant difference between amongst the lasers in terms of change in visual acuity.
A Japanese approach of selective PRP aimed at ischaemic areas only in pre-proliferative diabetic retinopathy (PPDR – their term, presumably severe NPDR ) delayed progression to PDR, with only 15% of the selective group developing PDR compared to 52% of those receiving no photocoagulation (p=0.03). The rationale is that only the ischaemic areas produce VEGF, and treating only those saves some peripheral vision.
In summary, recent evidence has shown a trend towards “lighter” photocoagulation, but in most studies, without loss of effectiveness. It is worth noticing that lighter photocoagulation can be given with argon machines.
Data on adverse events came from both RCTs and non-randomised studies, with a mixture of different types of lasers and different methods of photocoagulation, different levels of severity of diabetic retinopathy, different follow up times, and different methods of measuring outcomes.
PRP destroys retinal tissue and this can lead to symptoms due to the loss of function of the burned areas, including peripheral visual-field defects, reduced night vision, reduced colour vision, and decreased contrast sensitivity. Visual field defects can occur in up to 50% of treated patients, depending on intensity of PRP and level of testing. However, it does help preserve the more important central vision.
The most important adverse effect associated with PRP is macular oedema, which can lead to a reduction in visual acuity, mostly in the short term, though in one of the older trials, persistent visual acuity losses were attributed to treatment, of one line in 11% and two or more lines in an additional 3%, on the Snellen chart.
It appears in the UK, pattern scan lasers are now replacing single-spot argon lasers for PRP. This is not the case in all countries. The conclusion from the review of recent laser studies is that there have been advances in laser technologies but no convincing evidence as yet that modern lasers are more effective than the argon laser used in ETDRS.
Drug and laser combinations
The review included studies of the efficacy of drug and laser combination in patients with NPDR or PDR. The main interest was reduction in adverse effects, and in particular PRP-associated macular oedema.
Eleven trials compared the efficacy of anti-VEGFs or steroids used in conjunction with PRP. Seven studies used the anti-VEGFs ranibizumab or bevacizumab, and six were of triamcinolone (two trials included both an anti-VEGF drug and a steroid). Five studies included some patients with NPDR but the majority had PDR. Most trials had small numbers of patients and were short-term, but that should not be a problem because the macular oedema provoked by PRP occurs soon after PRP.
For the anti-VEGF drugs the evidence is fairly consistent – a single injection appears to reduce the risk of PRP-induced macular oedema.
In three trials, intra-vitreal triamcinolone reduced the risk of MO after PRP and improved best corrected visual acuity (BCVA) in patients with clinically significant macular oedema (CSMO), but in another it did not. However, intravitreal triamcinolone increased intraocular pressure (IOP), a well-known side-effect of steroids. One trial of a single sub-Tenon’s capsule injection of triamcinolone before PRP showed benefit in preventing visual loss at six months, without increasing IOP. Given the higher risk of adverse effects, anti-VEGF treatment might be preferable to steroids, though cost would need to be considered. Triamcinolone is not licensed for use in the eye.
Overall, adjuvant anti-VEGF or triamcinolone treatment reduced the adverse effects of PRP. The strength of the evidence base is that we have a set of randomised controlled trials. The limitations are their small size, and, for our purposes, that most patients had high risk PDR rather than severe NPDR. We also need more data on the value of anti-VEGF treatment for different patterns of macular oedema, such as foveal and extra-foveal.
One implication of modern laser methods and the use of anti-VEGF or steroid drugs may be a reduction in the risk of DMO when PRP is given in one session. The review concluded that PRP at the severe NPDR stage could be cost-effective but that there were many uncertainties, which could be resolved by a cost-effectiveness study as part of a randomised trial.
The current evidence base is insufficient to recommend that PRP be used for severe NPDR. However this assumes that patients with NPDR are followed up regularly and receive PRP once high-risk PDR develops. If regular follow-up is not provided, the case for PRP for severe NDPR becomes much stronger.
Royle P, Mistry H, Auguste P, Shyangdan D, Lois N, Waugh N. Treatment of non-proliferative diabetic retinopathy: a review of pan-retinal photocoagulation, other forms of laser treatment, and drug treatments: systematic review and economic evaluation. Health Technology Assessment 2014;19: no 51.