Lack of reproducibility of frequency versus time domain signal-averaged electrocardiographic analyses and effects of lead polarity in coronary artery disease.

Time domain signal-averaged electrocardiographic parameters have been found to be consistent over the short term. Similar data have not been verified for the frequency domain. In addition, the effects of lead polarity changes on time domain or frequency domain parameters are not known. Signal-averaged electrocardiography was performed on 15 patients using the Arrhythmia Research Technology model 1200 EPX with orthogonal X, Y and Z leads. Time domain variables included filtered QRS, root-mean-square voltage of the last 40 ms of the QRS and duration of signals less than 40 microV. Two frequency domain area ratios were analyzed: a 140-ms window including last 40 ms of QRS + 100 ms of ST; and a 140-ms window beginning at QRS onset (both 20 to 50/0 to 20 Hz). Values were compared for each lead, as well as for a vector composite and the arithmetic mean of the 3 leads. Each patient underwent 3 studies performed 5 minutes apart. The polarity of the X lead was reversed between the first and second studies, and then returned to standard for the third study. Time domain variables correlated closely among the studies irrespective of lead polarity, with r values of 0.993 to 1.000 (p less than 0.0001). Frequency domain parameter correlation was much poorer, with r values as low as 0.276. Frequency domain correlations between tests 1 and 2 and 2 and 3 (mean r = 0.778) were poorer than between tests 1 and 3 (mean r = 0.829), implying an effect of lead polarity change.(ABSTRACT TRUNCATED AT 250 WORDS)