Diabetic Retinopathy
Clinical Features and Management
Examination question · ~800 words
Discuss the clinical features, classification, investigations, and management of diabetic retinopathy, including diabetic macular oedema and proliferative diabetic retinopathy.
Try to outline your answer mentally before expanding sections below.
Diabetic retinopathy (DR) is the commonest ocular microvascular complication of diabetes mellitus and a leading cause of preventable blindness in the working-age population. It represents a spectrum from asymptomatic microaneurysms to sight-threatening diabetic macular oedema, proliferative diabetic retinopathy, vitreous haemorrhage, tractional retinal detachment, and neovascular glaucoma. Management requires systemic risk-factor control, stage-based retinal treatment, and timely use of anti-VEGF, laser, and vitreoretinal surgery.
DR affects approximately one-third of diabetics globally. After 20 years of diabetes, nearly all patients with type 1 diabetes and a majority of patients with type 2 diabetes develop some degree of DR. It is a major cause of new blindness in adults aged 20-65 years.
| Parameter | Type 1 DM | Type 2 DM |
|---|---|---|
| Time of onset | Usually after 7-10 years; virtually all by 20 years | May be present at diagnosis |
| PDR risk by 20 years | ~50% | ~10% |
| Macular oedema | Less common early | Major cause of visual loss |
| Renal association | Strong nephroretinal axis | Present but less tightly correlated |
Hyperglycaemia-to-Neovascularisation Cascade
| Step | Mechanism | Consequence |
|---|---|---|
| Pericyte loss | Sorbitol accumulation via aldose reductase pathway; advanced glycation end-product injury | Loss of capillary tone and autoregulation |
| Basement membrane thickening | AGE-mediated collagen cross-linking and fibronectin deposition | Reduced pericyte-endothelial interaction and increased permeability |
| Endothelial dysfunction | Oxidative stress, PKC-beta activation, eNOS dysfunction | Breakdown of the blood-retinal barrier |
| Capillary occlusion | Leucostasis, platelet aggregation, ICAM-1 upregulation | Non-perfusion, cotton-wool spots, IRMA |
| VEGF upregulation | Hypoxia-induced HIF-1 alpha release from inner retina | NVD, NVE, DME, NVI |
| Neurodegeneration | Ganglion cell apoptosis and inner retinal thinning may precede visible vascular signs | Reduced contrast sensitivity and early functional loss |
- Four major biochemical pathways: polyol pathway, AGE formation, PKC-beta activation, and hexosamine pathway.
- Earliest histological change: selective pericyte loss.
- Earliest clinical lesion: microaneurysm.
International Clinical DR Severity Scale
| Grade | Fundus Findings | Key Point |
|---|---|---|
| No apparent DR | No abnormalities | - |
| Mild NPDR | Microaneurysms only | Earliest clinical stage |
| Moderate NPDR | More than microaneurysms but less than severe NPDR: dot-blot haemorrhages, hard exudates, cotton-wool spots | Variable severity |
| Severe NPDR | 4-2-1 rule: >20 haemorrhages in all 4 quadrants; venous beading in 2+ quadrants; IRMA in 1+ quadrant | Any one criterion = severe NPDR; ~15% risk of high-risk PDR within 1 year |
| Very severe NPDR | Two or more severe NPDR criteria | Higher risk of progression |
| PDR | NVD or NVE with or without vitreous/preretinal haemorrhage | Neovascular stage |
High-Risk PDR: DRS Criteria
- NVD >= 1/4-1/3 disc area.
- Any NVD with vitreous or preretinal haemorrhage.
- NVE >= 1/2 disc area with vitreous or preretinal haemorrhage.
- High-risk PDR carries >25% risk of severe visual loss over 2 years without treatment.
Symptoms
| Stage | Typical Symptoms |
|---|---|
| Early NPDR | Usually asymptomatic |
| Moderate-severe NPDR | Blurred vision if DME is present; floaters if haemorrhage occurs |
| PDR | Sudden painless visual loss from vitreous haemorrhage; metamorphopsia if DME coexists |
| Advanced DR | Severe visual loss from TRD; painful red eye if rubeosis and neovascular glaucoma develop |
Fundus Signs
| Sign | Description | Exam Significance |
|---|---|---|
| Microaneurysms | Small red dots due to focal capillary wall outpouching | Earliest clinically visible lesion |
| Dot-blot haemorrhages | Deep intraretinal haemorrhages in inner nuclear and outer plexiform layers | Marker of NPDR severity |
| Hard exudates | Yellow-white, sharply demarcated lipoprotein deposits, often in circinate rings | Subfoveal hard exudate suggests chronic DME and poor visual prognosis |
| Cotton-wool spots | Fluffy white superficial nerve fibre layer infarcts due to axoplasmic flow interruption | Indicate retinal ischaemia; typically resolve over 4-12 weeks |
| IRMA | Dilated tortuous intraretinal channels bypassing occluded capillary beds | Does not leak on FFA; differentiates it from NVE |
| Venous beading | Sausage-link venous irregularity | Strong predictor of progression to PDR |
| NVD/NVE | New vessels on disc or elsewhere, growing on retinal surface or optic disc | Leak on FFA; predispose to VH, fibrosis, and TRD |
DME is the leading cause of moderate visual loss in DR, especially in type 2 diabetes. It results from breakdown of the inner and outer blood-retinal barriers with fluid accumulation in Henle's layer and the outer nuclear layer.
| Type | Definition | Management Relevance |
|---|---|---|
| Focal DME | Localised leakage from identifiable microaneurysm clusters | Responds to focal laser in selected non-centre-involving cases |
| Diffuse DME | Generalised breakdown of inner BRB with diffuse retinal thickening or cystoid spaces | Anti-VEGF preferred if centre-involving |
| CSME: ETDRS | Retinal thickening within 500 microns of foveal centre; hard exudates within 500 microns with adjacent thickening; or thickening >=1 disc area within 1 disc diameter of foveal centre | Classic treatment threshold |
| Centre-involving DME | Foveal centre/subfield involved on OCT | Anti-VEGF first-line when vision is reduced |
| Non-centre-involving DME | Thickening outside foveal centre | Observe or focal/grid laser depending on leakage and progression |
Fundus Fluorescein Angiography
| Finding | FFA Appearance |
|---|---|
| Microaneurysms | Early hyperfluorescent dots with late staining |
| IRMA | Hyperfluorescent channels without late leakage |
| NVD/NVE | Profuse leakage with late diffuse hyperfluorescence |
| Capillary non-perfusion | Hypofluorescent non-perfused retina; FAZ enlargement |
| DME | Late macular leakage; petaloid pattern in cystoid oedema |
| Hard exudates | Blocked fluorescence |
Optical Coherence Tomography
- Gold standard for diagnosing and monitoring DME.
- Documents central/foveal subfield thickness and response to therapy.
- Patterns: diffuse spongy oedema, cystoid macular oedema, subretinal fluid, vitreomacular traction.
- Poor prognostic biomarkers: DRIL, subretinal fluid in some contexts, ellipsoid zone disruption, chronic hard exudates.
Other Tests
| Investigation | Use |
|---|---|
| Ultra-widefield FFA | Peripheral ischaemia and NVE beyond vascular arcades |
| OCT angiography | Non-invasive FAZ and capillary plexus assessment; no leakage information |
| B-scan ultrasonography | Vitreous haemorrhage, TRD, or RRD when fundus view is obscured |
| HbA1c, BP, lipids, renal function | Systemic risk-factor assessment |
| Microperimetry | Macular sensitivity mapping |
Systemic control is the foundation of DR management. Ophthalmic treatment fails if hyperglycaemia, hypertension, dyslipidaemia, renal disease, anaemia, and pregnancy-related acceleration are ignored.
| Risk Factor | Target / Intervention | Evidence |
|---|---|---|
| HbA1c | Usually <7%; individualise in frail patients or high hypoglycaemia risk | DCCT, UKPDS |
| Blood pressure | Tight BP control, commonly <130/80 mmHg if tolerated | UKPDS, ACCORD Eye |
| Lipids | Statin as indicated; fenofibrate useful in selected patients with DR/dyslipidaemia | ACCORD Eye, FIELD |
| Renal disease | Screen and treat nephropathy; coordinate with physician/nephrologist | Nephroretinal association |
| Pregnancy | Pre-conception counselling and frequent retinal monitoring | Pregnancy may accelerate DR |
Anti-VEGF Therapy
Anti-VEGF therapy is first-line for centre-involving DME with reduced vision.
| Agent | Dose | Key Evidence | Key Point |
|---|---|---|---|
| Ranibizumab | 0.3/0.5 mg intravitreal | RIDE/RISE, RESTORE | Improves BCVA and reduces DME |
| Aflibercept | 2 mg intravitreal | VISTA/VIVID, DRCR Protocol T | Superior when baseline VA is 20/50 or worse |
| Bevacizumab | 1.25 mg intravitreal | DRCR Protocol T | Effective and low-cost; less effective than aflibercept in worse baseline VA subgroup |
| Faricimab | 6 mg intravitreal | YOSEMITE/RHINE | Dual VEGF-A/Ang-2 inhibition; extended dosing possible in many patients |
- DRCR Protocol T: aflibercept was superior to bevacizumab and ranibizumab when baseline VA was 20/50 or worse; at 20/32-20/40, all three had similar outcomes.
- Initial loading is usually monthly, followed by treat-and-extend or PRN based on OCT and visual response.
Steroids and Laser
| Treatment | Indication | Major Adverse Effects |
|---|---|---|
| Triamcinolone 4 mg | Pseudophakic eye, anti-VEGF non-response, inflammatory component | Cataract, IOP rise |
| Dexamethasone implant 0.7 mg | Pseudophakic eye, poor anti-VEGF response, pregnancy or cardiovascular contraindication to anti-VEGF | Cataract, IOP rise |
| Fluocinolone acetonide implant | Chronic refractory DME | High cataract and glaucoma risk |
| Focal/grid laser | Non-centre-involving CSME or focal leaking microaneurysms; adjunct in persistent DME | Paracentral scotoma, scar enlargement if poorly applied |
| Subthreshold micropulse laser | Selected non-centre-involving or mild DME | Less visible thermal damage |
Panretinal Photocoagulation
PRP ablates peripheral ischaemic retina, reducing VEGF drive. It remains the durable standard for high-risk PDR, especially where follow-up is uncertain.
| Indication | Timing | Rationale |
|---|---|---|
| High-risk PDR | Immediate PRP | DRS: reduces severe visual loss by about 50% |
| Severe NPDR in type 2 diabetes | Consider early PRP in selected high-risk or poorly compliant patients | ~15% develop high-risk PDR within 1 year; very severe NPDR has ~45-52% risk |
| Severe NPDR in type 1 diabetes | Close observation unless progression or poor follow-up | Avoid unnecessary peripheral field loss in patients who can be monitored |
| Pregnancy with severe DR/PDR | Prompt laser, often in second trimester if needed | Anti-VEGF relatively avoided; disease may accelerate |
- Typical PRP: 1200-1600 burns, 500 micron spot size, 0.1 second, one burn-width apart, sparing the macula and central 2 disc diameters.
- Complications: peripheral field constriction, reduced night vision, transient macular oedema, exudative detachment rarely.
Anti-VEGF for PDR
DRCR Protocol S showed ranibizumab was non-inferior to PRP for PDR visual acuity outcomes at 2 years, with less peripheral field loss and less DME. However, anti-VEGF monotherapy requires reliable follow-up; missed injections may result in rapid recurrence or worsening of neovascularisation. Anti-VEGF is especially useful as an adjunct before vitrectomy, in NVI/NVA, and when PDR coexists with CI-DME.
| Indication | Procedure | Timing |
|---|---|---|
| Non-clearing vitreous haemorrhage | Pars plana vitrectomy + endolaser PRP | Usually >1 month in type 1 DM or >3 months in type 2 DM, earlier if bilateral/one-eyed/occupational need |
| TRD involving or threatening macula | PPV + membrane dissection + endolaser | Urgent |
| Combined TRD + RRD | PPV + membrane dissection + endolaser + tamponade | Emergency |
| Dense VH preventing PRP in active PDR | PPV + intraoperative PRP | Urgent |
| Vitreomacular traction causing DME | PPV with posterior hyaloid/ILM management | Elective depending on visual impact |
Pre-operative anti-VEGF, commonly bevacizumab 1.25 mg given 3-7 days before PPV, reduces intraoperative bleeding and facilitates membrane dissection. Surgery should not be delayed after injection because rapid fibrovascular contraction can worsen TRD, especially if delayed beyond about 2 weeks.
| DR Stage | DME Status | Action |
|---|---|---|
| Mild NPDR | No DME | Systemic optimisation and periodic review |
| Mild/moderate NPDR | Non-centre-involving DME | Observe or focal/grid laser; anti-VEGF if worsening or vision affected |
| Any DR stage | Centre-involving DME with reduced VA | Anti-VEGF first-line; consider steroid or laser in selected cases |
| Severe NPDR | No DME | Close follow-up; consider early PRP in type 2 DM, very severe NPDR, or poor follow-up |
| Non-high-risk PDR | No DME | PRP or anti-VEGF depending on compliance, lens status, field needs, and access |
| High-risk PDR | No DME | Immediate PRP; anti-VEGF adjunct if active NV or haemorrhage risk |
| High-risk PDR | CI-DME | Anti-VEGF first or combined with prompt PRP |
| Advanced PDR | VH/TRD/RRD | PPV + membrane peel + endolaser |
| Patient Group | Initial Eye Examination | Follow-up |
|---|---|---|
| Type 1 DM | 5 years after diagnosis, or at puberty depending on guideline context | Annual if no DR; more frequent with DR |
| Type 2 DM | At diagnosis | Annual if no DR; more frequent with DR |
| Pre-existing DM planning pregnancy | Pre-conception counselling and retinal evaluation | Each trimester; more frequent if DR is present |
| Gestational diabetes alone | Routine DR screening usually not required unless pre-existing diabetes is suspected | Individualise if abnormal retinal findings or diabetes persists postpartum |
Teleophthalmology and validated AI-assisted screening systems expand access to DR screening, especially in resource-limited settings. Positive screening must be linked to definitive ophthalmic care.
| Trial | Focus | Key Finding |
|---|---|---|
| DCCT | Type 1 DM glycaemic control | Intensive therapy reduced new onset of DR by 76% in the primary prevention cohort and progression of established DR by 54% in the secondary intervention cohort |
| UKPDS | Type 2 DM glycaemia and BP control | Tight metabolic and BP control reduced microvascular complications |
| DRS | PRP for PDR | PRP reduced severe visual loss by about 50% in high-risk PDR |
| ETDRS | Laser for DME, aspirin safety, PRP timing | Defined CSME; focal laser reduced moderate visual loss; aspirin did not increase VH risk |
| DRCR Protocol T | Aflibercept vs bevacizumab vs ranibizumab for DME | Aflibercept superior when baseline VA was 20/50 or worse |
| DRCR Protocol S | Ranibizumab vs PRP for PDR | Ranibizumab non-inferior to PRP with less field loss, but requires reliable follow-up |
| RIDE/RISE | Ranibizumab for DME | Improved vision and reduced DME |
| VISTA/VIVID | Aflibercept for DME | Superior to laser for centre-involving DME |
| YOSEMITE/RHINE | Faricimab for DME | Non-inferior to aflibercept with extended dosing possible |
| PAGODA | Port delivery system with ranibizumab for DME | Demonstrated non-inferiority to monthly ranibizumab and supported FDA approval for DME in February 2025 |
| ARCHWAY | Port delivery system with ranibizumab for wet AMD | Non-inferior to monthly ranibizumab in neovascular AMD; should not be cited as DME/DR efficacy evidence |
| ReFineDR/DeFineDR | Finerenone post-hoc DR analysis | Numerical reduction in vision-threatening DR complications, but not statistically significant at the primary 2-year timepoint; hypothesis-generating |
- Faricimab: dual VEGF-A and Ang-2 inhibition; enables extended dosing in many DME patients.
- Retinal neurodegeneration: ganglion cell and inner plexiform layer thinning may precede clinically visible DR.
- AI-assisted screening: autonomous or semi-autonomous detection of more-than-mild DR is expanding population-level screening.
- Port delivery system: ARCHWAY applies to wet AMD; PAGODA supports DME use and led to FDA approval for DME in February 2025.
- Finerenone: post-hoc pooled analyses suggest possible reduction in vision-threatening DR complications, but DR progression was not a primary endpoint and the evidence remains hypothesis-generating.
Trap 1 — FALSE
“Cotton-wool spots alone classify a patient as diabetic retinopathy.”
Cotton-wool spots indicate retinal nerve fibre layer infarcts and may occur in hypertension, CRVO, HIV, anaemia, and other ischaemic states. DR diagnosis requires the diabetic clinical context and accompanying retinal findings.
Trap 2 — FALSE
“Aspirin is contraindicated in diabetic retinopathy because it increases vitreous haemorrhage.”
ETDRS showed aspirin did not increase vitreous haemorrhage risk; it should not be stopped when medically indicated for cardiovascular protection.
Trap 3 — FALSE
“IRMA shows profuse late leakage on FFA.”
IRMA is intraretinal and does not show profuse late leakage. NVE breaches the ILM and leaks profusely on FFA.
Trap 4 — FALSE
“Rubeosis iridis with NVA should be treated first with filtration surgery.”
Neovascular glaucoma from DR requires urgent anti-VEGF plus PRP. Filtering surgery is deferred until neovascular activity is controlled unless required for uncontrolled IOP.
Trap 5 — FALSE
“Focal laser is obsolete in DME.”
Focal/grid laser remains useful for selected non-centre-involving CSME, focal leaking microaneurysms, and as adjunctive therapy.
Trap 6 — FALSE
“Anti-VEGF monotherapy for PDR is suitable even when follow-up is unreliable.”
Anti-VEGF can control PDR but requires reliable follow-up. PRP is preferred when compliance is doubtful because it provides more durable VEGF-drive reduction.
Trap 7 — FALSE
“DME occurs only in advanced diabetic retinopathy.”
DME can occur at any stage of DR, including mild NPDR. DR severity does not reliably predict DME presence.
Trap 8 — TRUE
“Steroid implants are especially useful in pseudophakic eyes or when anti-VEGF is contraindicated.”
Intravitreal steroids are useful in pseudophakic patients, anti-VEGF non-responders, pregnancy, or situations where anti-VEGF is relatively contraindicated, but cataract and IOP rise must be monitored.
Trap 9 — FALSE
“Severe NPDR in type 2 diabetes has a standard 25% risk of progressing to PDR within 1 year.”
A better exam figure is ~15% risk of high-risk PDR within 1 year for severe NPDR; very severe NPDR carries a much higher risk of about 45-52%.
Trap 10 — FALSE
“ARCHWAY is the key DME trial for Susvimo.”
ARCHWAY studied the port delivery system in wet AMD. For DME, PAGODA demonstrated non-inferiority to monthly ranibizumab and supported FDA approval for DME in February 2025.
Q: What is the earliest histological change in diabetic retinopathy?
A: Selective pericyte loss.
Q: What is the earliest clinical lesion of diabetic retinopathy?
A: Microaneurysm.
Q: Differentiate IRMA from NVE on FFA.
A: IRMA is intraretinal and does not leak late; NVE breaches the ILM and shows profuse late leakage.
Q: What is the 4-2-1 rule?
A: Severe NPDR: >20 haemorrhages in all 4 quadrants, venous beading in 2+ quadrants, or IRMA in 1+ quadrant. Any one criterion is sufficient.
Q: When is aflibercept superior in DRCR Protocol T?
A: When baseline visual acuity is 20/50 or worse; with better baseline VA, aflibercept, ranibizumab, and bevacizumab have similar outcomes.
Q: Most common cause of sudden painless visual loss in PDR?
A: Vitreous haemorrhage.
Q: Why is the vitrectomy threshold shorter in type 1 diabetes than type 2 diabetes for non-clearing vitreous haemorrhage?
A: Type 1 patients often have organised gel vitreous with less spontaneous VH clearance, more aggressive fibrovascular proliferation, and greater need for early visual rehabilitation; hence PPV is considered after about 1 month rather than 3 months.
Q: What is high-risk PDR?
A: NVD >=1/4-1/3 disc area, any NVD with vitreous/preretinal haemorrhage, or NVE >=1/2 disc area with vitreous/preretinal haemorrhage.
Q: What is the treatment priority in PDR with rubeosis?
A: Urgent anti-VEGF plus PRP, with IOP control and glaucoma surgery only after neovascular activity is controlled if possible.
A 45-year-old type 2 diabetic patient on insulin presents with BCVA 6/60 OD and 6/9 OS. Fundus OD shows NVD 1/3 disc area, scattered preretinal haemorrhage, hard exudates at the fovea, and OCT shows centre-involving DME with foveal thickness 480 microns and subretinal fluid. HbA1c is 10.2%, BP is 154/96 mmHg, and he lives 80 km from the hospital. OS shows moderate NPDR without DME. Prioritise management and justify your plan.
Answer
OD has high-risk PDR with CI-DME. NVD 1/3 disc area with preretinal haemorrhage meets DRS high-risk criteria, while the OCT confirms visually significant DME.; Give intravitreal anti-VEGF first, preferably aflibercept because baseline VA is worse than 20/50, consistent with DRCR Protocol T. This treats CI-DME and induces short-term regression of neovascularisation.; Perform PRP urgently, ideally same sitting or within 1-2 weeks. Because the patient lives far away and follow-up is uncertain, complete PRP in fewer sessions if possible rather than relying on anti-VEGF monotherapy.; Continue anti-VEGF loading for DME with OCT and VA monitoring. If compliance becomes unreliable, durable PRP coverage becomes the priority for PDR safety.; OS has moderate NPDR without DME. No laser is indicated immediately; follow-up should be accelerated to 4-6 monthly because the fellow eye has high-risk PDR and systemic control is poor.; Systemic optimisation is mandatory: urgent physician referral for glycaemic control, BP control, lipid management including statin and possible fenofibrate, renal evaluation, and counselling about symptoms of vitreous haemorrhage and TRD.; Assess for NVI/NVA by slit lamp and gonioscopy. If present, treat with urgent anti-VEGF plus PRP and control IOP; filtration surgery is not first-line before neovascular regression.
References
- DCCT. Type 1 DM glycaemic control. Intensive therapy reduced new onset of DR by 76% in the primary prevention cohort and progression of established DR by 54% in the secondary intervention cohort
- ETDRS. Laser for DME, aspirin safety, PRP timing. Defined CSME; focal laser reduced moderate visual loss; aspirin did not increase VH risk
- DRCR Protocol T. Aflibercept vs bevacizumab vs ranibizumab for DME. Aflibercept superior when baseline VA was 20/50 or worse
- DRCR Protocol S. Ranibizumab vs PRP for PDR. Ranibizumab non-inferior to PRP with less field loss, but requires reliable follow-up
- FDA approval for DME in February 2025
- DCCT Research Group. N Engl J Med. 1993;329:977-986. Table 3 and abstract.
- ETDRS Report No. 12; internal consistency with DR-011 in this ledger.
- Roche press release 2025-02-04 (FDA approval of Susvimo for DME based on PAGODA; https://www.roche.com/media/releases/med-cor-2025-02-04); JAMA Ophthalmology 2025 PAGODA RCT (https://jamanetwork.com/journals/jamaophthalmology/fullarticle/2830898); Genentech press release Oct 2021 for wet AMD approval.