Files in this Data Supplement:

• Appendix - Microsoft Word) (22.5 kb). Short List of WHI Investigators.
• Table I - (Microsoft Word) (45.5 kb). Hazard Ratios and 95% Confidence Intervals for Incident Coronary Heart Disease† from Age-adjusted Single and Multiple ECG Variable Models by CHD Status at Baseline.
• Table II - (Microsoft Word) (48.0 kb). Hazard Ratios and 95% Confidence Intervals for Coronary Heart Disease Mortality From Age-Adjusted Single and Multiple ECG Variable Models by CHD Status at Baseline.
• Figure I - (TIFF) (7.74 MB). A schematic of ST-T segment of the ECG illustrating major repolarization variables used in risk evaluation. A = the mean ST segment amplitude; B = ST gradient expressing the linear increase in ST amplitude from the beginning of ST to the beginning of the T wave; C = the mean amplitude of the T wave (Tmean); D = T wave peak amplitude (Tpeak). In V5, Tpeak and Tmean are correlated at level of r = 0.92. Mean rather than peak amplitudes are obtained by waveform analysis as statistically independent components from orthonormal expansion.
• Figure II - (TIFF) (5.86 MB). A schematic of maximal QRS and T vectors (R and T), QRS/T angle (omega), and the spatial T vector loop. The percentage width and the length ratio of T vector loop can be conceived as the T wave roundness index, approximating the variance ratio of the second and the first principal components of the XYZ signal. Graphical illustration of the QRS and T wave nondipolar components of the 12-lead ECG is conceptually more problematic because of the dimensionality.