Retina · Evidence-based overview
Diabetic retinopathy: classification and stages
Clinical grading of NPDR and PDR, appropriate imaging, screening, and stage-based treatment—with diabetic macular oedema assessed separately.
Retinopathy stage and macular involvement are two separate clinical axes. Good visual acuity does not exclude advanced peripheral retinopathy; conversely, diabetic macular oedema may impair vision despite only mild peripheral retinopathy.[1]
1. Clinical framework
Diabetic retinopathy (DR) is classified by retinal vascular changes as non-proliferative (NPDR) or proliferative (PDR). Diabetic macular oedema (DMO) can occur at any retinopathy stage and is additionally assessed as involving or not involving the foveal centre. Keeping these axes separate prevents peripheral stage from being inferred from visual acuity alone.[1][5]
2. Severity scale
| Stage | Definition / typical findings | Clinical implication |
|---|---|---|
| No DR | No diabetic retinal lesions detected. | Screening interval according to diabetes type and duration, systemic risk, and the national programme. |
| Mild NPDR | Microaneurysms only. | No retinal treatment without a separate indication; address systemic risk factors and monitor the course. |
| Moderate NPDR | More than microaneurysms alone but less than severe NPDR; findings may include intraretinal haemorrhages, hard exudates, cotton-wool spots, mild venous abnormalities, or IRMA. | Risk-adapted follow-up; look for and classify DMO separately. |
| Severe NPDR | At least one 4-2-1 rule criterion without PDR: extensive intraretinal haemorrhages/microaneurysms in all 4 quadrants, definite venous beading in at least 2 quadrants, or prominent IRMA in at least 1 quadrant. | High progression risk and close follow-up. Do not derive treatment from a percentage alone. |
| Very severe NPDR | At least two 4-2-1 criteria. This ETDRS-derived subgroup is not shown separately in every simplified scale. | Particularly high progression risk; prompt retinal assessment. |
| PDR | Neovascularisation of the disc (NVD) or elsewhere (NVE), with or without preretinal/vitreous haemorrhage and fibrovascular proliferation. | Vision-threatening disease requiring treatment; PRP is the German guideline standard, while anti-VEGF can be an alternative in selected cases. |
Use risk figures correctly: The blanket claim “up to 50% develop PDR within one year” is misleading for severe NPDR. In ETDRS data summarised by the German S3 guideline, about 146 per 1,000 eyes with severe NPDR developed treatment-requiring PDR within one year; risk rises in very severe NPDR and with additional risk factors.[1]
3. Recognising lesions—especially IRMA versus neovascularisation
- Microaneurysms: the earliest clinically visible lesion; small red dots.
- Intraretinal haemorrhages: dot/blot haemorrhages in middle or deep layers, or superficial flame-shaped haemorrhages.
- Hard exudates: lipid deposits from chronic leakage; consider DMO when close to the macula.
- Cotton-wool spots: signs of focal nerve-fibre-layer ischaemia; not specific to diabetes.
- Venous beading and IRMA: markers of increasing ischaemia and progression risk.
- NVD/NVE: vessels on or above the internal limiting membrane; fluorescein leakage supports the diagnosis.
IRMA are intraretinal and typically show no or substantially less leakage than active neovascularisation. “No leakage” is not, however, an absolute stand-alone criterion. Fundus morphology, change over time, fluorescein angiography, and—when helpful—cross-sectional OCT/OCTA are interpreted together.[1]
4. Assessment and imaging
| Test | Use | Limitation |
|---|---|---|
| Visual acuity, slit-lamp examination, dilated fundus examination or validated photographic documentation | Baseline assessment of stage, anterior segment, haemorrhage, and other pathology. | Good visual acuity does not exclude severe peripheral DR. |
| Macular OCT | For symptomatic maculopathy and to guide and monitor DMO treatment. | Not a general screening test in asymptomatic eyes without macular fundus changes; review all B-scans, not only central thickness.[1] |
| Fluorescein angiography | Selected questions about non-perfusion, leakage, IRMA/NV, and planning focal laser. | Not required at every routine visit. |
| OCTA | Non-invasive adjunct for capillary networks and vascular morphology. | Not a stand-alone “decisive prognostic test”; consider artefacts and the absence of leakage information. |
| Gonioscopy and intraocular pressure | For advanced ischaemia, suspected rubeosis, or raised pressure; angle neovascularisation may be missed without gonioscopy. | Use when indicated; it does not replace retinal assessment. |
5. Screening and special situations
- Type 1 diabetes: under the commonly used international ADA framework, a first comprehensive eye examination within five years of onset; paediatric and national programmes may add age-related rules.[4]
- Type 2 diabetes: examination at diagnosis because hyperglycaemia may have been present unnoticed for some time.[4]
- Subsequent intervals: usually annual; a longer interval may be reasonable after repeatedly normal findings and good control, whereas established or progressive DR requires shorter intervals. In Germany, the NVL/screening programme is authoritative.[2][4]
- Pregnancy with pre-existing diabetes: counselling and assessment before conception or early in the first trimester, then stage-based follow-up during pregnancy and postpartum. Pregnancy can accelerate progression.[4]
- Gestational diabetes alone: does not carry the same retinopathy risk as pre-existing diabetes; individual comorbidities remain relevant.[4]
Sustained good glucose and blood-pressure control reduces progression risk. Rapid, marked improvement of previously poor glycaemia can temporarily be associated with “early worsening”. This is not a reason to withhold safe metabolic treatment, but it supports coordinated ophthalmic follow-up.[1]
6. Stage-based management
| Finding | Guideline-oriented approach | Key consideration |
|---|---|---|
| No, mild, or moderate NPDR without treatment-relevant DMO | Observe and address risk factors; no PRP or intravitreal medication solely because of this stage. | Set the interval according to the course and total risk.[1] |
| Severe NPDR without treatment-relevant DMO | Close follow-up; the German S3 guideline advises against routine anti-VEGF intravitreal treatment. | PRP may be considered in selected high-risk situations, such as unreliable follow-up, pregnancy, an adverse systemic profile, or extensive non-perfusion.[1] |
| PDR | Panretinal photocoagulation (PRP) should be performed. | Anti-VEGF may be used as an alternative in individual cases; this strategy requires reliable repeated monitoring and potentially repeated injections.[1] |
| PDR plus centre-involving DMO | Treat both axes; often anti-VEGF for DMO and PRP according to stage and course. | Details belong to the dedicated DMO page. |
| Dense vitreous haemorrhage, tractional detachment, or combined complication | Retinal/surgical assessment, ultrasonography when the fundus cannot be seen, and PPV according to findings. | Surgical timing and preoperative anti-VEGF are individual decisions; there is no rigid 3–7-day rule. |
7. What trials show—and what they do not
- Protocol S: Ranibizumab and PRP both enabled good five-year visual outcomes in suitable patients with PDR. The injection strategy required ongoing treatment and follow-up; loss to long-term follow-up limits precision.[7]
- CLARITY: In selected patients with active PDR, aflibercept was non-inferior to PRP at 52 weeks and produced a mean advantage of about 3.9 ETDRS letters. A selected population and one-year horizon do not show that injections permanently replace PRP for every patient.[8]
- Protocol W: Preventive aflibercept for moderate-to-severe NPDR without centre-involving DMO reduced anatomical vision-threatening events, but showed no proven visual-acuity benefit after four years. It does not support routine prophylactic injections for everyone.[9]
- Aspirin/ETDRS: Aspirin 650 mg daily did not materially alter progression or vitreous haemorrhage in the trial. This cannot be generalised to every anticoagulant; patients should not stop medication on their own.[1][10]
8. Discussion points and clinical perspectives
- Pragmatic view: PRP is often particularly robust when reliable long-term follow-up is uncertain. Anti-VEGF may be more attractive with coexisting DMO or a strong preference to avoid laser. This is individualised judgement, not a universal hierarchy.
- Earlier intervention: Some retina specialists discuss earlier PRP or medication in very severe NPDR. The German S3 guideline restricts this to risk situations and does not recommend routine anti-VEGF injections without relevant DMO.
- OCTA/FAZ: Enlargement or vascular loss may carry prognostic information, but should not be presented as the sole reason for or against treatment.
- Fenofibrate: Trials show less retinopathy progression but no clear vision-related benefit to date. Use, regulatory status, and systemic risks belong in a diabetology/internal-medicine decision, not a blanket ophthalmic prescription.[1]
9. Red flags and exam traps
- Sudden visual loss or new dense floaters: exclude vitreous haemorrhage or retinal detachment.
- Pain, rubeosis, or raised pressure: urgently assess for neovascular glaucoma; treat the ischaemic source, not only the pressure.
- Do not distinguish IRMA from NVE by one fluorescein feature alone.
- Do not treat DMO as the “highest stage” of DR; it is a separate axis.
- “High-risk PDR” is a historical DRS/ETDRS risk category, not a reason to leave unequivocally active PDR untreated in modern care.
10. Exam questions
What defines severe NPDR?
At least one 4-2-1 criterion without proliferative signs: extensive intraretinal haemorrhages/microaneurysms in four quadrants, definite venous beading in at least two quadrants, or prominent IRMA in at least one quadrant.
Why must retinopathy stage and DMO not be conflated?
Peripheral ischaemic-proliferative disease and macular leakage can differ independently in severity. Stage, visual acuity, and foveal involvement must therefore be documented separately.
When may PRP be considered in severe NPDR?
Under the German S3 guideline, in selected high-risk patients—for example with unreliable follow-up, pregnancy, an adverse systemic risk profile, or extensive non-perfusion.
Is anti-VEGF for PDR equivalent to “one laser treatment and done”?
No. RCTs support an effective alternative for suitable patients, but it requires reliable long-term follow-up and often repeated injections; PRP may also need supplementation.
What is diabetic retinopathy?
Diabetes can damage the small blood vessels in the retina. Vision often remains normal at first. Later, vessels may leak or abnormal new vessels may grow, bleed, and form scar tissue.
The stages in plain language
- Early stages: small vascular changes, usually without symptoms.
- Advanced non-proliferative stage: more haemorrhages and signs of reduced retinal blood flow.
- Proliferative stage: new, fragile vessels grow and may bleed or pull on the retina.
- Macular oedema: fluid in the centre of the retina; it can occur at any stage and affect central vision.
What protects the eyes?
- Attend eye examinations even when vision is good.
- Manage blood glucose, blood pressure, blood lipids, and kidney disease with the healthcare team.
- With pre-existing diabetes, plan pregnancy with an eye specialist whenever possible.
- Do not stop prescribed blood thinners because of retinopathy without medical advice.
When should I seek urgent help?
Seek urgent eye assessment for sudden visual loss, many new black spots or a dark curtain, severe eye pain, or a red eye with worsening vision.
Sources and evidence limitations
- S3 guideline · Germany · version 1.0 · 4 Aug 2025
DOG/BVA et al.: Special diagnosis and treatment of retinal complications in diabetes
Principal German reference for diagnosis and treatment; some recommendations rely on expert consensus or older ETDRS data. - S3/National Disease Management Guideline · Germany · continuously updated
NVL Type 2 Diabetes
Authoritative for screening and cross-sector care in Germany; intervals remain risk-adapted. - Preferred Practice Pattern · AAO · 2024
Diabetic Retinopathy Preferred Practice Pattern
US guideline; approvals and care pathways cannot be transferred unchanged to Germany. - Clinical guideline · ADA · 2026
Retinopathy, Neuropathy, and Foot Care: Standards of Care in Diabetes—2026
Current international framework for screening and systemic care; national programmes may differ. - International clinical consensus · 2003
Proposed International Clinical Diabetic Retinopathy and Diabetic Macular Edema Disease Severity Scales
Widely used simplified clinical scale; substantially based on ETDRS knowledge and not a replacement for full photographic research grading. - National guideline · NICE NG242 · 2024
Diabetic retinopathy: management and monitoring
UK care context; useful for international comparison but not a German legal or reimbursement basis. - Randomised clinical trial · PDR · 5 years · 2018
DRCR Protocol S: ranibizumab versus PRP
Long-term comparison; meaningful loss to follow-up and high trial adherence limit generalisability. - Randomised phase IIb trial · PDR · 52 weeks · 2017
CLARITY: aflibercept versus PRP
Showed non-inferiority and a mean visual-acuity advantage at one year; selected population and limited duration. - Randomised clinical trial · NPDR · 4 years · 2023
DRCR Protocol W: preventive aflibercept
Fewer anatomical complications, but no proven visual-acuity benefit after four years. - Randomised clinical trial · ETDRS Report 8 · 1991
Aspirin and diabetic retinopathy
Large trial of aspirin 650 mg; do not use as evidence for every anticoagulant or perioperative situation. - Patient information · RCOphth/RNIB · 15 Sep 2026
Understanding diabetes-related eye conditions
Useful for patient-centred communication; not a stand-alone treatment source. - Emerging evidence · phase III / registry · 2026
GLOW2: tarcocimab in diabetic retinopathy
Positive topline results have been reported, but they are not practice-changing until full peer-reviewed publication and regulatory assessment.
EyeWiki and lecture videos were not used as sole evidence. Claims were traced to guidelines, registries, and primary studies. Deliberations not directly based on evidence are explicitly labelled as discussion points.