Educational content. This page summarises a published consensus document for clinician education. It is not medical advice and not a protocol. Diagnosis and treatment decisions require a qualified clinician with the full clinical picture and local guidelines.
Why a fifth definition
The Universal Definition of Myocardial Infarction (UDMI) has been rewritten every six or so years since 2000. The Third UDMI in 2012 introduced the numbered types [3, 4]. The Fourth UDMI in 2018 split myocardial injury from infarction, added sex-specific troponin limits, and expanded to seven subtypes [5, 6].
The numbers worked well in research but poorly at the bedside. Three problems stood out in the 2026 task force's review [1]:
- Type 2 MI was a mixed bag. It grouped SCAD, embolism, and vasospasm (acute coronary pathologies) with tachyarrhythmia and anaemia (supply and demand problems). These patients need different tests and different treatment.
- Type 2 MI was often made without imaging, on symptoms or ECG changes alone. For many patients the label had no treatment consequence.
- Procedural thresholds were arbitrary. After cardiac surgery, 97.5% of patients exceed 10x the upper reference limit, so the type 5 threshold was nearly universal. A threshold of about 218x the upper reference limit (95% CI 40 to 318) was needed to identify higher peri-procedural mortality [1].
The Fifth UDMI, a joint ESC, ACC, AHA, and WHF statement published simultaneously in Circulation, JACC, European Heart Journal, and Global Heart, replaces the numbers with three clinical settings [1, 2].
Loading Fifth Universal Definition of MI (2026)...
The core definition
The skeleton is the same as 2018. MI requires acute myocardial injury plus at least one of the following [1]:
- Symptoms or other evidence of acute myocardial ischaemia
- New ischaemic ECG changes or pathological Q waves
- Imaging evidence of an acute coronary pathology
- New loss of viable myocardium or new regional wall motion abnormality in an ischaemic pattern
Acute myocardial injury is a rise and/or fall in cTnI or cTnT with at least one value above the sex-specific 99th percentile upper reference limit. The upper reference limit is the troponin value that 99% of a healthy reference population falls below, so anything above it counts as myocardial injury. The Fourth UDMI recommended sex-specific limits; the Fifth writes them into the definition, to avoid systematic under-recognition of injury and infarction in women [1].
Three settings replace seven types
| Primary MI | Secondary MI | Procedure-related MI | |
|---|---|---|---|
| Setting | Spontaneous, no other trigger | Another acute condition causes supply and demand mismatch | Coronary complication within 30 days of any cardiac procedure |
| Includes | Atherothrombosis, SCAD, embolism, vasospasm, and restenosis, stent thrombosis, or graft failure after 30 days | Tachyarrhythmia (most common), hypotension, hypoxia, anaemia, severe hypertension | PCI, CABG, TAVI, valve surgery, ablation |
| Design goal | Sensitivity | Specificity | Specificity |
| Confirmed by | Acute coronary pathology on angiography or intracoronary imaging, or new RWMA or loss of viability | Obstructive CAD (70% or more, or 50% or more and flow-limiting) without acute pathology, or new RWMA or loss of viability | Angiographic complication or new RWMA in that territory (both, if it arose during the procedure or the procedure was for AMI) |
| Old equivalent | Type 1, coronary part of type 2, late 4b and 4c | Imbalance part of type 2 | 4a, early 4b and 4c, 5 |
Source: Fifth UDMI boxes 3 to 5 and Table 1 [1].
Primary MI
Primary MI is simply what most clinicians mean by "MI". The word primary only distinguishes it from the other two settings, and in practice it can be called myocardial infarction [1].
The key change is scope. Every acute coronary pathology now counts: atherothrombosis, SCAD, coronary embolism, epicardial or microvascular vasospasm, and stent or graft failure more than 30 days after the procedure. The task force argues these patients follow the same initial pathway as atherothrombosis, so they belong together, and coronary angiography (with IVUS or OCT when needed) should define the mechanism. Reported frequencies vary widely: SCAD 0.1 to 4%, embolism 0.2 to 13%, epicardial vasospasm 1 to 20% [1].
When angiography and imaging are not available or not appropriate, primary MI can still be the final diagnosis on clinical, ECG, and troponin grounds, coded as "aetiology unknown".
Secondary MI
This is the biggest practical change for the ICU, the wards, and the emergency department. Under the Fourth UDMI, a patient with fast AF, a troponin rise, and some ST depression was routinely labelled type 2 MI. Under the Fifth UDMI that patient has acute myocardial injury unless imaging shows one of two things [1]:
- Obstructive CAD (70% or more stenosis, or 50% or more if flow-limiting) without an acute coronary pathology, or
- A new or presumed new regional wall motion abnormality, or loss of viable myocardium, in an ischaemic pattern
If imaging shows an acute coronary pathology, the diagnosis becomes primary MI. Sepsis and bleeding can trigger atherothrombosis, so primary MI stays on the differential.
The task force cites a prospective cohort of old type 2 MI patients who had systematic angiography: about 1 in 20 had atherothrombosis, and most had neither obstructive CAD nor new LV impairment. Applying the new criteria reduced the number of MI diagnoses and identified those at highest risk [1].
Timing is pragmatic. Invasive angiography is indicated when ischaemia persists or recurs. Otherwise imaging can wait until the acute illness is treated, and CCTA after recovery is reasonable. In frail patients or those with limited life expectancy, clinical judgment is acceptable; known CAD or injury large enough to affect cardiac function makes secondary MI more likely [1].
Procedure-related MI
One rule now covers every cardiac procedure, percutaneous or surgical, within 30 days [1]:
- The diagnosis is considered with symptoms, ischaemic ECG changes, Q waves, or unexplained deterioration under sedation, plus acute injury.
- It is confirmed by an angiographic coronary complication (side branch occlusion, dissection, no-reflow, stent thrombosis, graft failure, ostial obstruction during valve intervention) or a new RWMA in the territory involved.
- If the complication arose during the procedure, or the procedure was done for acute MI, both are required. Transient complications fixed on the table do not count by themselves.
No troponin multiple makes the diagnosis. Values above 5x the upper reference limit and still rising at 6 h after intervention, or above 35x the upper reference limit and rising at 24 h after surgery, are suggested as prompts to image, and the task force states these are arbitrary and unvalidated [1]. Troponin normally peaks 3 to 12 h after cardiac surgery; a peak at 12 to 24 h suggests a complication.
Events beyond 30 days, such as stent thrombosis at 8 months, are now primary MI rather than procedural complications. Events within 30 days in a different vessel are primary or secondary MI, depending on the setting.
Third vs Fourth vs Fifth at a glance
| Topic | Third (2012) | Fourth (2018) | Fifth (2026) |
|---|---|---|---|
| Classification | Types 1 to 5 | Seven subtypes: 1, 2, 3, 4a, 4b, 4c, 5 | Primary, secondary, procedure-related |
| Myocardial injury | cTn above 99th percentile upper reference limit, any cause | Acute vs chronic injury; non-ischaemic injury named | Sex-specific upper reference limit built into the definition; criteria for chronic injury refined |
| Troponin limit | 99th percentile upper reference limit | Sex-specific upper reference limit recommended | Sex-specific upper reference limit required |
| SCAD, spasm, embolism | Spasm and embolism in type 2 | Type 2 | Primary MI, each with its own code |
| Supply and demand MI | Type 2 | Type 2, often without imaging | Secondary MI, imaging confirmation |
| After PCI | 4a: above 5x the upper reference limit within 48 h plus criteria | 4a: above 5x the upper reference limit within 48 h; 4b, 4c any time | Imaging decides; same rule for all procedures within 30 days |
| After CABG | 5: above 10x the upper reference limit within 48 h plus criteria | 5: above 10x the upper reference limit within 48 h plus criteria | As above; 35x at 24 h only prompts imaging |
| Sudden death | Type 3 | Type 3 | Removed; classify by setting or post-mortem; BA41.Z if unevaluated |
| MINOCA | Not named | MI with non-obstructive coronary arteries | Injury with non-obstructive coronary arteries |
| Q wave | Alternative criteria | Alternative criteria | Classical: 40 ms or more and/or 25% or more of R wave |
| Coding | Not aligned with ICD-10 | Not aligned with ICD-10 | Proposed ICD-11 codes for each setting and mechanism |
Other changes worth knowing
MINOCA is renamed. The acronym now stands for myocardial injury with non-obstructive coronary arteries (no stenosis of 50% or more). It is a working diagnosis. When CMR is done, MI is the final diagnosis in only 22 to 27%; myocarditis and Takotsubo syndrome are more common. CMR within 2 weeks gives the highest yield [1].
Type 3 MI is gone. Sudden death is classified by its setting, ideally with autopsy or post-mortem CT or CMR. Without any evaluation, it is coded as unspecified MI [1].
Unrecognised MI should be confirmed by imaging, preferably CMR with LGE. It is present in about 1 in 5 people over 70 when CMR is done systematically [1].
ECG. ST elevation thresholds are unchanged. The document lists STEMI equivalents: new bundle branch block with ischaemia, Sgarbossa and modified Sgarbossa (ST/S ratio of 0.25 or less), de Winter, Wellens, Aslanger, and the South African flag sign. Up to 1 in 4 patients managed as NSTEMI have an occluded culprit artery [1].
Symptoms. The terms "typical" and "atypical" are discouraged, because "atypical" symptoms are more often reported by women and contribute to delayed diagnosis [1].
Chronic myocardial injury is flagged by two or more values above the sex-specific upper reference limit in a stable setting. Less than 20% change between values suggests stable injury. Halving eGFR roughly doubles basal troponin, but troponin elevation in CKD should not be dismissed as reduced clearance alone [1].
Age. One in 3 patients over 75 with possible MI have an elevated troponin, compared with 1 in 20 under 50. The task force calls for research to define age-adjusted upper reference limits but does not recommend them yet [1, 7].
Low-resource settings. Where neither biomarkers nor imaging are available, MI can be diagnosed from ischaemic symptoms with new ischaemic ECG changes or new Q waves [1].
Proposed ICD-11 codes
The task force worked with the WHO on codes that match the classification. They were under WHO review at publication [1].
| Code | Meaning |
|---|---|
| BA41.0 / BA41.1 | STEMI / NSTEMI (stem) |
| sixth digit 1 to 4 | Atherothrombosis, SCAD, embolism, vasospasm |
| sixth digit 5 to 7 | Restenosis, stent thrombosis, graft failure after 30 days |
| sixth digit 8 / 9 | Aetiology undetermined after imaging / unknown (no coronary imaging) |
| sixth digit A | Secondary MI, paired with the code for the acute condition |
| sixth digit B / C | Procedure-related, percutaneous / open surgery |
| BA40 | Unstable angina |
| BA41.Z | Unspecified MI (for example sudden death without evaluation) |
| BA50 | Unrecognised MI |
For example, NSTEMI due to SCAD is BA41.12, and STEMI from stent thrombosis on day 5 after PCI is BA41.0B.
What this means in the CICU
- Expect fewer MI diagnoses in septic, bleeding, or arrhythmic patients. "Troponin leak" becomes "acute myocardial injury" unless imaging shows CAD or a new wall motion abnormality.
- Order the echo. A new regional wall motion abnormality is now a diagnostic criterion for secondary and procedure-related MI, not just supporting evidence.
- Plan definitive imaging (angiography or CCTA) once the patient recovers, when it would change management.
- After cardiac surgery or PCI, a large troponin rise is a reason to image, not a diagnosis.
- Use sex-specific limits for diagnosis. Accelerated 0/1 h and 0/2 h rule-out pathways still use uniform thresholds, because those are for risk stratification rather than diagnosis.