An algorithm for QT dispersion analysis: Validation and application in chronic kidney disease

This work presents an algorithm for automatic analysis of QT dispersion (QTd) on 3 quasi-orthogonal leads (DI, aVF and V2) of ECG, based on the wavelet transform (WT). In a first step, QRS complexes and T waves of different morphologies are detected in each one of the three leads. Next, each QT interval is delineated by detecting and identifying the characteristic points as QRS onset (Qi) and T wave end (Te). Finally, mean of three consecutive QT intervals of the same beats in each lead is calculated to measure QTd. The accuracy of the algorithm developed has been validated with recordings of the manually annotated QT database, by comparing its results with manual annotations of an expert cardiologist, showing that the error obtained is within the tolerance range defined by each characteristic point, Qi and Te. The algorithm has been applied in a study to evaluate QTd in recordings of 7 normal subjects and 7 renal insufficiency patients from the PTB diagnostic ECG database. The results show that the QTd algorithm is effective in differentiating both groups. (p=0.004).