A Bayesian Lens on a Reversible “High-Risk” ECG

Interpreting ACS–Like ECG Patterns Through Pre-Test Probability

If you saw this ECG ,

aVR ST elevation.

V1 ST Segment elevated in comparison to V2.

Diffuse ST depression.

New right axis deviation.

your first instinct would probably be left main or proximal LAD occlusion.

Ours too.

But this case turned out to be a powerful reminder that ECG interpretation in Emergency Medicine is Bayesian — not binary.
The meaning of any ECG depends on the pre-test probability, the clinical context, and the physiology beneath it.

And in this patient, the entire picture shifted when we learned one critical detail:

He had an acute upper GI bleed earlier that day — approximately 1 liter.

ECG 1 was taken shortly after this massive blood loss.

And after correcting the reversible physiologic insult, the ECG normalized completely.

This post walks through the reasoning, the differentials, and what this ECG really means in an ED setting.

ECG 1: The Concerning Findings

The initial ECG showed:

• aVR ST elevation

• V1 > V2 ST segment level

• Diffuse ST depression

• New right axis deviation (Lead I isoelectric/negative vs. ECG 2)

This pattern is alarming because it mimics:

• Proximal LAD occlusion

• Left main disease

• Severe multivessel ischemia

But ECGs don’t diagnose ACS — they shift probability.
The patient’s physiology completes the story.

The Clinical Context:

This patient had a fresh bout of ~1 L blood loss from an upper GI bleed a few hours before arriving in the ED.

Acute hemorrhage →
↓ oxygen delivery
↓ preload
↓ coronary perfusion pressure
↑ myocardial oxygen demand
↑ sympathetic surge

This is the perfect storm for global subendocardial ischemia — a condition that mimics ACS on ECG but is fully reversible.

And in this case, it was.

ECG 2

After stabilizing the patient and correcting the physiologic derangement:

• ST depressions resolved

• aVR ST elevation normalized

• V1/V2 changes disappeared

• Axis returned to baseline

Understanding These ECG Patterns Through a Bayesian Lens

When encountering this pattern:

• aVR ST elevation

• Diffuse ST depression

• V1 > V2 ST elevation

• Axis shift

you must evaluate the pre-test probability of ACS

These ECG features have:

• High sensitivity for global ischemia

• Low specificity for ACS

• Moderate LR+ only when the patient already has a high ACS likelihood

So your clinical judgment determines how much the ECG moves your post-test probability.

A Practical ED Framework: How to Approach This ECG

1. If Pre-Test Probability of ACS Is HIGH

Treat as ACS until proven otherwise.

This includes patients with:

• Classic ischemic symptoms

• High-risk CAD profile

• Hemodynamic instability

• Concerning echo findings

• High clinical suspicion

This ECG may represent:

• Precordial swirl pattern

• Proximal LAD occlusion

• Left main equivalent

Immediate steps:

• POCUS for RWMA

• Repeat ECGs

• ACS pathway activation

• Stabilization + consults

(References: Harhash 2019; Bhatt 2022; Ricci 2025)

2. If Pre-Test Probability of ACS Is LOW

Rule out other causes first — especially reversible, physiologic ones.

Use:

• POCUS (heart, lungs, IVC)

• ABG/VBG (pH, lactate, bicarbonate)

• Electrolytes

• Hb level

• Focused history (bleeding, toxins, hypoxia, PE risk)

If no reversible cause emerges →
reconsider ACS and escalate accordingly.

(References: Bouzid 2022; Choi 2022; Y-Hassan 2020)

Differential Diagnosis: Why This ECG Is Not Always ACS

This pattern often reflects sub-endocardial ischemia or metabolic stress.

1️⃣ Perfusion Supply–Demand (S/D) mismatch

• Severe anemia

• Shock

• Hypoxia

• Tachycardia (can unmask CAD)

• Acute blood loss ← the cause in this case

(References: Bouzid 2022; Al-Zaiti 2020; Y-Hassan 2020)

2️⃣ Pulmonary Embolism

• RAD

• V1 > V2 elevation

• RV strain patterns

(References: Harhash 2019; Y-Hassan 2020)

3️⃣ Severe Aortic Stenosis

4️⃣ Aortic Dissection

5️⃣ Metabolic Causes

• Hypokalemia

• Severe metabolic acidosis

• Sodium channel blocker toxicity

All can mimic ischemia — and all can resolve with treatment.

(References: Y-Hassan 2020; Hong & Zeng 2022)

Serial ECGs + POCUS = Diagnostic Power Tools

Literature consistently supports:

• Serial ECGs identify transient or reversible changes (Bouzid 2022; Choi 2022)

• POCUS rapidly shifts probability up or down

• Dynamic improvement → strong evidence against ACS

• Persistent abnormalities → significantly raises ACS probability

This is bedside Bayesian reasoning in action.

Final Takeaway

An ischemic-looking ECG does not always mean ACS.
With the right clinical context — like acute hemorrhage — these changes may be entirely reversible.

What matters most is:

• The pre-test probability

• The physiology

• Bedside imaging

• Response to treatment

• Serial ECG trends

In Emergency Medicine, pattern recognition is essential.
But pattern + probability is what saves lives.

References:

1. Bouzid Z, Faramand Z, Martin-Gill C, Sereika S, Callaway C, Saba S, et al. Incorporation of serial 12-lead electrocardiogram with machine learning to augment the out-of-hospital diagnosis of non-ST elevation acute coronary syndrome. Ann Emerg Med. 2022. doi:10.1016/j.annemergmed.2022.08.005

2. Choi Y, Lee J. Dynamic changes in electrocardiographic findings between initial and follow-up electrocardiography: The role of the T/QRS ratio. Am J Emerg Med. 2022;54:8-14. doi:10.1016/j.ajem.2022.01.033

3. Bhatt D, Lopes R, Harrington R. Diagnosis and treatment of acute coronary syndromes: A review. JAMA. 2022;327(7):662-75. doi:10.1001/jama.2022.0358

4. Harhash A, Huang J, Reddy S, Natarajan B, Balakrishnan M, Shetty R, et al. aVR ST segment elevation: Acute STEMI or not? Incidence of an acute coronary occlusion. Am J Med. 2019;132(5):622-30. doi:10.1016/j.amjmed.2018.12.021

5. Ricci F, Martini C, Scordo D, Rossi D, Gallina S, Fedorowski A, et al. ECG patterns of occlusion myocardial infarction: A narrative review. Ann Emerg Med. 2025. doi:10.1016/j.annemergmed.2024.11.019

6. Al-Zaiti S, Besomi L, Bouzid Z, Faramand Z, Frisch S, Martin-Gill C, et al. Machine learning-based prediction of acute coronary syndrome using only the pre-hospital 12-lead electrocardiogram. Nat Commun. 2020;11:3966. doi:10.1038/s41467-020-17804-2

7. Hong J, Zeng Z. Predictive value of ST-segment deviation in aVR in patients suffering from acute coronary syndrome: A retrospective cohort study. Medicine (Baltimore). 2022;101(5):e29994. doi:10.1097/MD.0000000000029994

8. Y-Hassan S, Falhammar H. Cardiovascular manifestations and complications of pheochromocytomas and paragangliomas. J Clin Med. 2020;9(8):2435. doi:10.3390/jcm9082435