Identification of the culprit artery can be helpful in the management of inferior
infarction with ST-segment elevation myocardial infarction
. Some studies suggest that previously published algorithms intended to help identify the infarct-related artery are suboptimal. Our aim is to develop a better method to localise the culprit artery on the basis of the 12-lead ECG
Patients and methods
We analysed the ECG
and coronary angiograms of two different cohorts of patients with inferior
ST-segment elevation myocardial infarction
. Patients from the first cohort were labelled the derivative cohort (group A), whereas patients in the second cohort were labelled the validation cohort (group B). ST-segment elevation was measured in each lead, and a multiple logistic regression analysis was carried out to determine the best equation to predict the culprit artery. A derived algorithm was then applied to the validation cohort. Next, our algorithm was applied to the total cohort of both groups and compared with four different previously published algorithms. We analysed differences in sensitivity, specificity and area under the curve (AUC).
We included 252 patients in the derivative group and 90 in the validation group. The multiple models analysis concluded that the best model should include five leads. This model was validated by internal bootstrapping with 1000 repetitions in group A and externally in group B. The resultant algorithm was as follows: (ST-elevation
in III + aVF + V3) − (ST-elevation
in II + V6) less than 0.75 mm means that the culprit artery is the left circumflex
artery (Cx). If the result is at least 0.75, the culprit artery is the right coronary artery
. The total group of both cohorts comprised 342 patients, aged 61.2 ± 12.4 years, of whom 19.6% were female and 80.4% were male. The Cx was the culprit artery in 67 (19.6%) patients. Our algorithm had a sensitivity of 72.3, a specificity of 80.9 and an AUC of 0.766. The AUC value was better compared with the other algorithms.
The best algorithm to localise the culprit artery includes ST-elevation
in leads II and V6 related to Cx, and ST-elevation
in leads III, aVF and V3 related to right coronary artery
. Our algorithm has been validated internally and externally, and works better than other previously published algorithms.