Optimizing risk stratification: a modified signal-averaged ECG method

The signal-averaged ECG (SA-ECG) has established value in identifying patients a risk for serious ventricular arrhythmias following myocardial infarction. The standard SA-ECG has high negative (90-100%), but low positive (10-35%) predictive value. This latter characteristic appears, in part, related to the method by which the SA-ECG is derived. Given that the onset & offser of the QRS complex and the presence or absence of late potentials are inherently related to the signal to noise ratio, it may not be appropriate to derive these parameters using a single set of ECG epochs. To further addrms this issue, 78 individual SA-ECG recordings were analyzed at a target noise level of 0.5 pV and at a fmed number of templates (300). Wsing our mod@ed approach, a significant percentage of patients were classified as "indeterminate" in that the confidence interval for QRS width crossed the threshold level.