Towards Ambulatory Baroreflex Sensitivity: A Comparison between Indices of Arterial Line and Photoplethysmography (Preprint)

Abstract

The ability of the body to adjust heart rate to control blood pressure is represented as baroreflex sensitivity (BRS). BRS is traditionally quantified by measuring heart rate changes (obtained via an ECG) following a change in arterial pressure (conventionally measured through an arterial line). The invasiveness of arterial line necessitates alternatives, like the volume clamp method and the less invasive pulse photoplethysmography (PPG). The PPG method is also suitable for continuous and free-living conditions.

This study aims to better understand which PPG-based features for BRS determination perform best in comparison to BRS based on the volume clamp method and gold standard arterial line and in extension, the feasibility of assessing BRS patterns over 24h by means of a single volunteer.

Twenty eight male volunteers (age 52±7 years, BMI 27±4kg/m2) equipped with four sensing modalities: 1. arterial line, 2. infrared PPG, 3 volume clamp finger pressure and 4. ECG, performed a protocol of three repetitive sessions in supine position. For the extended feasibility on continuous BRS measurement, ECG and PPG data were acquired for 24 hours in free-living conditions from a normotensive male volunteer (33 years). BRS index was calculated in the low frequency window (0.04-0.15 Hz) taking the average over all trials per intervention per subject. A transfer function was estimated with systolic blood pressure or it’s surrogate as input and heart rate (from the ECG) as output.

PPG based BRS features, rise-decay time ratio (RDRatio) and pulse arrival time (PAT), show intra-subject precision of respectively 44% and 23% with inter-subject variation of 91% and 53% respectively. The correlation of BRSPAT,P and BRSRDRatio,P to the gold standard BRSSBP,A is 0.66 and 0.56 respectively. During intervention, the correlations remain high for BRSRDRatio,P (rest: 0.75, Paced-breathing: 0.50, handgrip: 0.46) and BRSPAT,P (rest: 0.69, Paced-breathing: 0.52, Handgrip: 0.62). On the 24h data the BRSPAT,P and BRSRDRatio,P show changes during the day which are expected considering the activity and correspond to the changes seen in the systolic blood pressure. During the night a cyclic rhythm can be seen for both BRSPAT,P and BRSRDRatio,P.

This study demonstrates PPG-based PAT and RDRatio BRS as suitable surrogates for gold standard BRS derived from arterial line, showing the highest correlation and comparable intra subject coefficient variation, as well as expected changes during controlled activities and a 24-hour feasibility test in free-living conditions.

Ziekenhuis Oost-Limburg in Genk, Belgium (ethical committee approval number 16/039U)