Objective The infarct-related artery (IRA) can be predicted by ECG, but there are little concern about why the IRA could not be predicted by ECG. In the present study, we attempted to understand why did we fail or be unable to predict the IRA by ECG based on the comparison between ECG records and coronary angiographic findings.

Methods All 18-lead ECG records were compared with their angiographic findings for analysis from 1024 consecutive patients with ST elevation myocardial infarction (STEMI) between October 2004 and July 2009. More than two continous 18-lead ECG records were performed within 12 h of the onset of symptoms in all patients. Patients with previous myocardial infarction, previous coronary artery bypass surgery, previous pacemaker implant surgery or ECG evidence of left bundle branch block and time from onset to the period of the angiography more than 12 h were excluded from this study.

Results Of all 1024 patients enrolled, the IRA were failed to be predicted by ECG in 170 patients: failure prediction in 96 patients and unable evaluation in other 74 patients. Of these 170 patients, IRA was left circumflex coronary artery in 76 (44.7%) patients, right coronary artery in 66 (38.8%) patients, left anterior descending branch in 20 (11.8%) patients, ramus medianus branch in 7 (4.1%) patients, and left main in 1 (0.6%) patient. Double-vessel and triple-vessel diseases were recorded in 27 (16%) patients and 47 (28%) patients respectively. Early repolarisation syndrome occurred in 8 (5%) patients, and dextrocardia1 patient (0.6%). Angiographic study showed acute occlusion of a small branch in 6 (3.6%) patients.

Conclusion Coronary collateral vessel can mislead judgements of the IRA. When the IRA cannot be determined by ECG, left circumflex coronary artery is most likely to be the culprit vessel. Occasionally, early repolarisation syndrome and anatomic variation of the coronary artery or heart and a small branch occlusion could be causes of misjudgements.