• Home
• Slide Presentations
• Publications
• Results FAQs
• Study Information
• Study Design
• Retinal Photography
• Baseline Characteristics
• Diabetic Retinopathy
• Committee Members
• Contact
• Information for Journalists

Retinal Photography and the assessment of Diabetic Retinopathy in the DIRECT Programme

Retinal photography
The gold standard for the assessment of diabetic retinopathy from static images is based on 7-field stereo photography (1). Fundus photographs are taken of seven overlapping standard fields (Figure 1).

Figure 1. 7-field 30° stereoscopic fundus imaging

a. Right eye		.b Left eye

The closed circle indicates the location of the optic disc. The cross indicates the location of the fovea.

In the DIRECT Programme film technology was used rather than digital imaging because access to high-quality digital imaging was not widely available at the time the Programme was designed. Film fundus photography has the advantage over digital photography of consistency; there is much more variation in digital imaging equipment and the inages produced by digital fundus cameras. In the DIRECT Programme one brand and type of slide transparency film was supplied to all the clinical centres and the films were processed in one of only two processing laboratories. All the participating clinical centres received training on the 7-field stereophotography technique.

Retinopathy assessments
Diabetic retinopathy was assessed according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) scale (2). The graders were masked to the identity of the patient for all assessments. The seven sets of photographs per eye were assessed as stereo pairs using viewing glasses to create a stereoscopic view and depth perception. Grading was via a multi-stage process:

• Photographs were primarily assessed by a senior assessor for quality and adherence to protocol, and a maximum ETDRS retinopathy level was allocated.
• A detailed grader next assessed the presence and severity of every lesion within each of the fields and summarised the overall severity of the lesions within that eye.
• If the detailed assessment result differed from the primary assessment, the photograph was reviewed.
• Independent adjudication occurred in the case of persistent differences in assessment.
• Finally, an ETDRS Retinopathy Level was allocated for each eye.

Features graded in the seven stereo fields included:

• Microaneurysms
• Haemorrhages
• Hard exudates
• Cotton wool spots
• Venous calibre abnormalities
• Venous sheathing
• Perivenous exudates
• Arteriolar abnormalities
• Intraretinal Microvascular Abnormalities (IRMA)
• Arteriovenous nicking
• Fibrous proliferation
• Retinal elevation
• Neovascualrisation
• Pre-retinal haemorrhage
• Vitreous haemorrhage
• Scars of photocoagulation
• Macular oedema area
• Macular oedema thickness

The overall grading score for all of the features was matched with severity of retinopathy as shown in table 1.

Table 1. The ETDRS scale and severity of retinopathy

ETDRS retinopathy level	Severity of retinopathy
10	None
20	Microaneurysms only
35	Mild NPDR
43	Moderate NPDR
47	Moderately severe NPDR
53a-d	Severe NPDR
53e	Very severe NPDR
61	Mild PDR
65	Moderate PDR
71,75	High-risk PDR
81,85	Advanced PDR

NPDR=non-proliferative diabetic retinopathy
PDR=proliferative diabetic retinopathy

Retinopathy status at each visit (after unmasking of patient identity) was reported to clinical centres as Right eye/Left eye summarised levels e.g. “mild to moderate non-proliferative”. Actual numerical ETDRS Levels and subcodes and clinically significant macular oedema (CSMO), reported on a numerical scale of 0–4, were used for data analysis. Worse eye/better eye retinopathy levels were used to analyse changes for the study endpoints.