Author: Claude (AI), from PubMed abstracts · Reviewed by: Dr Mina Abdelmseih · Reviewed: 2026-11-15 · Review due: 2027-11-15

Educational content only. Not personalised medical advice. Clinical decisions remain the responsibility of the treating clinician.

Diabetic Macular Oedema

Diabetic macular oedema (DMO) is a complication of diabetic retinopathy and one of the most common causes of visual impairment in people with diabetes [S3]. Intravitreal anti-VEGF therapy has replaced laser photocoagulation as the standard of care for oedema involving the centre of the macula [S9] [S3]. Its benefit depends on regular, long-term attendance, which many patients find hard to sustain [S5] [S4].

Key points - Risk-adapted eye examinations (every 2 years at low risk, every year or more often otherwise) detect retinopathy at treatable stages [S2]. - At 24 months there are no clinically important differences in vision between the main anti-VEGF drugs [S9]; aflibercept gave more gain than bevacizumab only when baseline vision was 20/50 or worse [S1]. - Starting with bevacizumab and switching to aflibercept if the response is poor gave visual results similar to aflibercept from the start over 2 years [S7]. - Intravitreal steroids give small gains but raise intraocular pressure and cause cataract in phakic eyes [S6]. - Macular laser still reduces visual loss compared with no treatment and is used for milder or non-central oedema [S3].

Evidence last checked: 2026-10-17 (PubMed abstracts).

Red flags

• Moderate or worse visual loss from centre-involving DMO. Eyes starting at 20/50 or worse gained the most from treatment (mean 11.8 to 18.9 letters at 1 year across drugs) [S1].
• Missed or delayed visits. Missed or delayed attendance ranged from 14% to 51%, and 25% of patients were lost to follow-up at 12 months [S5]. In Protocol T, vision fell by a mean 4.7 letters between years 2 and 5 once protocol care stopped [S4].
• Raised intraocular pressure after steroid. About 4 in 10 people given a dexamethasone implant needed pressure-lowering drops [S6].
• Poor-prognosis OCT signs. Hyperreflective retinal or choroidal foci, disorganisation of the retinal inner layers (DRIL) and a disrupted ellipsoid zone, external limiting membrane or cone outer segment line at baseline were associated with worse vision (moderate certainty) [S10].

Recognise it: who is at risk

• Retinopathy affects 9-16% of people with type 2 diabetes and 24-27% of those with type 1, as reported in a German national guideline [S2].
• Risk factors include diabetes duration, the degree of hyperglycaemia, hypertension and diabetic nephropathy [S2].
• DMO results from leakage from diseased retinal capillaries, with thickening of the central retina and poor central vision [S6].

Confirm it

• The key treatment decision is whether the oedema involves the fovea (centre of the macula): intravitreal drugs are recommended for clinically significant oedema with foveal involvement [S2] [S3].
• Trials measured central subfield thickness on OCT; YOSEMITE and RHINE defined absence of DMO as central subfield thickness below 325 µm [S8].
• Baseline OCT also gives prognostic information (see Red flags) [S10]. No biomarker reached moderate certainty for predicting better vision [S10].
• Document findings in a standard way so the ophthalmologist and the diabetes or primary care team share them [S2].

Manage it

Before acting on this section, check the national and local guidelines, product information and regulatory approvals that apply where you practise. See the medical disclaimer.

Tier A (primary care / no specialist)

• Arrange risk-adapted eye examinations: every 2 years at low risk and every year for others, sooner as retinopathy becomes more severe [S2].
• Improve the general risk profile: glycaemia, blood pressure and kidney disease [S2].
• Explain the disease and the purpose of repeated injections; patient understanding was an enabler of adherence [S5].
• Trace patients who miss treatment visits [S5] [S4].

Tier B (general eye clinic)

1. Macular laser. Compared with no treatment, laser reduced the chance of losing vision at 1 year (risk ratio 0.42) and increased resolution of clinically significant DMO at 3 years (risk ratio 1.55) [S3]. It remains an option for milder or non-central DMO [S3]. Subthreshold (especially micropulse) laser may be as effective as standard laser (low-certainty evidence) [S3].
2. Bevacizumab first. In Protocol AC, 1.25 mg bevacizumab first, with a switch to 2.0 mg aflibercept if criteria were met (doses as reported in [S7]), gave a mean gain of 14.0 letters vs 15.0 letters with aflibercept alone over 2 years (difference 0.8 letters, 95% CI -0.9 to 2.5); 70% of bevacizumab-first eyes were switched [S7]. This trial enrolled eyes with vision 20/50 to 20/320 [S7].
3. Steroids where anti-VEGF is unsuitable. The dexamethasone 0.7 mg implant improved vision slightly more than sham (mean difference -0.08 logMAR) but slightly less than anti-VEGF at 12 months (mean difference 0.07 logMAR), with about a fourfold increase in cataract progression and in pressure-lowering treatment [S6]. In pseudophakic eyes, cataract is not a concern [S6]. Weigh any benefit against pressure rise and the need for pressure-lowering treatment [S6]. Dosing beyond these reported values is not covered: follow local formularies.

Tier C (specialist centre)

1. Choice of anti-VEGF. In Protocol T (2015), aflibercept 2.0 mg, bevacizumab 1.25 mg and ranibizumab 0.3 mg (as reported in [S1]) improved vision by 13.3, 9.7 and 11.2 letters at 1 year. With baseline vision 20/32 to 20/40 the drugs did not differ; at 20/50 or worse aflibercept gained more [S1]. A 2023 Cochrane network meta-analysis found small, possibly clinically insignificant 12-month advantages for aflibercept, brolucizumab and faricimab over ranibizumab, and no clinically important differences at 24 months [S9].
2. Longer-acting regimens. Faricimab 6.0 mg treat-and-extend (up to every 16 weeks) matched aflibercept 2.0 mg every 8 weeks for vision at 2 years with fewer injections (median 10 and 11 vs 14); more than 60% of treat-and-extend patients were on 16-weekly dosing at week 96 [S8]. Aflibercept 8 mg every 12 or 16 weeks gave gains similar to aflibercept 2 mg every 8 weeks at 96 weeks (+8.8 and +7.5 vs +8.4 letters) with fewer injections (9.5 and 7.8 vs 13.8) [S11]. Both were industry-funded noninferiority trials [S8] [S11].
3. Long-term follow-up. At 5 years in Protocol T, mean vision remained 7.4 letters above baseline, and 68% of eyes needed further injections between years 2 and 5 [S4].
4. Steroid implants. The fluocinolone 0.19 mg implant improved vision slightly vs sham, but cataract progression occurred in about 8 in 10 participants [S6].

Refer when

Pitfalls

• Assuming aflibercept is needed for every patient: with mild visual loss the three drugs did not differ [S1], and bevacizumab first showed no significant difference in 2-year vision [S7].
• Stopping follow-up after the first 2 years: vision drifted down between years 2 and 5 [S4].
• Using steroids in phakic eyes without planning for cataract and pressure monitoring [S6].
• Treating real-world results as trial results: under-treatment is common outside trials [S9] [S5].

Evidence notes

• 13 of 23 trials in the Cochrane network meta-analysis had industry funding, and mortality estimates were very imprecise [S9].
• Most steroid trials had unclear or high risk of bias, and steroid vs anti-VEGF results were inconsistent [S6]. Laser trials were poorly reported; only one came from Africa [S3].
• In Protocol AC, serious adverse events were more frequent with aflibercept monotherapy (52% vs 36%) [S7]; the abstract does not attribute this to the drug.
• Not covered by these sources: drug cost and access, biosimilars, injection technique, vitrectomy for DMO, combination therapy, and DMO after cataract surgery.

Related articles

• Proliferative Diabetic Retinopathy: Laser, Anti-VEGF and Vitrectomy
• Anti-VEGF biosimilars and off-label bevacizumab
• OCT interpretation: macula and retinal nerve fibre layer
• Diabetic retinopathy screening programmes

References

See the numbered source list below. Each marker [S#] links to its PubMed and DOI record.