Non-invasive real-time autonomic function characterization during surgery via continuous Poincaré quantification of heart rate variability

Abstract

Heart rate variability (HRV) provides an excellent proxy for monitoring of autonomic function, but the clinical utility of such characterization has not been investigated. In a clinical setting, the baseline autonomic function can reflect ability to adapt to stressors such as anesthesia. No monitoring tool has yet been developed that is able to track changes in HRV in real time. This study is a proof-of-concept for a non-invasive, real-time monitoring model for autonomic function via continuous Poincaré quantification of HRV dynamics. Anonymized heart rate data of 18 healthy individuals (18-45 years) undergoing minor procedures and 18 healthy controls (21-35 years) were analyzed. Patients underwent propofol and fentanyl anesthesia, and controls were at rest. Continuous heart rate monitoring was carried out from before aesthetic induction to the end of the surgical procedure. HRV components (sympathetic and parasympathetic) were extracted and analyzed using Poincaré quantification, and a real-time assessment tool was developed. In the patient group, a significant decrease in the sympathetic and parasympathetic components of HRV was observed following anesthesia (SD1: p = 0.019; SD2: p = 0.00027). No corresponding change in HRV was observed in controls. HRV parameters were modelled into a real-time graph. Using the monitoring technique developed, autonomic changes could be successfully visualized in real-time. This could provide the basis for a novel, fast and non-invasive method of autonomic assessment that can be delivered at the point of care.

Keywords: Autonomic function; Heart rate variability; Intraoperative monitoring; Poincaré; Real-time monitoring.

Fig. 1

Poincare plots before (‘Stage 1’—left) and after (‘Stage 2’—right) anesthesia. Stage 1 anesthesia illustrates baseline data and it is defined as 5 min of recording prior to induction of anesthesia. Stage 2 anesthesia illustrates maintenance of anesthesia and it consists of 5 min taken during anesthesia. Changes in the plot shapes indicate changes in parasympathetic (SD2) and sympathetic (SD1) tone. HR(t) indicates the HR at the first beat, whereas HR(t + 1) indicates the HR of the next heartbeat; each data point therefore represents the relationship between HR of two successive heart beats

Fig. 2

Boxplots of average SD1 and SD2 before (‘Stage 1’—left) and after (Stage 2’—right) anesthesia, indicating changes in parasympathetic and sympathetic tone between the two stages. SD1 sympathetic function, SD2 parasympathetic function

Fig. 3

Boxplots of SD1 and SD2 from the resting controls taken at two time points during the recording, ‘Stage 1’ (left) and ‘Stage 2’ (right). SD1 sympathetic function, SD2 parasympathetic function

Fig. 4

Sliding window analysis of SD1 and SD2 of a patient undergoing anesthesia using a 20-s window, indicating parasympathetic and sympathetic tone in real-time. SD1 sympathetic function, SD2 parasympathetic function. HRV is measured in SD, represenitng geometrical measures of distribution of the data points along the identity line. LMA laryingeal mark airway, Fluclox flucloxacillin, NDAN ondansentron

Fig. 5

Sliding window analysis of SD1/2 of a control subject during a resting period using a 20-s window, indicating parasympathetic and sympathetic tone in real-time. SD1 sympathetic function, SD2 parasympathetic function. HRV is measured in SD, represenitng geometrical measures of distribution of the data points along the identity line

Fig. 6

Continuous monitoring data for heart rate and mean arterial pressure corresponding to subject undergoing anaesthesia whose HRV monitoring data is depicted in Fig. 4. LMA laryingeal mark airway, Fluclox flucloxacillin, NDAN ondansentron