Diabetic Retinopathy Classification: ETDRS, ICDR and Exam Traps

The first rule in a diabetic-retinopathy answer is to name the system you are using. The International Clinical Diabetic Retinopathy (ICDR) severity scale is a practical five-step clinical communication scale. ETDRS is a much more granular, photograph-based research severity scale. They are related, but their labels are not interchangeable.¹² This article is examination education for ophthalmologists and trainees, not screening, diagnostic or treatment advice for an individual patient.

The direct answer: ICDR before ETDRS

For most exam stems that ask “grade this retinopathy”, start with ICDR unless the stem explicitly invokes ETDRS levels, standard photographs or a trial. The ICDR scale makes one thing especially clear: PDR is defined by neovascularisation and/or preretinal or vitreous haemorrhage. Severe NPDR can have extensive intraretinal lesions, but it has no PDR feature.¹

The International Council of Ophthalmology paper that introduced the scale uses the familiar 4-2-1 rule for severe NPDR: haemorrhages and/or microaneurysms in all four quadrants, venous beading in at least two quadrants, or IRMA in at least one quadrant, provided PDR is absent.¹ In an SBA, read “any one of these” slowly. It is not a score that requires all three findings.

A lesion-to-grade sequence for photographs and vignettes

Do not begin with visual acuity or OCT when asked to grade retinopathy from a fundus image. Use this sequence.

1. Find PDR first. Look for NVD/NVE, preretinal haemorrhage or vitreous haemorrhage. If present, say PDR and then describe the lesion/location.
2. If no PDR, decide whether the lesions meet 4-2-1. Count quadrants only when the image/stem gives enough information.
3. If not severe, ask whether microaneurysms are the only lesion. If yes, mild NPDR.
4. If there are more lesions but not 4-2-1, use moderate NPDR.
5. Classify macular oedema separately. Retinopathy severity and macular thickening are not the same axis.

This method stops a common viva failure: describing a hard exudate or reduced acuity and jumping to “proliferative”. Hard exudates indicate chronic leakage; they do not by themselves establish retinal neovascularisation. Likewise, cotton-wool spots support ischaemic microvascular disease but do not automatically make the eye severe NPDR without the defined pattern.

Why IRMA is the recurrent trap

Intraretinal microvascular abnormalities (IRMA) are abnormal intraretinal vascular channels in areas of capillary non-perfusion. Neovascularisation grows on or through the retinal surface, commonly at the disc (NVD) or elsewhere (NVE), and has different clinical/angiographic behaviour. In an image question, an examiner may expect you to use location and morphology rather than a memorised sentence. If the stem says IRMA in one or more quadrants, it can satisfy one severe-NPDR 4-2-1 pathway; it does not itself make the diagnosis PDR.¹

The same discipline applies to venous beading. It is an important severe-NPDR component when present in two or more quadrants, not merely a decorative description of a tortuous vessel. Describe exactly what is shown and avoid upgrading a grade from a vague impression.

What ETDRS adds, and what it does not

The Early Treatment Diabetic Retinopathy Study (ETDRS) developed detailed photographic grading and severity scales for trial work. It is the source of much of the language in older landmark diabetic-macular-oedema literature and helps explain why retina papers can feel more granular than a clinic classification table.² ETDRS uses standardised fields and lesion-level comparisons; its severity levels should not be improvised from a casual one-line description.

For an exam answer, a safe distinction is:

“ICDR is the concise clinical severity scale I would use to communicate this fundus grade. ETDRS is a more detailed photographic research scale; I would use its precise level only when the image protocol and criteria support it.”

That answer demonstrates knowledge without inventing an ETDRS number. It also handles questions that deliberately compare “clinical classification” with “study grading”. A good candidate knows the direction of translation without pretending that every ICDR label has a one-to-one ETDRS equivalent.

Macular oedema terminology: say what the term was built for

Clinically significant macular oedema (CSME) is an ETDRS clinical-trial term, defined using retinal thickening and/or adjacent hard exudates in relation to the foveal centre. The landmark ETDRS report used it in a pre-OCT era and tied it to the study’s laser-treatment question.³ It remains a high-yield historical and examination term.

Contemporary literature often uses diabetic macular oedema (DMO/DME) and describes whether it is centre-involving, commonly using OCT. “Centre-involving DMO” and “CSME” therefore should not be used as automatic synonyms. One is a modern anatomical/imaging descriptor; the other is a specific historical clinical definition. When the stem gives OCT, state the compartment and centre involvement as described. When it gives ETDRS wording, use CSME accurately. Local protocols determine real clinical decisions; this article does not attempt to reproduce them.

A disciplined image and viva answer

Use seven sentences, in order:

1. State laterality and image adequacy.
2. Describe microaneurysms, haemorrhages, exudates, cotton-wool spots, venous changes and IRMA.
3. State whether there is NVD/NVE or preretinal/vitreous haemorrhage.
4. Apply ICDR grade with the criterion that earned it.
5. Describe macular involvement separately, including OCT evidence if supplied.
6. Name the imaging limitation or needed correlation when relevant.
7. Stop before making a patient-specific management prescription.

For example: “This is a gradable right-eye photograph with haemorrhages/microaneurysms in all four quadrants and no neovascularisation shown, so it meets ICDR severe NPDR by the four-quadrant criterion. Macular oedema cannot be confirmed or excluded from this photograph alone.” That last sentence often earns more credit than a guessed treatment.

Ten classification MCQs with explanations

1. Microaneurysms only =? Mild NPDR. Why: that is the exact ICDR mild criterion.¹
2. Haemorrhages in four quadrants, no NVD/NVE =? Severe NPDR. Why: one 4-2-1 pathway is sufficient.
3. Venous beading in two quadrants, no PDR feature =? Severe NPDR. Why: it meets the “2” limb.
4. IRMA in one quadrant with no NVD/NVE =? Severe NPDR. Why: it meets the “1” limb, not PDR.
5. New vessels elsewhere =? PDR. Why: NVE is neovascularisation.
6. Preretinal haemorrhage with no visible NVE =? PDR. Why: ICDR includes preretinal/vitreous haemorrhage.¹
7. Hard exudates alone make PDR? No. Why: they indicate leakage history, not neovascularisation.
8. ETDRS is best described as? A detailed photographic research severity system. Why: avoid treating it as a simple five-row chart.²
9. CSME and centre-involving DMO are identical terms? No. Why: they arise from different clinical/imaging frameworks.³
10. Visual acuity grades NPDR severity? No. Why: acuity is clinically important but does not replace retinal lesion classification.

Ten more “exam trap” prompts

11. Does severe NPDR mean PDR is imminent in every individual? No. Why: severity communicates risk, not a deterministic time course.
12. Can an eye have PDR with relatively little background haemorrhage? Yes. Why: classify what is present; do not demand a staged visual narrative.
13. Does absence of visible NVD exclude NVE? No. Why: inspect disc and elsewhere.
14. Does IRMA cross into the vitreous? No; it is intraretinal. Why: this distinguishes it conceptually from surface neovascularisation.
15. Is “moderate NPDR” a failure to classify? No. Why: it is an intentional ICDR category between mild and severe.
16. Can a single non-macular photograph classify DMO fully? No. Why: macular assessment needs appropriate macular information.
17. What should you say if the image is ungradable? That grading is limited by image quality. Why: never fabricate lesions from blur.
18. Do standard ETDRS photographs equal a routine fundus-camera image? No. Why: the system’s precision depends on standardised photographic grading.
19. Does the word “proliferative” refer to exudate proliferation? No. Why: it refers to pathologic retinal/disc neovascularisation.
20. Most useful final line after the grade? State macular status separately and the modality used. Why: it avoids category mixing.

A 30-minute revision drill

Draw the ICDR table from memory in three minutes. Spend ten minutes looking at labelled images in an approved reference and force yourself to describe lesions before naming grades. Then answer the twenty prompts without notes. For every error, write only one correction: “system error”, “4-2-1 error”, “IRMA/NV error”, or “macular terminology error”. Re-test the same error labels in a mixed retina block. The goal is accurate language under time pressure, not merely recognising a familiar photograph.

Worked classification stems

Stem 1: the tempting “moderate” answer

A photograph shows multiple blot haemorrhages and microaneurysms in every quadrant. No NVD, NVE, preretinal haemorrhage or vitreous haemorrhage is described. Answer: severe NPDR by the four-quadrant haemorrhage/microaneurysm component of 4-2-1. It is tempting to call this moderate because the examiner did not mention venous beading or IRMA. That is exactly the trap: one qualifying limb is enough.¹

Stem 2: the tempting “PDR” answer

An ultra-widefield image is described as having IRMA temporally and venous beading in two quadrants, with no neovascularisation. Answer: severe NPDR. IRMA is severe disease but remains intraretinal. Calling it PDR confuses degree of ischaemic retinopathy with the separate event of neovascularisation.

Stem 3: the terminology question

A historical trial question gives retinal thickening within a specified distance of the foveal centre and asks for its named criterion. Answer: use CSME if the stated ETDRS clinical definition is met. Do not replace it with “centre-involving DMO” just because current OCT-based practice uses that phrase. The wording signals which framework the examiner expects.³

Stem 4: insufficient evidence

A non-stereoscopic photograph shows exudates close to the macula but no OCT and no clear information about retinal thickening. Answer: grade retinopathy from visible lesions but do not conclusively classify oedema from inadequate macular information. Naming uncertainty is better than inventing a centre-involving diagnosis.

How to build an image notebook

On the left page, paste or sketch a labelled image from a legitimate teaching source. On the right, write only: lesions, quadrants, PDR present/absent, ICDR grade, macular information available, and one possible distractor. Do not write treatment plans. After ten cases, shuffle the right pages and identify the grade from lesion lists alone. Then reverse it: look at the image for sixty seconds and dictate the seven-sentence answer structure. This drills both recognition and language, which are different examination tasks.

Why the systems are not competitors

ICDR deliberately trades detail for reproducible communication. ETDRS preserves more lesion-level and photographic detail for research-grade categorisation. In a real paper, the selected system reflects the question being asked; in an exam, the selected system reflects the wording. “ETDRS versus ICDR” is therefore not a request to choose a winner. It is a request to show that you understand scope, granularity and the limits of translation.¹²

One practical safeguard follows from that distinction: never write a mixed label such as “ETDRS severe NPDR, ICDR level 53” unless the source image and question explicitly provide the conversion context. Give the requested system first, then add the other only as a qualified comparison. This small discipline prevents a technically knowledgeable answer from becoming internally inconsistent.

In a single-best-answer paper, underline the system named in the stem before inspecting the options. It is a simple action that avoids a preventable terminology error.

For broader retina revision, use the verified retina study guide. Retina Deciphered and the site’s free high-yield MCQs are existing OphthaMCQ resources; they are not official diabetic-retinopathy guidance.

Sources

Footnotes

1. Wilkinson et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

2. ETDRS Report 10: Grading diabetic retinopathy from stereoscopic colour fundus photographs. ↩ ↩² ↩³ ↩⁴

3. ETDRS Report 1: Photocoagulation for diabetic macular edema. ↩ ↩² ↩³