ST-segment elevation in aVR: is it predictive for clinical or angiographic outcome in chest pain patients?

Background
The European Society of Cardiology guideline recommends assessing the presence of ST-segment elevation (STE) in chest pain patients. The diagnostic accuracy of STE is limited, a considerable number of patients with acute coronary occlusion (ACO) do not have the required amount of STE. Beyond conventional ST-segment elevation myocardial infarction (STEMI) criteria, a few electrocardiogram (ECG) patterns are defined as equivalents. The definitions vary in the literature, but STE in lead aVR in combination with multi-lead depression is widely accepted. It was associated with ACO in either left main coronary artery (LMCA) or proximal left artery descending (LAD). STE in aVR might also appear in combination with STE in lead V1. But these ECG patterns are not specific, they were also described in patients with rather oxygen demand-supply mismatch than ACO. There is uncertainty about the predictive value of aVR-STE in chest pain patients.

Objective
We investigated the predictive value of aVR-STE in an all-comer German metropolitan chest pain unit (CPU) cohort

Methods
All consecutive patients presenting to a metropolitan CPU in Germany from 01.11.2018 to 30.06.2021 were screened. Patients with any aVR-STE on 12-lead ECG at admission were included. All ECGs were classified by two independent and trained physicians. They were blinded to the clinical outcome. Discrepancies were resolved by consensus. The ECGs were categorized as follows:

I)               aVR-STE, any

II)              aVR-STE ≥1mm

III)            aVR-STE ≥1mm + multi-lead depression (ST segment depression (STD) ≥6 leads)

IV)            aVR-STE ≥1mm + V1-STE ≥1mm

ACO was the modelled outcome and considered the following parameters: identification of a culprit lesion, infarct artery flow and peak creatine kinase (>upper limit of normal). 

Results
4440 patient admissions were registered in the CPU, 4121 had traceable ECGs and were screened. Of these, 145 (3.5%) patients had any aVR-STE and were included. The mean age of patients was 67 (±15.9) years, and 65.2% were male. Overall, 57 (39.3%) patients were ruled-out by serial hs-cTnT measurement and 63 (43.4%) patients were referred for angiography. The remaining 25 (17.2%) patients were managed conservatively. ACO prevalence was 12.4%, correspondingly the ACO prevalence was 2.5% in the overall CPU cohort.

The aVR-STE clusters are presented in Figure 1 and isolated  aVR-STE ≥1mm was the most prevalent finding. The diagnostic performance of all clusters is summarized in Table 1. Overall, all criteria had high negative predictive values (87.7 – 92.5%), but the positive predictive values were really low ranging from 12.9% to 18.5%. As aVR-STE criteria were associated with LMCA or proximal LAD lesions before, further angiographic analysis was performed among ACO patients. Only 50% of culprit lesions were located in LMCA or proximal LAD. The remaining were located in the proximal left cirucumflex artery (16.7%), right coronary artery (22.2%), distal segments or sight branches of left coronary arteries (11.1%).

Conclusion 
ACO was more prevalent among patients with aVR-STE compared to the overall CPU cohort. aVR-STE in combination with any other criterion resulted in acceptable specificity, but the diagnostic performance predicting ACO remained poor. Only every second patient with aVR-STE and ACO had a culprit lesion located in LMCA or proximal LAD. aVR-STE criteria are barely useful to detect ACO and do not qualify for triage to emergency coronary angiography as stand-alone tool.