Raja Selvaraj

Body surface projection of action potential duration alternans: A combined clinical–modeling study with implications for improving T-wave alternans detection

BACKGROUND Action potential duration alternans (APDA) can vary regionally in magnitude and phase. The influence of APDA heterogeneity on T-wave alternans (TWA) has not been defined. OBJECTIVE Our objectives were: (1) to determine how APDA affects the magnitude and spatial distribution of TWA, and (2) to optimize electrocardiographic (ECG) lead configuration accordingly to improve TWA detection. METHODS Global, regional, and discordant APDA were simulated in a 257-node heart model. Using a forward solution, body surface potentials were derived at 300 points on the thorax and TWA was computed at each point. In 22 patients with cardiomyopathy (left ventricular ejection fraction 28% +/- 6%), TWA was measured from a 114-electrode body surface map using the spectral method during atrial pacing at 110 beats/min. RESULTS An increase in global APDA from 4 to 12 ms resulted in an increase in maximum TWA from 10 to 30 microV. TWA magnitude varied with the size and location of the alternating myocardium, but was largest with discordant APDA compared with regional or global APDA. Irrespective of the location or phase of APDA, TWA was largest over the precordium and correlated with T-wave amplitude in the simulation (R(2) = 0.56 +/- 0.24, P <.01) and clinical study (R(2) = 0.45 +/- 0.23, P <.02). A novel lead configuration (12 precordial leads + limb leads) significantly improved maximum TWA detection compared with the conventional 12-lead ECG+ Frank lead configuration. CONCLUSION TWA magnitude is dependent on the interaction of concordant and discordant alternating sources within the heart. Maximum TWA consistently localizes to the precordium and a novel lead configuration using 12 precordial leads improves TWA quantification.

T wave alternans evaluation using adaptive time-frequency signal analysis and non-negative matrix factorization.

Each year 400,000 North Americans die from sudden cardiac death (SCD). Identifying those patients at risk of SCD remains a formidable challenge. T wave alternans (TWA) evaluation is emerging as an important tool to risk stratify patients with heart diseases. TWA is a heart rate dependent phenomenon that manifests on the surface electrocardiogram (ECG) as a change in the shape or amplitude of the T wave every second heart beat. The presence of large magnitude TWA often presages lethal ventricular arrhythmias. Because the TWA signal is typically in the microvolt range, accurate detection algorithms are required to control for confounding noise and changing physiological conditions (i.e. data nonstationarity). In this study, we address the limitations of two common TWA estimation methods, spectral method (SM) and modified moving average (MMA). To overcome their limitations, we propose a modified TWA quantification framework, called Adaptive SM, that uses non-linear time-frequency distribution (TFD). In order to increase the robustness of TWA detection in ambulatory ECGs, we also propose a new technique, called non-negative matrix factorization (NMF)-Adaptive SM. We present the analytical background of these methods, and evaluate their accuracy in detecting synthetic TWA signal in simulated and real-world ambulatory ECG recordings under conditions of noise and data non-stationarity. The results of the numerical simulations support the effectiveness of the proposed approaches for TWA analysis, which may ultimately improve SCD risk assessment.

Short-term memory and restitution during ventricular fibrillation in human hearts: an in vivo study.

Background—Action potential duration (APD) variation is an important determinant of wave break and reentry. The determinants of APD variability during early ventricular fibrillation (VF) in myopathic human hearts have not been studied. The objective of this study was to study the role of APD restitution and short-term cardiac memory on variation in human VF. Methods and Results—The study consisted of 7 patients (67±9 years old) with ejection fraction <35%. Monophasic action potentials were recorded from the right and/or left ventricular septum during VF. APD60/90 was measured in sinus beat preceding induction of VF, and its amplitude was used to define 60%/90% repolarization in VF. The monophasic action potential upstroke (dV/dtmax) was used to characterize local excitability. Simple linear regression showed that variability in APDn60 was determined by APD/diastolic interval restitution (R2=0.48, P<0.0001) and short-term memory (APD60 n−1, n−2, n−3, n−4; R2=0.55, 0.40, 0.33, and 0.27 respectively; P<0.001). Using multiple stepwise regression, short-term memory and restitution accounted for 62% of variance in APD60 (P<0.001). Individually, memory effect had the greatest contribution to APD variability (R2=0.55, P<0.0001). Conclusions—In early human VF, short-term memory and APD/diastolic interval restitution explain most of the APD variability, with memory effects predominating. This suggests that in early human VF, short-term cardiac memory may provide a novel therapeutic target to modulate progression of VF in myopathic patients.

Microscopic systolic pressure alternans in human cardiomyopathy: Noninvasive evaluation of a novel risk marker and correlation with microvolt T-wave alternans

BACKGROUND In patients with severe left ventricular (LV) dysfunction, visible pulsus alternans coincides with visible T-wave alternans (TWA), but a similar relationship has not been described for nonvisible microscopic systolic pressure alternans (MSPA) and microvolt TWA (MTWA). OBJECTIVE The purpose of this study was to determine the prevalence of MSPA and its relationship to MTWA in patients with cardiomyopathy. METHODS Using the spectral method, MSPA was measured from finger pressure during incremental atrial pacing and then validated against MSPA measured from the LV pressure in 12 patients. In 23 other patients with cardiomyopathy (LV ejection fraction <40%), noninvasive MSPA and MTWA were measured simultaneously during incremental atrial pacing. RESULTS MSPA (<1 mm Hg) was detected in 80% of patients with cardiomyopathy and in 43% of controls. The presence of finger MSPA showed 100% positive concordance with LV MSPA; however, finger MSPA was 20% larger due to peripheral augmentation. Finger MSPA was highly concordant (96% positive concordance and 90% negative concordance) with MTWA. The magnitudes of MSPA and MTWA showed a linear correlation (R = 0.66, P <.001), and the k value, a measure of signal-to-noise ratio, was significantly larger for MSPA compared to MTWA (108 ± 88 vs 24 ± 48, P <.001). Premature beats resulted in concordant and coincident changes in MSPA and MTWA. During follow-up (454 ± 274 days), 2 (8.6%) patients experienced ventricular tachycardia, and both manifested MSPA and MTWA during pacing at 600 ms. CONCLUSION MSPA can be detected noninvasively in patients with cardiomyopathy and is coupled to MTWA, suggesting a common mechanism. The high signal-to-noise ratio of MSPA may provide a novel robust metric of sudden cardiac death risk in these patients.

Adrenergic stimulation increases repolarization dispersion and reduces activation–repolarization coupling along the RV endocardium of patients with cardiomyopathy

AIMS Dispersion of repolarization (DOR) in the human heart is minimized by activation-repolarization coupling. Adrenergic stimulation can be proarrhythmic in patients with impaired left-ventricular function and its effect on repolarization dispersion has not been systematically investigated. Our objective was to study the effect of dobutamine on repolarization dispersion and activation-repolarization coupling in patients with cardiomyopathy. METHODS AND RESULTS Activation recovery intervals (ARI) and activation times (AT) were measured from unipolar electrograms at 10 sites along the apicobasal right ventricle (RV) in 14 patients with cardiomyopathy (LVEF < 40%). These measurements were made during control, dobutamine 2.5-5.0 microg/kg/min, and a recontrol phase while maintaining constant heart rates with atrial pacing. Dispersion of repolarization was calculated from the total recovery time (TRT, AT+ARI). Activation-repolarization coupling was assessed by linear regression of ARI and AT. Dispersion of repolarization across all 10 sites and between adjacent sites increased with dobutamine compared with control (whole DOR: range 15 +/- 2 vs. 12 +/- 2 ms, P = 0.06 and standard deviation 5.5 +/- 0.9 vs. 4.3 +/- 0.9 ms, P = 0.04; adjacent DOR: 5.9 +/- 0.8 vs. 4.5 +/- 0.6 ms, P = 0.04). This was associated with shallower ARI/AT slopes (-0.3 +/- 0.2 vs. -0.8 +/- 0.2, P = 0.05) and a decrease in ARI-AT correlation (R(2) 0.4 +/- 0.1 vs. 0.6 +/- 0.1, P = 0.05) with dobutamine compared with control. CONCLUSION Adrenergic stimulation increases apicobasal RV DOR and reduces coupling between activation and repolarization in patients with cardiomyopathy. This may provide a mechanism for the proarrhythmic potential of heightened adrenergic states in these patients.

Modulated dispersion of activation and repolarization by premature beats in patients with cardiomyopathy at risk of sudden death

Premature beats can trigger ventricular arrhythmias in heart disease, but the mechanisms are not well defined. We studied the effect of premature beats on activation and repolarization dispersion in seven patients with cardiomyopathy (57 ± 10 yr, left ventricular ejection fraction 31 ± 7%). Activation time (AT), activation-recovery interval (ARI), and total repolarization time (TRT) were measured from 26 unipolar electrograms during right ventricle (RV) endocardial (early) to left ventricle epicardial (late) activation in response to RV apical extrastimulation (S1S2). Early TRT dispersion increased significantly with shorter S1S2 (1.0 ± 0.2 to 2.3 ± 0.4 ms/mm, P < 0.0001), with minimal change in late TRT dispersion (0.8 ± 0.1 to 1.0 ± 0.3 ms, P = 0.02). This was associated with an increase in early AT dispersion (1.0 ± 0.1 to 1.5 ± 0.2 ms/mm, P = 0.05) but no change in late AT dispersion (0.6 ± 0.1 to 0.7 ± 0.2 ms/mm, P = 0.4). Early and late ARI dispersion did not change with shorter S1S2. AT restitution slopes were similar between early and late sites, as was slope heterogeneity. ARI restitution slope was greater in early vs. late sites (1.3 ± 0.6 vs. 0.8 ± 0.6, P = 0.03), but slope heterogeneity was similar. With shorter S1S2, AT-ARI slopes became less negative (flattened) at both early (-0.4 ± 0.1 to +0.04 ± 0.2) and late (-1.5 ± 0.2 to +0.3 ± 0.2) sites, implying less activation-repolarization coupling. There was no difference in AT-ARI slopes between early and late sites at short S1S2. In conclusion, high-risk patients with cardiomyopathy have greater TRT dispersion at tightly coupled S1S2 due to greater AT dispersion and activation-repolarization uncoupling. Modulated dispersion is more pronounced at early vs. late activated sites, which may predispose to reentrant ventricular arrhythmias.

Effect of noise on T-wave alternans measurement in ambulatory ECGs using modified moving average versus spectral method.

Background: The modified moving average (MMA) and spectral method (SM) are commonly used to measure T‐wave alternans (TWA), but their accuracy has not been compared in ambulatory electrocardiograms (ECGs) where TWA signal‐to‐noise ratio is low. Our objective was to compare the effect of noise and signal nonstationarity on the accuracy of TWA measurement using MMA versus SM when applied to synthetic and ambulatory ECGs.

Chasing Red Herrings: Making Sense of the Colors While Mapping

A 50-year-old man presented with sustained wide complex tachycardia and was treated with cardioversion. He gave a history of 2 such episodes within the past 7 years, requiring cardioversion each time. Ten years previously, he had undergone surgical repair of tetralogy of Fallot. Review of operative notes indicated that the procedure included patch closure of the ventricular septal defect, right ventricular outflow tract resection, repair with a transannular patch, and pulmonary valve replacement with a porcine valve. Editor’s Perspective see p 557 ECG recorded during the tachycardia showed a regular wide complex tachycardia at a rate of 240 beats per minute of left bundle branch block morphology with left axis deviation. ECG in sinus rhythm showed PR interval of 200 ms, right bundle branch block with normal axis, and QRS duration of 170 ms. Echocardiography showed mild right ventricular dilatation with normal left and right ventricular function. The ventricular septum was intact, there was no pulmonary regurgitation, and the peak gradient across the pulmonary valve was 32 mm Hg. Mapping was performed using an electroanatomic system (CARTO 3, Biosense Webster, Diamond Bar, CA). Wide complex tachycardia at a cycle length of 260 ms with the same morphology as the clinical tachycardia was induced during catheter placement and could not be pace terminated. ventriculo-atrial dissociation and negative His-ventricular interval confirmed ventricular tachycardia. Surface ECG was used …

Antidromic His Capture during Entrainment of Orthodromic AVRT

A narrow QRS tachycardia with eccentric atrial activation is presented with features favoring an orthodromic atrioventricular re‐entrant tachycardia including an extranodal paraHisian response, and a short corrected post‐pacing interval to tachycardia cycle length difference following right ventricular entrainment. However, during entrainment, the H‐H interval was entrained by the pacing train several beats prior to the A‐A interval which would suggest an atrioventricular nodal re‐entry tachycardia. We discuss the diagnosis and its mechanism. (PACE 2010; 33;1153–1156)

A tale of four atrioventricular intervals

The introduction of new pacing algorithms for preserving intrinsic atrioventricular conduction has made the interpretation of normal pacemaker function at times challenging. Electrocardiographic (ECG) findings in a patient with apparent pacemaker malfunction is presented and interpretation given. During this process, the managed ventricular pacing algorithm is described to interpret the ECG findings and the potential pitfalls of this algorithm is discussed.