Noise in the signal-averaged electrocardiogram and accuracy for identification of patients with sustained monomorphic ventricular tachycardia after myocardial infarction

Abstract

Late potentials are detected at various noise levels in clinical studies. The aim of this study was, in a case-control design, to assess the effect of residual noise level on the identification of patients with sustained monomorphic ventricular tachycardia after myocardial infarction. Electrocardiograms from 16 patients with prior myocardial infarction and documented sustained monomorphic ventricular tachycardia and 41 patients with prior myocardial infarction and without ventricular tachycardia, were analysed by two signal averaging procedures to noise level 0.2 and 0.4 muV. Standard time domain parameters were measured. Two definitions of late potential were analysed: (1) if any two of the following criteria were present (signal-averaged QRS duration > 120 ms, late potential duration > 40 ms, and root-mean-square voltage of the terminal 40 ms of the filtered QRS < 25 muV); or (2) if the signal-averaged QRS duration > or = 120 ms. Overall the signal-averaged electrocardiogram performed better at noise level 0.4 muV compared to noise level 0.2 muV with respect to identification of patients with or without ventricular tachycardia after myocardial infarction. Reducing noise level from 0.4 to 0.2 muV increased the sensitivity, but the consequence was a substantial decrease in specificity. Our data indicate that when a high sensitivity is the goal, the definition based only on signal-averaged QRS duration > or = 120 ms should be applied; sensitivity was 88% and specificity 59% at noise level 0.4 muV. If a high specificity is the goal, the definition should be based on the definition with two abnormal parameters; sensitivity was 69% and specificity 68% at noise level 0.4 muV.