(1490) Heart Rate Variability Discriminates Outcomes After Pediatric Cardiac Arrest

Introduction: Determination of outcomes after pediatric cardiac arrest is complex and relies on clinician integration of multiple data streams over the first several days of hospitalization. However, consistent, effective, and early prognostication remains limited. This study used a data science approach to assess the performance of age-adjusted integer heart rate variability (HRVi) in discriminating death and neurologic outcomes in children post-cardiac arrest.

Methods: This is a single-center, retrospective review of patients admitted to the pediatric ICU (PICU) after in or out-of-hospital cardiac arrest (2008 – 2019) with at least 12 hours of heart rate data during the first 24 hours of admission. Patients were excluded if they died within 24 hours of arrest . HRVi was defined as the standard deviation of heart rate sampled every 1 second over 5 minutes. The primary outcome was in-hospital mortality. The secondary outcome was neurologic status on discharge assessed by Glasgow Outcome Score (GOS) and Pediatric Cerebral Performance Category (PCPC) scores.

Results: Seventy-eight patients met inclusion/exclusion criteria. The median age was 4.4 years (IQR 1.3-13.0). Thirty-one patients (40%) died during hospitalization, and 18 (23%) died within the first week of admission. Fifty-seven patients (73%) had poor GOS scores, and 58 patients (74%) had poor PCPC scores. Median HRVi in the first 12 hours of PICU admission was significantly different between survivors and non-survivors (p < 0.01). The median 12-hour HRVi had an AUROC of 0.74 (95% CI 0.62 – 0.87) and an AUPRC of 0.44 for discriminating death in the first week. Median 12-hour HRVi remained significantly associated with mortality after adjusting for age, location of cardiac arrest and illness severity by PRISM III scoring (p=0.017). Subsequent univariate analyses also showed a significant association between 12-hour median HRVi and poor neurologic outcomes by PCPC scoring (p=0.02) with an association that approached significance for outcomes by GOS (p=0.05).

Conclusions: Median HRVi over the first 12 hours of PICU admission after cardiac arrest was found to discriminate death and poor neurologic outcomes. Future validation studies are needed to assess the performance of this physiologic biomarker in predicting outcomes after cardiac arrest.