Magnitude, mechanism, and reproducibility of QT interval differences between superimposed global and individual lead ECG complexes

Abstract

Background: The global QT interval, emerging as a standard measurement provided by digital electrocardiographs, is defined by the earliest QRS onset and latest T-wave offset that occur in any of the standard leads. Differences between global ECG measurements and those from individual ECG leads have implications for the redefinition of normal values, for recognition of disease, and for drug safety. This study sought to quantify the differences between global QT intervals measured from 12 superimposed ECG leads with QT intervals and from single lead complexes, to examine the separate effects of QRS onset and T-wave offset on these differences, and to examine the reproducibility of these measurements.

Methods: QTo intervals (Q onset to T offset) from 50 digitized ECGs sampled at 500 Hz were examined by computer assisted derivation of representative complexes from standard leads II, V(2), and V(3), by both baseline and tangent methods. Global QTo intervals were measured from superimposition of the representative complexes of all 12 leads. A time-coherent matrix of waveform onset and offset points allowed direct comparison of the components of the differences.

Results: Global QTo and Bazett-adjusted global QTc were greater than each of the baseline and tangent measurements in representative leads II, V(2), and V(3), with mean differences ranging from 8 to 18 ms. QRS onset was earlier in the global complex than in each of the representative leads, with mean differences of 3-5 ms, whereas T-wave offset was significantly later in the global complex than in each of the representative leads, with mean differences of 5-11 ms. Remeasurement of all ECGs after an interval of 6 months confirmed the relative magnitudes of the global and individual lead QTo durations and small mean differences between pairs (-0.9 to 2.7 ms). Although global QTo had the largest mean difference (only 2.7 ms), it had the smallest standard deviation of the mean difference and lowest coefficient of variability (1.58%) of all measurements.

Conclusion: Global QT measurements are systematically larger than measurements from representative complexes of individual leads. These differences result from the combined effects of earlier QRS onset and later T-wave offset in the global complex, with T-wave offset the more dominant component of the difference.

Figure 1

Superimposed representative complexes from 12 leads (top), with temporally aligned individual representative complexes from lead II (middle) and lead V₂ (bottom). The lines represent global QRS onset, global QRS offset, and global T‐wave offset as represented by the superimposed complexes. Note that apparent QT in each of the single leads is less than the global QT measurement.

Figure 2

Mean (standard error) Bazett‐corrected QTc from global superimposition and from individual representative leads II, V₂, and V₃, using baseline and tangent methods of QT measurement. QTob = Bazett corrected QTc; Repbaseline(2, V₂, V₃) = representative lead (II, V₂, V₃) by baseline method of QT measurement; Reptangent (2, V₂, V₃) = representative lead (II, V₂, V₃) by tangent method of QT measurement.

Figure 3

Bland–Altman plot illustrating the relation of the differences between global QTo and representative lead II QTo (baseline method) to underlying QTo value expressed as the average of the two measurements. Data are shown with the mean regression line and 95% boundary. Despite the systematically higher values for global QT, there is no relation of the differences to the underlying QT measurements (R²= 0.00).

Figure 4

Mean (standard error) difference between global QTo and representative lead II QTo, with components of QTo difference attributable to earlier onset (Rep2‐Global QRSonset) and later offset (Global‐Rep2 Toffset) of global QTo relative to lead II.