Fig 1.
Alignment process and difference calculation.
The upper panel shows the sliding and superimposition of the red QRST complex that was to be compared with the fixed blue QRST complex. The middle panel shows the alignment of the 2 QRST complexes and their difference in one lead in a short segment of the QT interval. The lower panel shows the calculation of the signal difference in the specific segment displayed in the middle panel. This procedure was used for 1) characterizing the stability of each 50s-segment of the recording (consisting of 5 consecutive 10s-signal-averaged complexes; saQRST complexes); 2) selecting the segment with least variability; 3) from the selected segment choose the 10s-saQRST complex which was most similar to the average complex of the selected 50s-segment. Comparisons were made at each 2ms-step (time resolution at sampling rate 500Hz) and the absolute amplitude difference over the QT interval was defined (entire QRST complex). The amplitude resolution was 2.5 μV and for each comparison all differences ≥5μV were used to calculate an instability value (no unit because small differences are not included).
Fig 2.
Flow-chart showing the series of instability calculations.
The flow-chart describes the process for arriving at an instability value for a 50s qualified segment consisting of 5 consecutive 10s-signal-averaged-complexes (saQRST complexes) of the recording applying the methods described in Fig 1. This process was repeated for all qualified 50s segments of an individual recording and the segment with least variability was referred to as the selected segment from which the most representative 10s-saQRST complex was chosen for the calculation of 28 vectorcardiographic parameters.
Table 1.
Demographic and clinical characteristics of the population sample.
Table 2.
Vectorcardiographic based parameters in 1080 women and men.
Table 3.
Comparison of the time-dependent variability for the representative algorithm selected vs. the 4th 10s-saQRST complex of the recording.
Table 4.
Assessment of fully automatic vs. manually edited annotation points.
Table 5.
Recordings with high instability values.
Fig 3.
The relation between the instability value and the range for 3 vectorcardiographic parameters.
Panels a-c show graphs of the relation between the ranges (maximum-minimum values) of vectorcardiographic parameters among the 5 consecutive 10s-saQRST complexes in the selected 50s-segment and its instability value (no unit); the QT interval (panel a; rs = 0.06; NS; rs2<0.01), the Mean QRS-T angle (panel b; rs = 0.15; p<0.001; rs2 = 0.02), and the Ventricular gradient (panel c; rs = 0.51; p<0.001; rs2 = 0.265). rs is the Spearman rank order correlation coefficient. More examples are shown in S5 Fig.
Fig 4.
The lower the T wave amplitude the greater the QT range.
This graph shows the inverse relation between the ranges in QT (maximum-minimum value) in the selected 50s-segment and the T-wave amplitude. The lower this amplitude the greater the differences in Tend and therefore also in the QT interval (rs = -0.48; p < 0.001; rs2 = 0.23). rs is the Spearman rank order correlation coefficient.