Retina · Evidence-based overview

Diabetic macular oedema: diagnosis, treatment and evidence limits

Medical review: 30 September 2026 · 22:48 CEST (Europe/Berlin)Published: 30 September 2026About 8 minutes

Clinical takeaway

Treatment of diabetic macular oedema (DMO) depends on foveal involvement, best-corrected visual acuity, trajectory and associated findings. OCT oedema alone does not automatically justify an injection. Anti-VEGF is generally the first pharmacological option for foveal DMO with relevant visual impairment; monitored observation can be appropriate when vision remains good. [1][2]

Definition and scope

Disruption of the blood–retinal barrier allows fluid to accumulate in the macula. VEGF, inflammatory signalling and neurovascular changes interact; this mechanism does not reliably identify the best drug for an individual eye.

DMO can occur at different retinopathy stages. Document retinopathy stage and macular involvement separately. Full staging and PRP are covered on the canonical DR page. This overview primarily concerns adults with DMO.

Diagnosis and imaging

Assess symptoms, best-corrected acuity in both eyes, previous treatment, lens status, intraocular pressure, fundus findings, comorbidities and ability to attend follow-up. Consider alternative causes including vein occlusion, postoperative/inflammatory oedema and neovascular AMD. [1]

Test/finding Role and limitation
OCT Foveal involvement, intra-/subretinal fluid and change over time; inspect all B-scans, not just central thickness. Check segmentation.
Vitreomacular traction/epiretinal membrane Assess mechanical contribution; an interface finding alone is not an automatic surgical indication.
DRIL, ellipsoid zone, external limiting membrane May relate to visual function/prognosis; not validated rules for choosing steroid over anti-VEGF.
Fluorescein angiography (FA) Selective use for suspected ischaemia, diagnostic uncertainty or planned focal laser; not required before every injection.
OCT angiography (OCTA) Adjunctive perfusion assessment; does not show leakage. Segmentation, motion and shadow artefacts limit FAZ/vessel-density measurements.

OCT thickness is an anatomical surrogate. Reduced fluid does not guarantee proportional visual recovery. Hyperreflective foci (HRF), subretinal fluid and cyst size do not reliably diagnose “steroid-responsive inflammation”. A 2025 meta-analysis found no biomarker with at least moderate certainty associated with better visual acuity or its improvement after treatment. [1][5][6]

Guideline-based treatment decisions

Situation Approach
Non-foveal oedema Monitor; offer focal laser individually for appropriate, vision-threatening clinically significant findings. No blanket injection indication.
Foveal oedema, good vision Consider monitored observation; watch for visual deterioration and ensure access to rescue treatment.
Foveal oedema, relevant visual impairment Anti-VEGF where morphology suggests potential functional benefit; discuss individual risks and treatment burden.
Insufficient response Reassess diagnosis, treatment intensity, trajectory and causes of visual loss; consider agent switch or steroid implant individually.
Relevant traction Vitreoretinal assessment; consider vitrectomy when functional/morphological findings support surgery.

This approach follows the German S3 guideline. Ischaemia and structural damage may limit visual potential, but cannot be reliably quantified by a single automated OCTA value. [1][6]

Primary trials: findings and interpretation

Trial Population/design, result and limitation
Protocol V [2] Randomised trial, 702 adults, centre-involved DMO with acuity 20/25 or better. At two years, ≥5-letter loss: 16% aflibercept, 17% laser, 19% observation; no significant difference. Observation included acuity-triggered rescue injections and reliable follow-up; do not extrapolate to visually impaired eyes.
Protocol T [3] Randomised trial, 660 participants. With baseline acuity 20/50–20/320, two-year gains were +18.1/+13.3/+16.1 letters with aflibercept/bevacizumab/ranibizumab. Aflibercept outperformed bevacizumab, but not ranibizumab. Better baseline vision showed no significant pairwise differences. Protocol-specific care; not a ranking of all current agents.
Protocol AC [4] Randomised trial, 312 eyes, acuity 20/50–20/320. Initial aflibercept versus bevacizumab with defined switching: +15 versus +14 letters averaged over two years; difference 0.8 (95% CI −0.9 to 2.5). 70% switched. Evidence for a supervised step strategy, not indefinite bevacizumab despite inadequate response.

Randomisation and standardised care strengthen the evidence. Treatment was not masked in every respect; intensive trial follow-up may differ from routine care. Non-significance does not establish universal equivalence. Agent/dose selection, approval and prescribing information must reflect current local practice.

Medicines, safety and monitoring

Choose anti-VEGF according to baseline vision, previous treatment, benefits/risks, approval, access and follow-up feasibility. Bevacizumab is used off-label and requires appropriate counselling. Loading, PRN and treat-and-extend depend on the agent and protocol; there is no universal “3–6 injections” regimen. [1][3][4]

Steroid implants may be considered after insufficient anti-VEGF response or when anti-VEGF is unsuitable. Cataract and raised intraocular pressure are major risks. Review glaucoma status, previous steroid response and product-specific contraindications; plan pressure and lens monitoring. “Any glaucoma is an absolute contraindication” is too broad. Additional OCT drying alone does not justify combination treatment. [1]

Before switching, look beyond residual fluid: review acuity/OCT trajectory, adequate treatment exposure, adherence, cataract, ischaemia and traction. A particular cyst diameter or HRF finding does not establish treatment failure. [1][5]

Systemic care

Coordinate individual glucose, blood-pressure and lipid targets and renal assessment with diabetes/primary care. These measures support long-term retinal health but do not replace necessary local treatment. Tailor follow-up to activity, treatment and risk. [1]

What supports practice, and what remains uncertain?

Supported: function-based decisions, anti-VEGF for visually significant foveal DMO, and observation of suitable eyes with good vision and rescue access. Uncertain: reliable biomarker-guided drug selection and a universal optimal switch time. Anatomical improvement or longer achievable intervals alone do not establish superior visual function. [1][2][3][4][5]

Patient explanation and warning signs

Diabetes can make vessels leak at the centre of detailed vision. Depending on vision and examination, monitoring, injections or other treatments may help. Observation means planned visits, not abandonment of care.

Arrange prompt assessment for a new marked decline in vision. Sudden severe visual loss or a dark curtain requires urgent examination. Increasing pain, redness or visual decline after an injection requires immediate ophthalmic assessment.

Educational information; not a substitute for individual examination or treatment advice.

References

  1. DOG/BVA/DDG. S3: Spezielle Diagnostik und Therapie der Netzhautkomplikationen des Diabetes. Version 1.0, 04.08.2025. Official guideline.
  2. Baker CW et al. Protocol V. JAMA. 2019. DOI: 10.1001/jama.2019.5790. PMID 31037289.
  3. Wells JA et al. Protocol T, two-year results. Ophthalmology. 2016. DOI: 10.1016/j.ophtha.2016.02.022. PMID 26935357.
  4. Jhaveri CD et al. Protocol AC. N Engl J Med. 2022. DOI: 10.1056/NEJMoa2204225. PMID 35833805.
  5. Baseline OCT Biomarkers Associated with Visual Acuity in Diabetic Macular Edema: A Systematic Review and Meta-analysis. 2025. PMID 40803536.
  6. OCTA segmentation-error study in DME. 2019. PMID 31036586.
  7. AAO. Diabetic Retinopathy Preferred Practice Pattern. 2025. DOI: 10.1016/j.ophtha.2024.12.020. PMID 39918521. Bibliographic/context reference; detailed recommendations are anchored to source 1 because PPP full text was unavailable during this audit.