BACKGROUND: Many variants of the spatial QRS-T angle (QRS-Ta) are in use. We aimed to identify the best QRS-Ta for all-cause mortality prediction among different variants.

METHODS: 6667 individuals from the Inter99 General Population Study were followed for a median of 12.7 years. Vectorcardiograms were calculated using the Kors and Inverse Dower matrices. The QRS-Ta was calculated using both mean and peak vectors of the QRS- and T-loops. Hazard ratios (HR) for abnormal QRS-Tas were calculated using a Cox's Proportional Hazard Model.

RESULTS: The highest HR and largest AUC for all-cause mortality was obtained with the Kors matrix and the mean vector (HR = 2.2, 95% confidence interval: [1.38;3.43] p < 0.001, in men). There was interaction with the orientation of the QRS-T plane.

CONCLUSION: For optimal prediction of all-cause mortality, the mean vectors in the QRS- and T-loops of the Kors-derived vectorcardiogram should be used. QRS-T plane orientation affects mortality prediction.