Differential electrocardiographic changes related to the culprit coronary artery in patients with acute myocardial infarction and left bundle branch block

Abstract

Abstract Background The electrocardiographic (ECG) recognition of acute myocardial infarction (AMI) in patients with concurrent left bundle branch block (LBBB) is hampered by background changes in QRS complex duration and ST segment, potential directly caused by LBBB. Currently available ECG algorithms were constructed to suspect the presence of AMI but they not take into a count the culprit coronary artery and this fact could explain their low sensitivity. Objectives To analyze the influence of the culprit coronary artery on the ECG changes and on the diagnostic yielding of ECG algorithms in patients with acute myocardial infarction and LBBB. Methods We analyzed the index ECG of 240 patients with acute chest pain and LBBB referred to emergent coronary angiography. Patients were classified in 3 groups according their coronary anatomy: a) acute coronary occlusion (n=48); b) critical non occlusive lesion (n=45); and c) absence of coronary lesions (n=76). In all ECGs we measured the ST-segment shifts and QRS duration and compared the pattern of changes in relation to the culprit coronary artery. Results Location of the culprit coronary artery induced differential ECG patterns: a) mean ST segment elevation ≥3mm in leads V2-V3 in patients with LAD occlusion (n=27); b) absence of ST segment depression in leads V5-V6 in LCX occlusion (n=11); and c) mean ST segment elevation ≥1.5 mm in leads III-aVF and attenuation of the ST elevation in leads V1-V2 (mean <1mm) in cases of RCA occlusion (n=10). Patients with critical non occlusive lesions (n=45) did not show particular ECG changes. QRS complex duration was longer in patients with LCX occlusion. Sensitivity of the Sgarbossa score was around 50%, sensitivity of the Smith score was >70% in all coronary groups and sensitivity of Barcelona score was really good for LCX and RCA occlusion. The combination of the three algorithms increased the sensitivity in all groups (>74%) but more markedly in patients with LCX occlusion (90.9%) and RCA occlusion (90%). All the ECG algorithms lose sensitivity for the critical non occlusive lesions group. Conclusions Patients with acute myocardial infarction and LBBB present distinct patterns of QRS and ST segment changes and different diagnostic yielding of ECG algorithms in relation to the location of the occluded coronary artery. Data suggest that ECG prediction of the culprit coronary artery in patients with AMI and LBBB might be possible using artificial intelligence systems in larger series of patients. Funding Acknowledgement Type of funding sources: None.