Model Based Analysis of QT Variability Independent of Heart Rate and Respiration

Abstract

Electrical instability in the ventricles of the heart predisposes patients to abnormal heart rhythms (ventricular arrhythmia) and sudden cardiac death (SCD). Current risk assessment strategies for identifying patients at high risk of SCD are mainly based on markers of structural dysfunction. However, most deaths occur in those deemed at low risk by traditional risk markers, stipulating the need for improvement of existing risk assessment strategies. Malignant neural modulation of the repolarisation process also contributes to arrhythmogenesis and increased risk of SCD. Improving the ability to sense instability in the repolarisation process and quantifying neural contribution to arrhythmia can improve the ability to predict the onset of ventricular arrhythmias and improve risk stratification strategies. The main aim of this thesis is to quantify repolarisation variability independent of heart rate and respiration, explore its prognostic value for improving risk prediction and examine its relationship to sympathetic neural activity. Parametric power contribution analysis methods and autoregressive modelling are used to quantify repolarisation variability independent of heart rate and respiration. Temporal ECG features that best capture the influence of sympathetic activation on repolarisation independent of other factors are identified by comparing LF powers of several features following increase in sympathetic activation elicited by orthostatic stress. Measurements of QTV independent of heart rate and respiration from single leads (II and V5) are compared to that of multi-lead approaches in a group of coronary artery disease patients to investigate the suitability of singlelead ECG for the analysis of repolarisation variability independent of heart rate and respiration. The influence of sympathetic activation elicited by orthostatic stress on this fraction of repolarisation variability is investigated in healthy adults and adolescents. Finally, survival analysis is used to explore the prognostic value of this fraction in a large cohort of myocardial infarction patients. Using stepwise multivariable Cox regression analysis, a QTV risk score was developed, and its predictive value was investigated. Results show that the variabilities of QTend and RTend increased with sympathetic activation elicited by tilt independent of heart rate and respiration. The improved model fit or signal-tonoise ratio in the multi-lead approaches did not lead to significant differences in QTV compared to single leads. Measurements of the intervention effect and pathophysiological group differences in QTV independent of heart rate and respiration were consistent in single and multi leads. Results also show that QTV independent of heart rate and respiration increased with sympathetic activation elicited by tilt and exhibited a low frequency peak consistently. Finally, this fraction of QTV was elevated in non-surviving myocardial infarction patients and exhibited a clear LF peak indicating a rhythmical source. Cox proportional hazard model analysis shows that QTV independent of heart rate and respiration is predictive of mortality. A QTV risk score that includes QTV independent of heart rate and respiration is found to be predictive of mortality independent of traditional risk markers. Further the proposed QTV risk score was able to identify a subgroup of patients at higher risk of mortality within a group deemed as low risk using traditional risk markers. In conclusion, results from this thesis show that QTV independent of heart rate and respiration originates from a rhythmical source, is increased by sympathetic activation, and might help improve stratification of patients at higher risk of mortality in combination with existing risk assessment strategies.