Subtraction electrocardiography: Detection of ischemia-induced ST displacement without the need to identify the J point

• Published in: Journal of electrocardiology Vol. 49; no. 3; pp. 316 - 322
• Main Authors: ter Haar, C. Cato, Man, Sum-Che, Maan, Arie C., Schalij, Martin J., Swenne, Cees A.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.05.2016
• Subjects: Algorithms; Cardiovascular; Diagnosis, Computer-Assisted - methods; Electrocardiogram; Electrocardiography - methods; Humans; J-point identification; Middle Aged; Myocardial Ischemia - diagnosis; Pattern Recognition, Automated - methods; Reproducibility of Results; Sensitivity and Specificity; ST displacement; Subtraction Technique; Electrocardiogram; ST displacement; J-point identification
• ISSN: 0022-0736; 1532-8430; 1532-8430
• DOI: 10.1016/j.jelectrocard.2016.01.004

When triaging a patient with acute chest pain at first medical contact, an electrocardiogram (ECG) is routinely made and inspected for signs of myocardial ischemia. The guidelines recommend comparison of the acute and an earlier-made ECG, when available. No concrete recommendations for this comparison exist, neither is known how to handle J-point identification difficulties. Here we present a J-point independent method for such a comparison. After conversion to vectorcardiograms, baseline and acute ischemic ECGs after 3minutes of balloon occlusion during elective PCI were compared in 81 patients of the STAFF III ECG database. Baseline vectorcardiograms were subtracted from ischemic vectorcardiograms using either the QRS onsets or the J points as synchronization instants, yielding vector magnitude difference signals, ΔH. Output variables for the J-point synchronized differences were ΔH at the actual J point and at 20, 40, 60 and 80ms thereafter. Output variables for the onset-QRS synchronized differences were the ΔH at 80, 100, 120, 140 and 160ms after onset QRS. Finally, linear regressions of all combinations of ΔHJ+… versus ΔHQRS+… were made, and the best combination was identified. The highest correlation, 0.93 (p<0.01), was found between ΔH 40ms after the J point and 160ms after the onset of the QRS complex. With a ΔH ischemia threshold of 0.05mV, 66/81 (J-point synchronized differences) and 68/81 (onset-QRS synchronized differences) subjects were above the ischemia threshold, corresponding to sensitivities of 81% and 84%, respectively. Our current study opens an alternative way to detect cardiac ischemia without the need for human expertise for determination of the J point by measuring the difference vector magnitude at 160ms after the onset of the QRS complex. •In this study, we measured ischemic ST displacements relative to the non-ischemic baseline ECG of the same individual.•We measured these displacements at various points in time relative to the baseline and ischemic J points and relative to the baseline and ischemic onset-QRS instants.•Our results suggest that ischemia diagnosis can be based on ST displacements that are measured at a fixed time interval after the baseline and ischemic onset-QRS instants.•These results imply that the difficult issue of J point localization in ischemic ECGs can be overcome by subtracting the baseline ECG from the ischemic ECG.