Pediatric heart rate variability preceding in-hospital cardiac/pulmonary arrest

• Published in: Journal of electrocardiology Vol. 69; pp. 81 - 82
• Main Authors: Zhang, Yu-He, Helfenbein, Eric D., Babaeizadeh, Saeed, Chan, Jeffery, Knight, Lynda J., Su, Felice
• Format: Journal Article
• Language: English
• Published: New York Elsevier Inc 01.11.2021; Elsevier Science Ltd
• Subjects: Algorithms; Cardiac arrhythmia; Electrocardiography; Heart rate; Pediatrics
• ISSN: 0022-0736; 1532-8430
• DOI: 10.1016/j.jelectrocard.2021.11.007

Background In-hospital cardiac/pulmonary arrest (CPA) can lead to death if not treated immediately and contributes to poor outcomes even after successful resuscitation. Predictive indications for CPA could identify patients at risk and allow early intervention. A reduction in heart rate variability (HRV) often occurs prior to adult cardiac arrest and sepsis in preterm infants. Studying pediatric HRV changes immediately preceding in-hospital CPA has the potential to enhance risk stratification. Methods Continuous pediatric patient-monitoring data was acquired using Philips Data Warehouse Connect, and included waveforms, beat-classifications, and R-R intervals measured by the arrhythmia algorithm with 0.5 ms resolution using parabolic ECG interpolation. Record length varied between a few and 24 h. An initial database for preliminary study included 23 patient records with CPA and 23 without CPA as reference. The PhysioNet Cardiovascular Signal Toolbox provided HRV time-domain (TD) and frequency-domain (FD) calculations using 5-minute windows stepped 30-seconds. The Lomb-Scargle Periodogram was used for FD. HRV statistic trend graphs and histograms were visually compared between CPA and reference patients. The quantitative comparison between groups used 60-minute mean HRV statistics up to 5-minutes before CPA and randomly chosen for the reference group. Results The TD and FD HRV statistics providing predominant indications of lower HRV for CPA patients are shown in the table with p-values. A decrease in the HRV statistic trends before CPA was observed in some cases. Drastic increases in HRV (not included in statistics) were observed immediately (<5 min) before CPA events but were mostly related to sudden bradycardias, ectopic beat interruptions, or other irregular rhythms. Conclusions In our preliminary database, a significant difference with lower HRV was observed between in-hospital pediatric patients with impending CPA events compared to a reference set. In addition, a temporal reduction of HRV was observed in some cases prior to CPA events but requires further study and characterization. HRV measures NNiqr, SDNN, FD stats, and SD2 showed the predominant lower values for CPA patients while RMSSD and SD1 showed moderate differences. Larger databases allowing patient age, diagnosis, surgical, pharmacological, and sleep-state matching need to be developed and acute events separated from declining multi-organ failure CPA. HRV measurements may ultimately prove to be one significant input to complex multi-parameter predictive models for pediatric in-hospital CPA.