Changes in paced signals may predict in-hospital cardiac arrest.

Abstract

An increasing number of patients with chronic illnesses have implanted cardiac rhythm devices such as pacemakers and implantable cardioverter-defibrillators (ICDs). This study was conducted to identify potentially useful predictors of in-hospital cardiac arrest (I-HCA) within paced electrocardiogram (ECG) signals from cardiovascular patients with implanted medical devices. In this retrospective study of 17 subjects, full-disclosure ECG traces prior to the time of documented I-HCA were analyzed to determine R-R intervals and QRS durations (QRSd). Ventricular paced QRSd prolongation was observed prior to I-HCA in 10/16 (63%) subjects. QRSd was significantly greater immediately preceding cardiac arrest than during each of the 8hours prior to cardiac arrest (P<0.05). Heart rate changes (measured using standard deviation) within 15 minutes of cardiac arrest were significantly greater in subjects with pulseless electrical activity (PEA)/asystolic arrest compared to those with cardiac arrests due to ventricular tachycardia/ventricular fibrillation (VT/VF) (10.13vs 3.31; P=0.024). Significant differences over the 8hours preceding cardiac arrest in heart rate (74vs 86beats/min; P=0.002) and QRS duration (172ms vs 137ms; P<0.001) were observed between subjects with initial rhythms of VT/VF and those with initial rhythms of PEA/asystole. Patterns of diagnostic ECG features can be extracted from the telemetry data of patients with implanted medical devices prior to adverse events including I-HCA. The detection of these significant changes might have an immediate prognostic impact on the timely treatment of some patients at risk of adverse events.