Geometrically accurate transaortic mapping of left ventricular endocardial activation during surgery

Influence of filtering techniques on the time-domain analysis, diagnosis, and clinical use of signal-averaged electrocardiogram

In order to investigate the effect of different filtering techniques on the time-domain analysis of signal-averaged electrocardiogram (SAECG), recordings of 1,192 subjects were analyzed using Butterworth and Del Mar filters, both set at 40-250 Hz high and low pass frequencies. The recordings were taken from six clinically defined groups: (a) survivors of acute myocardial infarction (n = 553); (b) patients with sustained symptomatic postinfarction ventricular tachycardia (n = 89); (c) patients with hyperthropic cardiomyopathy (n = 219); (d) patients with dilated cardiomyopathy (n = 76); (e) direct relatives of patients with dilated cardiomyopathy (n = 170); and (f) normal healthy volunteers (n = 85). The study investigated differences between the SAECG results reported with both filters in three individual aspects: (1) numerical values of individual time-domain SAECG variables; (2) differences in SAECG findings of patients with postinfarction ventricular tachycardia and pair matched patients with uncomplicated follow-up after acute infarction; and (3) the power of SAECG findings to predict high risk of arrhythmic complication (sudden death and/or sustained ventricular tachycardia) among survivors of acute myocardial infarction. Compared with the Butterworth filter, the Del Mar filter led to a systematic difference of +8% in total QRS duration, was equally powerful in distinguishing between the pair matched patients with and without postinfarction ventricular tachycardia, and was statistically significantly more powerful in identifying those survivors of acute infarction who were at high risk of arrhythmic complications. The study concludes that the use of different filters may produce discordant results of SAECG analysis. Normal and abnormal values for various types of SAECG recording and analysis have to be established individually for different equipment and different software settings. These optimal cut-offs of SAECG variables should also take into account the clinical characteristics of patient groups.