Principal component analysis of the t-wave for mortality Prediction in hemodialysis patients.

PRINCIPAL COMPONENT ANALYSIS OF THE T-WAVE FOR MORTALITY PREDICTION IN HEMODIALYSIS PATIENTS Patients undergoing hemodialysis (HD) therapy often experience alterations in cardiac excitability and have accounting for an estimated 3-year cumulative probability of cardiovascular death of 39.5% of total deaths [1]. Abnormalities in ventricular repolarization and its dispersion could be a cause of HD-induced arrhythmogenic effect. Nowadays, no ECG-derived parameter has been proven to predict the risk of cardiovascular death. QT dispersion (QTd) has been proposed, however, some concerns have been raised about uncertainty of the QT dispersion measurement and technical difficulties in measuring the QT interval. Principal component analysis (PCA) of the T-wave vector applied to 12-lead recordings has been proposed to obtain an ECG marker of vulnerability to ventricular arrhythmias and of cardiovascular mortality [2]. Several studies showed that the ratio of the second to first eigenvalues (PCA ratio) more accurately represents repolarization abnormalities than QTd in a large general population sample [3,4]. The aim of this study was to explore the predictive value of the PCA ratio parameter for all-cause and cardiac mortality in a retrospective study on HD patients. METHODS The selected subjects were 122 patients (46 women and 76 men, mean age 77±10) in whom digital ECG recordings were available for the analysis from previous clinical studies. Standard holter 12-lead recordings (H-12 Holter, Mortara Instrument Inc., Milwaukee, Wisconsin, USA) were collected starting with the dialysis session. ECGs were sampled at 180 Hz or 1kHz and stored to a PC hard disk for subsequent analysis. PCA is an established method for representing data and, when applied to T-wave, it describes features of repolarization in a manner that is less dependent on precise determination of T-wave offset. Singular value decomposition was applied to the covariance matrix of the raw ECG data corresponding to T-wave from the eight independent ECG leads. Then, the main eigenvectors of the spatial T-wave were computed. The first eigenvector accounts for most of the energy in repolarization when applied to the normal T-wave vector, whereas inhomogeneous repolarization, if present, is indicated by a relevant contribution of the second and third components. Thus, the ratio of the second to first eigenvalues of the spatial T-wave vector (PCA ratio) generated from the 12-lead digital ECG serves as a measure of T-wave complexity or heterogeneity of repolarization, with increasing values referring to higher amount of complexity. As shown in fig. 1 the PCA ratio provides information that can be visualized by analogy as the long and short axes of the three-dimensional T-wave loop and provides an estimate of the relative fatness of the T-wave loop relative to its peak amplitude, in which a fatter loop with a higher PCA ratio reflects more complex Twave morphology. A median value of PCA was computed for each patient throughout the whole ECG recording. Following the Strong Heart Study [3] a threshold for PCA ratio in men and women, independently of gender, was defined as 28%. Deaths were identified in an ongoing surveillance in each dialysis center and were verified through review of medical records. Deaths were classified as cardiac if caused by myocardial infarction, sudden death from CHD, or congestive heart failure by an independent review panel of physicians unaware of PCA ratio findings. After a maximum follow-up of 5 years, patients were censored as dead or alive considering the days from the date of the first ECG recording. Patients were then divided in two groups depending on the median PCA ratio value. Endpoints were all-cause mortality and cardiac mortality. Mortality rates were calculated and plotted according to the Kaplan-Meier analysis. P<0.05 was considered significant. RESULTS AND DISCUSSION During the fo...