Filtering approach based on empirical mode decomposition improves the assessment of short scale complexity in long QT syndrome type 1 population.

This study assesses the complexity of heart period (HP) and QT variability series through sample entropy (SampEn) in long QT syndrome type 1 individuals. In order to improve signal-to-noise ratio SampEn was evaluated over the original series (SampEn0) and over the residual computed by subtracting the first oscillatory mode identified by empirical mode decomposition (SampEn(EMD1R)). HP and QT interval were continuously extracted during daytime (2:00-6:00 PM) from 24 hour Holter recordings in 14 non mutation carriers (NMCs) and 34 mutation carriers (MCs) subdivided in 11 asymptomatic (ASYMP) and 23 symptomatic (SYMP). Both NMCs and MCs belonged to the same family line. While SampEn0 did not show differences among the three groups, Samp(EnEMD1R) assessed over the QT series significantly decreased in ASYMP subjects. SampEn(EMD1R) identified a possible factor (i.e. the lower short scale QT complexity) that might contribute to the different risk profile of the ASYMP group.

A quantitative assessment of T-wave morphology in LQT1, LQT2, and healthy individuals based on Holter recording technology.

BACKGROUND: The clinical course and the precipitating risk factors in the congenital long QT syndrome (LQTS) are genotype specific. OBJECTIVES: The goal of this study was to develop a computer algorithm allowing for electrocardiogram (ECG)-based identification and differentiation of LQT1 and LQT2 carriers. METHODS: Twelve-lead ECG Holter monitor recordings were acquired in 49 LQT1 carriers, 25 LQT2 carriers, and 38 healthy subjects as controls. The cardiac beats were clustered based on heart-rate bin method. Scalar and vectorial repolarization parameters were compared for similar heart rates among study groups. The Q to Tpeak (QTpeak), the Tpeak to Tend interval, T-wave magnitude and T-loop morphology were automatically quantified using custom-made algorithms. RESULTS: QTpeak from lead II and the right slope of the T-wave were the most discriminant parameters for differentiating the 3 groups using prespecified heart rate bin (75.0 to 77.5 beats/min). The predictive model utilizing these scalar parameters was validated using the entire spectrum of heart rates. Both scalar and vectorcardiographic models provided very effective identification of tested subjects in heart rates between 60 and 100 beats/min, whereas they had limited performance during tachycardia and slightly better discrimination in bradycardia. In the 60 to 100 beats/min heart rate range, the best 2-variable model identified correctly 89% of healthy subjects, 84% of LQT1 carriers, and 92% of LQT2 carriers. A model including 3 parameters based purely on scalar ECG parameters could correctly identify 90% of the population (89% of healthy subjects, 90% of LQT1 carriers, and 92% of LQT2 carriers). CONCLUSION: Automatic algorithm quantifying T-wave morphology discriminates LQT1 and LQT2 carriers and healthy subjects with high accuracy. Such computerized ECG methodology could assist physicians evaluating subjects suspected for LQTS.