19 * Prognostic utility of microvolt T wave alternans in hypertrophic cardiomyopathy improves when assessed with QRS fractionation

Abstract

Introduction: Microvolt T wave alternans (TWA) can be arrhythmogenic in patients with cardiomyopathy, particularly when discordant TWA develops from delayed conduction. We hypothesized that QRS fractionation, as a measure of delayed conduction, will improve the accuracy of TWA in predicting ventricular arrhythmias (VA) in patients with hypertrophic cardiomyopathy (HCM). Methods: We prospectively evaluated 97 HCM patients with a preexisting ICD (age 51±15 yrs, 62 male). High resolution (1kHz sampling) 12-lead Holter recordings were made during sinus rhythm and ventricular pacing at 100, 110 and 120 bpm for 3 minutes at each rate. TWA was measured during ventricular pacing using the spectral method that quantifies the maximum alternans magnitude (Valt) and a noise threshold (k). +TWA was defined as Valt >0 µV with k >3. QRS fractionation was measured in each precordial QRS complex during sinus rhythm using custom software that quantifies the number of low amplitude peaks of <25ms duration (QRSp). Ventricular arrhythmia (VA) was assessed as cardiac arrest or appropriate ICD therapy during followup after ICD implant. Results: Over a median follow-up period of 41 (22-98) months, 20% of patients had VA. +TWA was found in 77% of patients. +TWA (21 vs 19%, p=0.54) and Valt (8.4 [2.5-18.6] vs 6.4 [2.5-10.5] µV, p=0.39) were similar between +VA and –VA patients. Sensitivity of +TWA for distinguishing +VA vs –VA was 79%, but specificity and accuracy were poor at 24 and 35%, respectively. QRSp was greater in +VA vs -VA patients (2 [1-3] vs 1[0-2], p=0.024). ROC curve analysis (area under curve 0.66, p=0.03) showed that a QRSp ≥ 2 was the optimal cutpoint for distinguishing +VA vs –VA (sensitivity 58%, specificity 65%, accuracy 64%). The proportion of patients with both QRSp ≥ 2 and +TWA was 31%. QRSp ≥2 and +TWA together achieved higher specificity (74%) and accuracy (70%) with similar sensitivity (53%) when compared to QRSp ≥ 2 alone. Patients with both QRSp ≥ 2 and +TWA had greater VA events compared to those without (33 vs 13%, p=0.03). Conclusions: In HCM, the presence of both QRS fractionation and +TWA identifies patients at risk of VA more accurately than either measure alone. The conduction delay causing QRS fractionation may promote more arrhythmogenic discordant TWA; thereby increasing arrhythmia vulnerability.