Multiple Autonomic and Repolarization Investigation of Sudden Cardiac Death (MARIA-SCD): a 10 years prospective, single blinded study in patients with reduced ventricular function and controls

Abstract

Background: Left ventricular ejection fraction (LVEF) at a single time point is the standard method to identify patients with risk of sudden cardiac death (SCD). Autonomic evaluation and assessments of cardiac electrical substrate have been introduced successfully to predict mortality. Progression of autonomic dysfunction might help identifying high-risk patients. Objectives: A combined assessment of autonomic tone and changes in cardiac electrical substrate at multiple time points should add prognostic information to standard methods using LVEF. Methods: We enrolled patients with ischemic (ICM) and dilated cardiomyopathy (DCM) with LVEF ≤ 50 % and control subjects (CSU) with LVEF > 50 %. The protocol included initial testing (I1) after establishing an optimal medical treatment, a second testing after 3 years (I2), and a final control (maximum 10 years). We assessed autonomic function by pharmacological baroreflex testing (BRS), by short-term spectral analysis of heart rate variability (LF/ HF) and long-term time domain analysis (SDNN). Cardiac electrical substrate was evaluated by exercise Microvolt T-wave alternans (MTWA), signal-averaged ECG (SAECG), and corrected QT-time (QTC). Results were interpreted blinded, and categorized using pre-specified cut-points. The primary outcomes were cardiac death (CD), arrhythmic death (AD) and resuscitated cardiac arrest (RCA). We compared adjusted multivariate models based on combinations of parameters with an univariate “LVEF only” model to identify patients with high risk of CD, AD and RCA using a single time point (I1). We performed similar analysis for multiple time points (I1 and I2) using a time varying covariate model. Results: We studied 210 patients (age 59 ± 10 years, 82 % male): Group 1: n = 120 ICM, Group 2: n = 60 DCM, Group 3: n = 30 CSU. The median follow-up was 7 years (0.4-9.7). CD, AD and RCA was observed in 31, 16 and 18 % of ICM and DCM patients, respectively. In the single time point model analysis, MTWA, BRS, SDNN, LF/ HF, QTC added significant information regarding CD and AD and MTWA only for AD/RCA. In the time varying covariate model, MTWA (p < 0.001) and BRS (p = 0.042) added significant information regarding CD and AD and MTWA only (p < 0.001) for AD/RCA. Conclusion: In initial single time point models, MTWA, BRS, SDNN, LF/HF and QTC were significantly related to cardiac and especially arrhythmic mortality. In addition to LVEF, MTWA adds the most powerful information in multiple risk assessment of cardiac patients during ultra-long follow-up.