Diabetic Retinopathy: At What Stage Does It Need Treatment, and What Is the Difference Between Laser, Injections, and Surgery?

Research-informed explainer — last updated April 12, 2026

Diabetic retinopathy affects approximately 93 million people worldwide and is the leading cause of new blindness among working-age adults — but in most cases, vision loss is preventable if the disease is caught and treated before it reaches advanced stages. The choice between laser photocoagulation, anti-VEGF injections, and vitrectomy surgery depends on which stage is present, whether the macula is involved, and whether abnormal new blood vessels have already begun to grow.

This article draws on research from five specialist physicians with direct contributions to diabetic retinopathy science. Tien Wong, M.D., at Houston Methodist Hospital, is co-author of the global DR prevalence study (4,798 citations), the definitive Lancet clinical review of diabetic retinopathy (3,016 citations), a landmark AI deep learning detection paper in JAMA (2,225 citations), and the 2021 burden projection through 2045 (2,023 citations). Susan Bressler, M.D., at Johns Hopkins Bayview Medical Center, led analysis of the DRCR.net Protocol I trial that established anti-VEGF as superior to laser for center-involving diabetic macular edema (1,425 citations) and published on persistent macular thickening after anti-VEGF (314 citations). Joan Miller, M.D., at Massachusetts General Hospital and Massachusetts Eye and Ear, published the foundational vitreous VEGF study in proliferative DR (1,290 citations) — the mechanistic discovery explaining why anti-VEGF drugs work in the first place. Jeffrey Heier, MD, at Ophthalmic Consultants of Boston, was an investigator on the VISTA and VIVID aflibercept for DME trials (792 and 609 citations respectively). Theodore Leng, MD, Professor of Ophthalmology at Stanford's Byers Eye Institute, published the automated deep learning DR detection system (1,307 citations), which is reshaping how screening reaches patients in primary care settings.

How diabetic retinopathy develops

Chronically elevated blood sugar damages the small blood vessels of the retina, causing them to leak, become blocked, or — in advanced disease — trigger the growth of fragile new vessels that bleed and scar. The process has two main phases:

Non-proliferative diabetic retinopathy (NPDR): Ranges from mild (a few microaneurysms visible on exam) through moderate to severe. In severe NPDR, widespread vascular damage and ischemia create the biological conditions that drive the next phase. Diabetic macular edema (DME) — swelling of the central retina due to leaking vessels — can occur at any stage of NPDR and is the most common cause of vision loss in diabetic patients.

Proliferative diabetic retinopathy (PDR): The retina responds to ischemia by releasing VEGF, which triggers growth of new, abnormal blood vessels (neovascularization) on the retinal surface and into the vitreous. As Dr. Miller's 1994 paper first demonstrated, VEGF levels in the vitreous humor of eyes with proliferative DR are dramatically elevated compared with non-diabetic eyes — establishing the molecular rationale for anti-VEGF treatment.

Who is at risk and how widespread is this?

Dr. Wong's global prevalence study found approximately 93 million people with DR, 17 million with proliferative DR, and 21 million with DME worldwide as of 2010 data. Longer duration of diabetes is the strongest risk factor — after 20 years of type 1 diabetes, nearly all patients have some degree of retinopathy. Poor glycemic control and elevated blood pressure are the most important modifiable risks. The 2021 burden projection estimated that 160.5 million people will have DR by 2045 as the global diabetes epidemic continues to grow.

Treatment by stage

Mild to moderate NPDR: Observation with annual dilated eye exams. No ocular treatment is indicated. The most important intervention is systemic: tighter blood sugar control (target HbA1c below 7%) and blood pressure management can slow progression and, in some cases, cause partial regression.

Severe NPDR: Panretinal photocoagulation (PRP) laser is traditionally recommended to reduce the risk of progression to PDR. Evidence now also supports anti-VEGF injections as an alternative that may be superior for some patients, particularly those who develop PDR. The DRCR.net Protocol W trial found that preventive anti-VEGF treatment in eyes with severe NPDR reduced the risk of developing vision-threatening PDR.

Diabetic macular edema with center involvement: Anti-VEGF injections are first-line treatment. Dr. Bressler's Protocol I analysis was pivotal: ranibizumab with prompt or deferred laser was superior to laser alone, with mean visual acuity gains of approximately +9 letters at one year versus +3 letters with laser. The VISTA/VIVID trials for aflibercept (in which Dr. Heier participated) showed gains of +10.7 letters with every-two-month dosing versus +0.2 letters with laser at one year. Importantly, Dr. Bressler's follow-up work on persistent macular thickening found that approximately 40% of DME patients still have residual fluid after multiple injections of any anti-VEGF agent — a finding that guides decisions about switching drugs or adding a steroid implant.

Proliferative diabetic retinopathy: Panretinal laser photocoagulation destroys peripheral ischemic retina, reducing VEGF stimulus. Anti-VEGF injections can treat PDR directly, reduce neovascularization before surgery, and manage coexisting DME. The choice depends on whether DME is present, patient adherence capacity, and whether the vitreous is clear enough for laser delivery.

Tractional retinal detachment or vitreous hemorrhage: Vitrectomy surgery removes hemorrhage, releases tractional membranes pulling on the retina, and allows completion of PRP. Recovery from vitrectomy takes weeks to months; visual outcomes depend on how long the retina has been detached and whether the macula is involved.

Comparison table

The role of AI in screening

Dr. Leng's automated deep learning system published in 2017 achieved sensitivity and specificity for referable DR comparable to trained ophthalmologists when reading retinal photographs. Dr. Wong led a parallel JAMA study validating a Google deep learning system across multiethnic populations in Singapore and the United States. These systems are now being deployed in primary care offices and diabetes clinics, where patients can be screened during routine visits without needing to see an ophthalmologist — dramatically expanding the ability to catch DR before it reaches a treatment-requiring stage.

Questions to ask your doctor

• What stage of diabetic retinopathy do I have, and what is my risk of progressing to the next stage?
• Do I have center-involving diabetic macular edema, and if so, which anti-VEGF drug do you recommend?
• How often do I need injections, and for how long?
• If I have not responded to one anti-VEGF drug, what is the next step?
• Is panretinal laser photocoagulation still indicated for my level of disease, or would you treat with injections instead?
• What HbA1c and blood pressure targets are most important for protecting my eyes?

The bottom line

Diabetic retinopathy progresses silently through stages, and treatment thresholds are well-defined: center-involving DME requires prompt anti-VEGF injections, severe NPDR warrants preventive intervention, and proliferative DR demands aggressive treatment to prevent catastrophic hemorrhage or detachment. Anti-VEGF injections have largely replaced laser as first-line therapy for macular disease, delivering consistently better visual outcomes — but laser and vitrectomy remain essential tools for advanced-stage disease. Annual eye exams are not optional for people with diabetes; they are the primary defense against the blindness that this condition can otherwise cause silently.