Abstract 202: Automated Heart Rate Complexity Calculation: Reference Ranges in Trauma and Cardiology Patients

Abstract

Objective: Heart rate complexity (HRC), measured by sample entropy (SampEn), and multiscale entropy (MSE), has been useful for detection of injury severity in off-line analysis of animal and human data. However there are no reported reference ranges for these metrics, especially as relevant to automated analysis of the electrocardiogram (ECG) Materials and methods: We developed algorithms for automated R-wave detection and HRC and heart rate variability computation. We used them to analyze a mixed data set of ECGs of various lengths from the US Army Trauma Vitals database (TV, n=71), and the Physionet Holter monitor database (n=30). The TV set was further subdivided into injured patients that received lifesaving interventions (LSI, i.e. intubation, chest tube, transfusion, n=49) or who did not (no LSI, n=22). The Physionet data were separated into these groups: (1) normal sinus rhythm, untreated (n=5); (2) normal sinus rhythm, medicated (n=9); and (3) arrhythmias, medicated (n= 16). Group 3 included patients with intermittent atrial fibrillation or ventricular tachycardia while on anti-arrhythmics. One-way analysis of variance with Tukey’s procedure was performed. Data are means ± SD, unitless unless noted otherwise. Results: see table 1 Conclusions: In agreement with previous off-line manual analysis of ECG, automated HRC analysis showed lower values in more severely injured subjects with LSIs. The reference ranges for HRC and select heart rate variability metrics will be useful for prospective evaluation of patients using automated analysis of the ECG. Table 1.