What is diabetic retinopathy?

Diabetic retinopathy is damage to the small blood vessels of the retina caused by chronic high blood glucose. In the early (non-proliferative) stages, capillaries develop microaneurysms, leak fluid, protein and blood, and progressively close off, leaving areas of retina without blood supply. In the advanced (proliferative) stage, the ischaemic retina releases VEGF, driving growth of fragile new vessels that bleed into the vitreous and form fibrous tissue that can pull the retina off — causing sudden, severe and sometimes permanent vision loss. Separately, leakage at the macula causes diabetic macular oedema, which is the commonest cause of visual loss in diabetes and can occur at any stage of retinopathy.

Why do I need eye screening if my vision is fine?

Because sight-threatening diabetic retinopathy causes no symptoms at all until damage is already done. Proliferative retinopathy and macular oedema can both be advanced in an eye that still reads the bottom of the chart. By the time blurring, floaters or a dark patch appear, vision has often already been irreversibly lost. Annual screening with a retinal photograph — through dilated pupils — detects the disease at a stage when laser or injection treatment reliably prevents blindness. This is why screening is offered to everyone with diabetes from diagnosis (from age 12 in type 1), regardless of symptoms, regardless of how good the vision is, and regardless of how well controlled the diabetes seems.

How is diabetic retinopathy treated?

Treatment depends on what is threatening vision. Diabetic macular oedema involving the centre with reduced vision: intravitreal anti-VEGF injections (aflibercept, ranibizumab, bevacizumab, faricimab) are first-line and can improve vision, not merely stabilise it; intravitreal steroid implants are used in non-responders or where frequent injections are impractical; focal laser is still used for some non-centre-involving oedema. Proliferative retinopathy: panretinal photocoagulation laser, which causes the new vessels to regress and has prevented severe visual loss for five decades; anti-VEGF can be used as an alternative or adjunct but depends on sustained attendance. Advanced complications — non-clearing vitreous haemorrhage or tractional retinal detachment — require vitrectomy surgery. Alongside all of this, glycaemic, blood pressure and lipid control alter the underlying trajectory.

Will better diabetes control reverse my retinopathy?

Better control substantially slows progression and reduces the risk of new retinopathy — the DCCT and UKPDS trials showed large, durable benefits, with a legacy effect persisting for years after the trial period. It does not reverse established structural damage such as scarring or areas of closed capillaries. One important caveat needs to be handled properly rather than avoided: a large, rapid fall in HbA1c — for example when starting insulin, after bariatric surgery, or on starting a GLP-1 agonist — can cause a transient WORSENING of retinopathy over the following months. This is not a reason to accept poor control; it is a reason to check the retina before and during rapid intensification, so that any deterioration is detected and treated promptly. Long-term, good control is unambiguously protective.

Can diabetic retinopathy cause blindness?

Yes — it is the leading cause of preventable blindness in working-age adults worldwide. But the word "preventable" is the important one. With systematic annual screening and timely treatment, severe visual loss from diabetic retinopathy is largely avoidable: laser treatment reduces severe visual loss in proliferative disease by more than half, and anti-VEGF therapy improves vision in a substantial proportion of eyes with macular oedema. The people who go blind from diabetic retinopathy are overwhelmingly those who were never screened, were lost to follow-up, or presented only once symptoms appeared. This is why the health-system failures — no organised screening programme, no reliable recall, no accessible treatment — are the actual cause of most diabetic blindness, not the biology of the disease.

How does GMJ News cover diabetic retinopathy?

GMJ News covers global burden and projections, the asymptomatic nature of sight-threatening disease, national screening programme design and impact, international retinopathy grading and the 4-2-1 rule, diabetic macular oedema and the shift from laser to anti-VEGF (DRCR Protocol T), panretinal photocoagulation and Protocol S, vitrectomy indications, DCCT/UKPDS control evidence and the early-worsening phenomenon, retinopathy in pregnancy and screening schedules, and AI-based retinal screening with its real-world caveats. Directly relevant to PHIG's diabetes, eye health and screening programme work.

Home › Topics › Diabetic Retinopathy

Diabetic Retinopathy

GMJ News knowledge hub · last reviewed September 2026 · Georgian Medical Journal

Summary Statistics Glossary GMJ news FAQ Documents Organizations Research

Diabetic retinopathy — microvascular damage to the retina from chronic hyperglycaemia — is the leading cause of preventable blindness in working-age adults globally, affecting approximately one third of all people with diabetes and around 103 million people worldwide, with numbers projected to rise to 160 million by 2045 as diabetes prevalence increases (WHO). The defining clinical fact is that sight-threatening diabetic retinopathy is almost entirely asymptomatic until vision is already lost — proliferative retinopathy and diabetic macular oedema can be advanced while the patient reads a normal Snellen line — which is why systematic annual retinal screening with digital photography, backed by prompt laser or anti-VEGF treatment, is one of the highest-yield preventive programmes in medicine and the intervention that reduced diabetic retinopathy from the commonest cause of blindness certification in working-age adults in England to a position it no longer holds.

Key messages

Leading cause of preventable blindness in working-age adults — ~103 million affected

Diabetic retinopathy affects approximately one third of all people with diabetes — around 103 million people globally, projected to reach 160 million by 2045. It is the leading cause of preventable blindness in working-age adults worldwide. Duration of diabetes is the strongest predictor: after 20 years, nearly all patients with type 1 and over 60% with type 2 diabetes have some retinopathy.

Sight-threatening retinopathy is ASYMPTOMATIC until vision is already lost

The single most important clinical fact. Proliferative diabetic retinopathy and diabetic macular oedema can be advanced while the patient reads a normal Snellen line and reports no visual symptoms whatsoever. By the time a patient notices blurring, floaters or vision loss, irreversible damage has usually occurred. This is why systematic ANNUAL RETINAL SCREENING — not symptom-triggered referral, and not visual acuity testing — is the entire foundation of preventing diabetic blindness. Patients who say their vision is fine must still be screened.

Screening programmes are among the highest-yield interventions in medicine

Systematic digital retinal photography screening with organised failsafe, grading and referral pathways transformed outcomes: in England, diabetic retinopathy ceased to be the leading cause of certifiable blindness in working-age adults for the first time in at least five decades following the introduction of national screening. Screening is cost-effective in almost every modelled setting. Extended intervals (every 2 years) are now used in some programmes for patients with two consecutive screens showing no retinopathy and good metabolic control — but only within a quality-assured programme with reliable recall.

Diabetic macular oedema — anti-VEGF replaced laser as first-line

DMO (retinal thickening from leakage at the macula) is the commonest cause of visual loss in diabetes. Historically treated with focal/grid macular laser, which stabilised but rarely improved vision. Intravitreal anti-VEGF (aflibercept, ranibizumab, bevacizumab, faricimab) is now first-line for centre-involving DMO with reduced acuity, and IMPROVES vision in a substantial proportion. DRCR Protocol T showed aflibercept superior at worse baseline acuity. Intravitreal steroids (dexamethasone implant, fluocinolone acetonide) are used in anti-VEGF non-responders, pseudophakic eyes, and where injection frequency is impractical — accepting cataract and IOP-rise risks.

Panretinal photocoagulation remains the mainstay for proliferative disease

PRP — scattered laser burns to the peripheral retina — reduces the ischaemic drive producing VEGF, causing neovascular regression. It has prevented severe visual loss for five decades (Diabetic Retinopathy Study, ETDRS) and remains the standard for high-risk proliferative diabetic retinopathy, particularly where reliable follow-up cannot be guaranteed. Anti-VEGF (DRCR Protocol S) achieves comparable or better visual outcomes with less field loss — but only with sustained adherence to injections; loss to follow-up on anti-VEGF monotherapy is catastrophic in a way that completed PRP is not. In many settings PRP is therefore the safer choice.

Glycaemic and blood pressure control alter the trajectory — plus the "early worsening" caveat

DCCT and UKPDS established that intensive glycaemic control substantially reduces retinopathy incidence and progression, with a durable legacy effect. Blood pressure control (UKPDS) is comparably important. Fenofibrate (FIELD, ACCORD-Eye) reduces retinopathy progression independently of lipid effect. Important clinical caveat: rapid intensification of glycaemic control — particularly a large fall in HbA1c over a short period, as with insulin initiation, bariatric surgery or GLP-1 initiation — can cause transient EARLY WORSENING of retinopathy. Patients with existing significant retinopathy should have retinal assessment before and during rapid intensification, not as a reason to avoid control but to monitor and treat.

Key statistics

~103M

people with diabetic retinopathy globally; projected 160 million by 2045

WHO/IDF

~1 in 3

people with diabetes have some degree of diabetic retinopathy

IDF/WHO

Asymptomatic

sight-threatening retinopathy causes NO symptoms until vision is already lost — screen regardless

RCOphth/AAO

Anti-VEGF

replaced macular laser as first-line for centre-involving diabetic macular oedema

DRCR Protocol T

DCCT/UKPDS

intensive glycaemic and BP control substantially reduce retinopathy progression

NEJM/BMJ

Early worsening

rapid HbA1c reduction can transiently worsen retinopathy — assess retina before intensification

DCCT/RCOphth

Diabetic retinopathy — prevalence by diabetes duration

Source: WESDR/IDF. Duration is the strongest predictor; near-universal retinopathy after 20 years in type 1 diabetes.

Glossary of key terms

Retinopathy grading
Ophthalmology/ScreeningInternational Clinical Diabetic Retinopathy scale: No apparent retinopathy. Mild NPDR (non-proliferative): microaneurysms only. Moderate NPDR: more than microaneurysms but less than severe. Severe NPDR (the 4-2-1 rule): severe intraretinal haemorrhages in all 4 quadrants, OR venous beading in 2+ quadrants, OR intraretinal microvascular abnormalities (IRMA) in 1+ quadrant — approximately 50% progress to proliferative disease within one year. PDR (proliferative): neovascularisation of the disc (NVD) or elsewhere (NVE), vitreous or preretinal haemorrhage. Separately graded: diabetic maculopathy/macular oedema, which can coexist with any retinopathy grade and is the commonest cause of visual loss.

Diabetic macular oedema
OphthalmologyRetinal thickening and hard exudate deposition at the macula due to breakdown of the blood-retinal barrier with leakage from microaneurysms and capillaries. Classification: centre-involving (affecting the central subfield — treated with anti-VEGF if acuity reduced) versus non-centre-involving (observed or treated with focal laser). OCT is the diagnostic standard, quantifying central subfield thickness and identifying intraretinal cysts, subretinal fluid, disorganisation of retinal inner layers (DRIL — a predictor of poor visual outcome) and ellipsoid zone disruption. Systemic contributors to DMO that must be addressed alongside ocular treatment: poor glycaemic control, hypertension, fluid overload, renal impairment, sleep apnoea, and glitazone therapy.

Panretinal photocoagulation (PRP)
Laser/OphthalmologyTypically 1,200-2,000 or more scatter burns applied to the peripheral retina, usually over 2-4 sessions, ablating ischaemic peripheral retina to reduce the VEGF drive and induce regression of neovascularisation. Established by the Diabetic Retinopathy Study and ETDRS as reducing severe visual loss by over 50% in high-risk PDR. Side effects, which require explicit consent discussion: peripheral visual field loss (relevant to driving licence standards); reduced night vision; possible worsening of macular oedema (pre-treat or co-treat DMO); and reduced contrast sensitivity. Its decisive practical advantage over anti-VEGF monotherapy is that once completed it is durable and does not depend on the patient returning — a critical consideration in populations with poor follow-up.

Vitrectomy in diabetic eye disease
Vitreoretinal surgeryPars plana vitrectomy is indicated for: non-clearing vitreous haemorrhage (earlier in type 1 diabetes and in bilateral cases); tractional retinal detachment involving or threatening the macula; combined tractional-rhegmatogenous detachment; dense premacular subhyaloid haemorrhage; and severe fibrovascular proliferation despite PRP. Anti-VEGF given a few days preoperatively reduces intraoperative bleeding and simplifies membrane dissection — but must not be given too far in advance, as it can precipitate crunch (acute contraction of fibrovascular tissue causing tractional detachment). Outcomes are substantially better when PRP has been completed and when surgery is not delayed until the retina is chronically detached.

AI-based retinal screening
Technology/ScreeningAutonomous and assistive AI systems for diabetic retinopathy screening are among the most mature clinical AI applications: several have regulatory clearance (including FDA De Novo authorisation for autonomous detection), and validation studies report sensitivity and specificity comparable to human graders for referable retinopathy. Their potential value is greatest exactly where the burden is greatest — settings with insufficient trained graders and ophthalmologists. Practical caveats that determine real-world success: performance varies with camera type, image quality, pupil dilation and population; ungradable image rates matter as much as accuracy; algorithmic performance can degrade across ethnic groups not represented in training data; and an AI result is useless without a functioning referral, treatment and failsafe pathway behind it.

Diabetic retinopathy in pregnancy
Obstetrics/OphthalmologyPregnancy is an independent risk factor for retinopathy progression, and progression can be rapid. Contributing factors: hormonal and haemodynamic changes; and, importantly, the rapid intensification of glycaemic control typically undertaken at the start of pregnancy, which carries the early-worsening effect. Guidance: women with pre-existing type 1 or type 2 diabetes should have retinal assessment at pre-conception counselling, at booking (first trimester), and again at 28 weeks — with additional assessment at 16-20 weeks if any retinopathy is present at booking. Gestational diabetes does NOT require retinal screening. Progression during pregnancy frequently regresses partially in the year after delivery, and pregnancy is not a contraindication to necessary laser treatment.

Latest GMJ coverage

NEJM Images Show Advanced Diabetic Eye Disease Progression and Treatment Response

13/06/2026

Diabetic Retinopathy

Key messages

Key statistics

Glossary of key terms

Latest GMJ coverage

Frequently asked questions 12 Q&A — structured for Google featured snippets and AI discovery

Knowledge hub: guidelines, conventions and reports

Organizations working in migration and health

Related health topics

Company