Population screening for diabetic retinopathy without mydriasis: a scoping review

Next update: Update not planned

Contact for queries and feedback

Introduction

Diabetic retinopathy is a common complication for people with diabetes. With rising diabetes levels in Wales, the prevalence of diabetic retinopathy and related health problems is expected to increase, placing more demand on hospital eye services.

Diabetic Eye Screening Wales aims to detect retinopathy early to prevent sight loss. Currently, screening involves dilating the pupils with eye drops before taking retinal images. While this improves image quality, it can cause temporary blurred vision, making driving inadvisable and creating anxiety for some patients. This may act as a barrier to attendance.

This review was designed to inform the evaluation of Topcon NW500 non-mydriatic cameras, which capture retinal images without pupil dilation. The review assessed whether these or similar devices can provide images of sufficient quality for grading diabetic retinopathy and identified factors linked to poor image quality, to guide evaluation in Wales.

Main points

How up to date is this evidence?

We produced this scoping review in 2025 and searched for evidence published between 2015 and July 2025.

What we found

• We included 15 studies in the review: eight on diagnostic accuracy and nine on factors affecting image quality.
• No studies evaluated the Topcon NW500 specifically, and none were conducted in Wales or the UK.
• Diagnostic accuracy of non-mydriatic 45° fundus cameras varied widely. Only one study met British Diabetic Association thresholds for sensitivity (≥80%) and specificity (≥95%), but methodological concerns limit confidence.
• Considerable variation in study aims, camera models, imaging protocols, and reference standards reduces comparability and generalisability.
• Evidence on factors linked to ungradable images was limited:
  • Older age was most consistently associated with poor image quality.
  • Single studies suggested associations with cataract, high spherical equivalent, reduced visual acuity, male sex, ethnic minorities (excluding white minorities), smoking, and high service throughput.
• Most studies had quality concerns

What this means

• There is no evidence on the diagnostic accuracy of the Topcon NW500, so a robust local evaluation is justified.
• Findings from other non-mydriatic cameras show wide variation in accuracy, reinforcing the need for local testing.
• Characteristics linked to poor image quality were mostly single-study findings and should be applied cautiously. However, they may help identify which groups to focus on in the evaluation.
• Collecting these characteristics during evaluation could help identify patients who may still need pupil dilation, improving service efficiency and patient convenience.

Technical information

We used scoping review methods, including comprehensive searches of databases and grey literature, to identify, assess, and summarise evidence for the following questions:

1. Can non-mydriatic retinal imaging using the Topcon NW500 or similar devices produce images of sufficient quality for grading diabetic retinopathy without pupil dilation?
2. What factors are associated with poor-quality non-mydriatic retinal images?

Glossary

Mydriasis: The process of dilating (enlarging) the pupil using eye drops to improve image quality during eye screening.

Non-mydriatic: Means “without dilation.” Refers to taking retinal images without using eye drops to enlarge the pupil.

Fundus camera: A specialised camera used to take pictures of the back of the eye (retina) to check for eye diseases.

Collaboration details

This is an internal Public Health Wales collaborative review between the Evidence Service and Screening Services division, commissioned by Diabetic Eye Screening Wales.