Diabetic eye screening every 2 years: the evidence, and who waits

In North East London's records, a 2-year gap between diabetic eye screens would have pushed 39% of sight-threatening diagnoses in people with 2 clear screens back a screen, a simulation published in Diabetologia found. Since October 2023, England has been moving people whose last 2 routine screens found no retinopathy onto diabetic eye screening every 2 years instead of every year.

It follows a 2016 recommendation from the UK National Screening Committee, and Scotland, Wales and Northern Ireland made the change before England. NHS England tells people on the longer interval to see their optometrist straight away if their sight changes.

Who is on diabetic eye screening every 2 years

In England it takes no retinopathy and no maculopathy in either eye (R0M0) at the last 2 routine screens. One microaneurysm makes a screen R1, and R1 alone means yearly screening again, while maculopathy (M1) means OCT surveillance. Pregnant women are screened more often. A review in Eye Open adds that people with type 1 diabetes starting a hybrid closed loop system are exempt, those starting an insulin pump stay yearly for a year, and the change has been staggered since a 2024 software upgrade.

What the evidence behind the change showed

The UK National Screening Committee relied on the Four Nations study, published in Diabetes Care. It pooled 2005 to 2012 grading results for 354,549 people in 7 programmes: the national ones of Scotland, Wales and Northern Ireland, and 4 in England. Of people with no retinopathy at 2 screens at least 12 months apart, 0.3% to 1.3% progressed to referable retinopathy within 2 years, and fewer than 0.3% to proliferative disease. With background retinopathy in both eyes at both screens, the figures were 13% to 29% and up to 4%.

It was retrospective, and its authors wrote that it could not directly show the effect of changing the interval. The nearest test is Liverpool's ISDR trial, which randomised 4,534 people to yearly screening or to 6, 12 or 24 months set by a risk calculator using clinical data. Attendance at first follow-up was equivalent (83.6% against 84.7%), sight-threatening retinopathy was found in 1.4% against 1.7% over 2 years, and the individualised arm needed 43.2% fewer appointments.

What London's records added

Pitt and colleagues took North East London's records from 2012 to 2023, when everyone was still invited yearly, and found 105,083 people with 2 clear screens. Over a mean of 5 years, 3,694 developed sight-threatening retinopathy, meaning a screening grade of R2, R3 or M1. The paper does not say how many were proliferative. Replaying the records on a 2-year cycle cut appointments by 54%, but 1,444 of those diagnoses, 39%, would have come a screen later.

The risk was uneven before the interval changed. Against the lowest-risk group, white people aged 41 or over with diabetes for under 10 years, almost every group was significantly more likely to have a sight-threatening diagnosis at an appointment: South Asian people 1.61 to 3.12 times, Black people 2.27 to 4.10, highest at 40 or under. The ratios were much the same with yearly screening, but a 2-year gap lets more build up: for Black people, the chance at each appointment rose from 0.8% to 2.4%, against 1% for the lowest-risk group. A points score using age, diabetes duration and ethnicity cut delayed diagnoses to 25% while saving 43% of appointments, and from 34% to 22% when tested on South East London's records.

Why it matters in practice

The College of Optometrists' guidance covers the year between screens:

• Ask whether the patient is in a screening programme and when they last attended, and say if you think it is overdue.
• Keep sight-test recall as for someone without diabetes, whatever their screening interval, but use your judgement for younger people and people of Black and Asian ethnicities.
• Dilating solely to look for retinopathy is not needed where a screening programme is in place.
• Report relevant findings to whoever is responsible for the patient's care.

Beyond the guidance, record the result of the last screen as well as its date. Background retinopathy in someone on the 2-year cycle is a relevant finding to report, though a report does not by itself change their screening interval. Younger patients may matter twice: in London, people aged 40 or under were particularly at risk beyond 3 years, and a September review of 14 UK studies from 2006 to 2021 found younger age, in bands such as 12 to 39, and deprivation the most consistent markers of missed or late screening.

The caveats

• The Four Nations study used 2005 to 2012 data, and its authors said the spread between programmes may partly reflect grading.
• London's 39% is a simulation on yearly-screening records, so it cannot show how people respond to fewer invitations.
• ISDR tested a clinical-data calculator, not England's rule, in one programme whose participants were 94.6% white, over 2 years.
• The points score needs recorded ethnicity, which some programmes hold incompletely.
• London's funders were Wellcome, the Health Foundation, the NHS Transformation Directorate and the NIHR. Its authors also research automated grading, covered on our news page.

Sources

1. 1.A simple score-based strategy to improve equity of the UK biennial diabetic eye screening protocol among people deemed as low risk · Diabetologia, 12 March 2025
2. 2.Progression of diabetes retinal status within community screening programs and potential implications for screening intervals · Diabetes Care, 18 December 2014
3. 3.Safety and cost-effectiveness of individualised screening for diabetic retinopathy: the ISDR open-label, equivalence RCT · Diabetologia, 4 November 2020
4. 4.Diabetic eye screening intervals extended for people at lowest risk · NHS England (GOV.UK), 27 September 2024
5. 5.Updates in the English Diabetic Eye Screening Programme · Eye Open, 9 February 2026
6. 6.Examining patients with diabetes mellitus (Guidance for Professional Practice) · College of Optometrists
7. 7.Diabetic eye screening attendance among minority ethnic and other inequality-affected populations in the United Kingdom: a systematic review · Vision, 4 September 2026