Introduction Thoracic impedance (TI) drops measured by implantable cardioverter‐defibrillators (ICDs) have been reported to correlate with ventricular tachycardia/fibrillation (VT/VF). The aim of our study was to assess the temporal association of decreasing TI trends with VT/VF episodes through a longitudinal analysis of daily remote monitoring data from ICDs and cardiac resynchronization therapy defibrillators (CRT‐Ds). Methods and Results Retrospective data from 2384 patients were randomized 1:1 into a derivation or validation cohort. The TI decrease rate was defined as the percentage of rolling weeks with a continuously decreasing TI trend. The derivation cohort was used to determine a TI decrease rate threshold for a ≥99% specificity of arrhythmia prediction. The associated risk of VT/VF episodes was estimated in the validation cohort by dividing the available follow‐up into 60‐day assessment intervals. Analyses were performed separately for 1354 ICD and 1030 CRT‐D patients. During a median follow‐up of 2.0 years, 727 patients (30.4%) experienced 3298 confirmed VT/VF episodes. In the ICD group, a TI decrease rate of >60% was associated with a higher risk of VT/VF episode in a 60‐day assessment interval (stratified hazard ratio, 1.42; 95% confidence interval (CI), 1.05–1.92; p = .023). The TI decrease preceded (40.8%) or followed (59.2%) the VT/VF episodes. In the CRT‐D group, no association between TI decrease and VT/VF episodes was observed (p = .84). Conclusion In our longitudinal analysis, TI decrease was associated with VT/VF episodes only in ICD patients. Preventive interventions may be difficult since episodes can occur before or after TI decrease.

Introduction Implantable cardioverter defibrillators (ICDs) provide daily values of thoracic impedance (TI) that are inversely correlated with fluid accumulation in the lungs. Since sustained ventricular arrhythmias (SVA) are known to have a short-term relationship with heart failure exacerbations, our objective was to assess TI trends temporally related to SVA episodes. Methods This study analyzed data daily transmitted from patients with ICD or cardiac resynchronization therapy defibrillator (CRT-D) of the nationwide Home Monitoring Expert Alliance network. Device-detected SVAs were adjudicated for appropriateness. Patients were randomly split into a derivation and validation cohort. To identify the most significant TI trend (TI-index), several numerical TI transformations were tested in a cross-sectional analysis of the derivation cohort modelling the odds of first SVA with univariate logistic regressions. In the same cohort, the threshold of the selected transformation was identified to maximize the projected specificity. The relative risk of SVA for TI-index above threshold was estimated in the validation cohort by applying Cox proportional hazard models stratified by individual patients to 60-day duration windows. Analyses were performed separately in the ICD and CRT-D groups. Results The study cohort included 2,384 patients with 69 years of age (interquartile range: 60, 77); 19% were women, 42% had coronary artery disease, and 43% had a CRT-D. After a median follow-up of 2.0 (1.0, 3.4) years, there were 3,298 appropriate SVA episodes in 727 patients (30%). The derived IT-index consisted of the percentage of 6-day intervals of the rolling average of TI values showing monotone decrease in the last 82 days. The increase in the risk of SVA was 3% (p<0.0001) per unit of percentage. The threshold of 60% of intervals with monotone decrease was associated with 99.9% projected specificity and 70.3% accuracy. In the validation cohort, the TI-Index was associated with a 42% increased risk of SVA episodes in the ICD group (hazard ratio [HR] 1.42, 95% confidence interval [CI] 1.05-1.92, p=0.023). The TI-index exceeded the 60% threshold before the episode in 38% of the detected episodes. The association was not significant in the CRT-D group (HR 0.96, 95% CI 0.62-1.47, p=0.84). Conclusions In our analysis of remote monitoring data, a specific monotonic decreasing trend of TI was temporally associated with SVA in patients with ICD. Careful monitoring of TI can identify a period of susceptibility to ventricular arrhythmias that deserves more intensive attention.