STEMI Mimics: ECG Practice

"ST elevation equals STEMI" is a useful first approximation and a poor final position. ST elevation is an electrocardiographic finding with a long differential, and coronary occlusion is a clinical event that does not always produce the textbook finding. Those two facts generate errors in opposite directions, and the same practice fixes both.

The aim here is not a list of conditions to memorise. It is a set of questions to ask of any trace with ST elevation, and a description of the categories of ECG that are worth deliberately seeking out until they stop surprising you.

If you want to test this against real traces, work through mimic cases in the app.

Why the label is the wrong starting point

The clinically important question is not "does this ECG meet criteria" but "is a coronary artery occluded right now, and does this patient need reperfusion". Those questions have different answers more often than is comfortable.

Criteria exist because they are reproducible and they perform reasonably at a population level. They were never intended as the sole determinant of a decision in a single patient, and guideline documents are explicit that the ECG is interpreted alongside the presentation, the trajectory and prior traces.

Practically, that means holding two separate conclusions at the same time. First: what does this trace show, described precisely. Second: what is the probability of occlusion given the trace and everything else I know. Collapsing those two into one label is where most errors begin.

Categories that imitate occlusion

Grouping the mimics by mechanism is more durable than a list, because you can reason about an unfamiliar trace if you know which mechanisms are available.

Baseline repolarisation variants

Some hearts produce ST elevation as their normal state, most often in the earlier precordial leads, and it can be substantial. The features that point this way are stability over time, a concordant and proportionate appearance, absence of reciprocal change, and — decisively — an identical previous trace. This category is the single strongest argument for chasing down an old ECG before committing.

Altered depolarisation

Bundle branch block, ventricular pacing, ventricular rhythms and pre-excitation all change the QRS, and repolarisation follows the QRS. Applying narrow-complex thresholds to these traces produces both false positives and false negatives. Specific approaches exist for assessing ischaemia in the presence of altered conduction; they are worth learning properly rather than approximately, and where you are not confident, the patient's presentation and serial traces carry more weight than your reading.

Pericardial and myocardial inflammation

Inflammatory processes can produce widespread ST elevation that does not respect coronary territories. The distribution is the clue: territorial change with reciprocal depression points one way, diffuse change without a reciprocal pattern points another. Neither is definitive, and the two conditions can also coexist.

Metabolic and electrolyte causes

Marked electrolyte disturbance can distort repolarisation dramatically and produce traces that look alarming and territorial. The tell is usually elsewhere on the ECG — conduction changes, T wave morphology, QRS width — which is why running the full reading order matters rather than jumping to the ST segments.

Structural and pressure causes

Ventricular hypertrophy, aneurysmal change after previous infarction, and raised intracranial pressure all have recognised repolarisation effects. Aneurysmal change is a particular trap because it is genuinely fixed: the trace looks like an infarct because there was one, years ago.

Technical causes

Lead misplacement and artefact. Boring, common, and responsible for a meaningful share of unnecessary escalations. Repeating the ECG with careful placement costs three minutes.

Occlusion that does not look like it

The reverse error receives less teaching time and does more harm. Several situations produce genuine occlusion with a trace that fails criteria.

• Territories the standard leads represent poorly. Some regions of myocardium are seen only indirectly, and their infarction can present as depression or as T wave change in the leads that face the opposite wall rather than as elevation anywhere.
• Very early change. Occlusion is a process. A trace recorded in the first minutes may show only subtle T wave changes, and the diagnostic appearance may develop over the following half hour. The countermeasure is serial ECGs at defined intervals, not a single trace.
• Elevation that is real but subthreshold. Millimetre criteria are thresholds, not physiological boundaries. Change that is proportionally significant for the size of the QRS may not reach an absolute cut-off.
• Occlusion with altered conduction. As above — the same feature that generates false positives also masks true ones.
• Patterns recognised as high risk without meeting elevation criteria. Several described morphologies are associated with critical coronary lesions and do not produce elevation. They are worth learning by name from a current cardiology source and worth acting on when the clinical picture supports them.

The unifying countermeasure for all five is the same: when the story is convincing, a non-diagnostic ECG is a reason to repeat it and to keep looking, not a reason to stand down.

Five questions to ask of any elevated trace

Run these in order, before you commit to a label. They take under a minute and they catch most of both error types.

1. Is the QRS normal? If not, the thresholds you are about to apply do not apply. Establish this before looking at the ST segments at all.
2. Is the change territorial? Contiguous leads corresponding to a vascular distribution, with reciprocal depression in the opposing leads. Territorial plus reciprocal is a strong pattern; diffuse without reciprocal is a different problem.
3. Is there a previous trace? The highest-value question on the list. A finding that is unchanged from two years ago is a baseline; the same finding that is new is an event. Chase this before escalating and before standing down.
4. Does the ECG match the patient? Not just the symptoms — the whole picture, including how unwell they look, their trajectory, and what else could explain both the trace and the presentation. A dramatic trace in a comfortable patient with an alternative explanation deserves scrutiny.
5. What will the next ECG tell me, and when am I recording it? Naming the interval converts uncertainty into a plan. Evolution over fifteen minutes is one of the most discriminating pieces of information available to you, and it costs nothing.

What changes when reperfusion is hours away

In a hospital with a catheter lab, an uncertain trace can be resolved by a rapid specialist opinion and a short journey. At sea, offshore, or at a remote site, the same uncertainty attaches to a much heavier decision: a treatment with real risk, or an evacuation that takes hours and commits significant resources.

Three things follow. Serial ECGs become more valuable, not less, because time is the diagnostic tool you still have. Remote specialist advice should be sought earlier and with a described trace rather than a label, so the person advising can form their own view. And the documentation of your reasoning matters more, because the decision will be reviewed by people who were not there and could not see the patient.

The interpretation skill is identical. What changes is the cost of being wrong in either direction, which is an argument for describing findings precisely rather than reaching for a label early.

A practice plan

Mimics are learned by volume and by consequence. A structure that works:

Deliberately seek out the categories above. Work through a teaching library by mechanism rather than at random, so you meet each category enough times to recognise it. Five examples of each beats fifty random traces.

Always commit before revealing. Write the description, the differential, your confidence, and your action. Then check. Comparing your reasoning to the answer is what produces learning; comparing your verdict alone does not.

Practise with the clinical context attached and then without it. Reading a trace blind trains pattern recognition. Reading it with the story trains the actual bedside task. You need both, and most people only ever do the first.

Keep a list of the ones that caught you. Not the topic — the reason. "Applied narrow-complex thresholds to a paced rhythm" is a rule. "Missed a mimic" is not.