Novel system predicts cardiorespiratory deterioration in children with single ventricle

A real-time computer algorithm has been shown to provide 1 to 2 h of advanced warning for 62 percent of all cardiorespiratory deterioration events in children with single-ventricle physiology during the interstage period, reports a study. Furthermore, the new system generates only one alarm at the bedside per patient per day.

The authors carried out this retrospective study from prospectively collected physiological data of patients with single-ventricle physiology. Deterioration events referred to a cardiac arrest that required cardiopulmonary resuscitation or an unplanned intubation. Physiological metrics were drawn from electrocardiogram (heart rate, heart rate variability, ST-segment elevation, and ST-segment variability) and photoplethysmogram (peripheral oxygen saturation and pleth variability index).

Additionally, the authors used a logistic regression model to separate the physiological dynamics of the predeterioration phase from all other data generated by the participants. They split data 50/50 into model training and validation sets to enable independent model validation.

Two hundred thirty-eight patients were admitted to the cardiac intensive care unit and stepdown units of Texas Children’s Hospital over 6 years. Nearly 300,000 h of high-resolution physiological waveform and vital sign data were obtained through the Sickbay software platform.

Overall, 112 cardiorespiratory deterioration events were recorded, with 72 children having at least one deterioration event. The optimized algorithm generated sensitive and specific risk index metric for the identification of impending events 1 to 2 h in advance of overt extremis (receiver-operating characteristic curve area, 0.958, 95 percent confidence interval, 0.950 to 0.965).

“Patients with single-ventricle physiology have a significant risk of cardiorespiratory deterioration between their first and second stage palliation surgeries,” the authors said.