Subtle anterior STEMI calculator (4-variable)

Clinical background

With ST elevation in V2 to V4 and chest pain, the question sometimes arises whether the pattern represents a subtle LAD occlusion or a benign normal variant, formerly called early repolarisation. This is one of the most difficult ECG diagnostic problems in the emergency department: an incorrect judgement in either direction has serious consequences. Withholding emergency coronary intervention in an occluded LAD means lost myocardium, whereas an unnecessary catheterisation of a patient with a normal variant carries risk and wastes resources.

The appearance of the ST elevation alone is insufficient to separate these two conditions. The formula rests on the insight that a normal variant and a subtle LAD occlusion differ along several measurable dimensions simultaneously: the magnitude of the ST elevation, the length of the QT interval, the QRS amplitude and the R wave amplitude. By combining these variables into a weighted sum, one obtains a measure that outperforms any single ECG parameter.

Calculating the subtle anterior STEMI formula (4 variables)

The formula is calculated as:

$$\text{Score} = 1{.}062 \times \text{STE60V3} + 0{.}052 \times \text{QTc} - 0{.}151 \times \text{QRSV2} - 0{.}268 \times \text{RAV4}$$

where STE60V3 is the ST elevation 60 ms after the J point in V3 (mm), QTc is the computer-calculated Bazett-corrected QT interval (ms), QRSV2 is the QRS amplitude in V2 (mm) and RAV4 is the R wave amplitude in V4 (mm). A value of 18.2 or higher favours LAD occlusion.

The formula was derived by Driver et al. at Hennepin County Medical Center in Minneapolis [1]. The study reviewed 355 consecutive cases of proven LAD occlusion and excluded those constituting obvious STEMI on predefined ECG criteria: a straight or convex ST segment in V2 to V6, a summed inferior ST depression of 1 mm or more, anterior ST depression, Q waves, terminal QRS distortion or ST elevation above 5 mm. This left 143 subtle LAD occlusions. The control group consisted of 171 patients presenting to the emergency department with chest pain, in whom serial troponins excluded myocardial infarction and a cardiologist judged the ECG to show a normal variant with at least 1 mm ST elevation in V2 and V3. The outcome measure was binary: proven LAD occlusion versus normal variant.

The formula is a further development of an earlier 3-variable formula from the same group, which used STE60V3, QTc and RAV4 with a threshold of 23.4. Adding the QRS amplitude in V2 improved discrimination significantly: the AUC increased from 0.954 to 0.969 (p = 0.009) [1].

Interpretation in practice

A value just below the threshold does not mean that occlusion is excluded. The formula supports clinical judgement; it does not replace it. In particular, where clinical suspicion is high, ST changes are dynamic or biomarkers are rising, the patient should be managed as an acute myocardial infarction irrespective of the formula result.

Validation and performance

In the derivation cohort the formula achieved, at a threshold of $\geq 18{.}2$, a sensitivity of 88.8%, a specificity of 94.7% and an overall accuracy of 92.0% [1]. The positive likelihood ratio was 16.9 (95% CI: 8.9 to 32) and the negative likelihood ratio 0.12 (95% CI: 0.07 to 0.19). At the same specificity as the 3-variable formula (90.6%), sensitivity increased from 86% to 90.2% at a threshold of 17.75, and the formula correctly identified 6 additional infarctions.

A prospective validation was carried out by Sert et al. in Turkey between April 2021 and January 2023 [2]. The study included 232 patients, 116 with anterior myocardial infarction and 116 with benign variant ST elevation, matched for ST elevation of at least 1 mm in the anterior leads. At a threshold of $\geq 18{.}2$ on the first ECG, sensitivity was 82.7%, specificity 85.3% and diagnostic accuracy 83.6%. Performance improved with serial ECGs: on the third ECG, taken 20 minutes after arrival, sensitivity was 83.6%, specificity 89.6% and accuracy 86.2%.

The external validation thus gives lower figures than the derivation cohort, particularly for sensitivity on the initial ECG. This is to be expected: the derivation cohort was selected with strict exclusion criteria and the measurements were made by experienced readers. In the Turkish cohort the patients were unselected emergency department attenders and the measurements reflect more realistic conditions. That serial ECGs improved performance suggests that the formula works best when applied to repeated recordings rather than to a single initial ECG.

Limitations

The formula applies only to patients with ST elevation in V2 to V4 and chest pain in whom the differential diagnosis is a normal variant versus subtle LAD occlusion. It should not be used in obvious STEMI, left ventricular hypertrophy, left bundle branch block or a pericarditis pattern, since these conditions alter the variables on which the formula rests and can give misleading results.

The derivation cohort systematically excluded obvious STEMI using specific ECG criteria. Applying the formula to ECGs with a straight or convex ST segment, Q waves, terminal QRS distortion or ST elevation above 5 mm is not meaningful, since these patterns already diagnose STEMI independently of any formula.

An important limitation is that the formula relies on a computer-calculated QTc using Bazett's correction. Bazett's method is known to overcorrect at high heart rates and undercorrect at low ones, which can affect the score in patients with marked tachycardia or bradycardia. In addition, computerised QT measurement can vary between ECG machines, which makes the formula sensitive to the technical equipment used.

The external validation by Sert et al. showed lower sensitivity on the initial ECG than the derivation study, underlining that the formula must not be used as the sole basis for excluding LAD occlusion. A value below 18.2 reduces the probability but does not eliminate it.

The formula has so far been validated in only one external prospective study of limited size [2]. Further independent validations in different populations are needed before the formula can be recommended as a routine tool.

References

1. Driver BE, Khalil A, Henry T et al. A new 4-variable formula to differentiate normal variant ST segment elevation in V2-V4 (early repolarization) from subtle left anterior descending coronary occlusion. J Electrocardiol 2017;50(5):561-569. PMID: 28460689
2. Sert ET, Gül M, Kokulu K et al. Usefulness of the four-variable formula on serial electrocardiograms in detecting subtle anterior myocardial infarction. Am J Emerg Med 2023;73:83-87. PMID: 37633077