BEAT-TO-BEAT VARIABILITY OF INVASIVE PULSE WAVE VELOCITY: TECHNICAL AND CLINICAL ASPECTS

Objective: Invasive foot-to-foot pulse wave velocity (PWV) is the reference standard for validating new non-invasive PWV-measuring devices. Because of acute haemodynamic variability and data processing issues, PWV shows beat-to-beat variability, which may have important clinical and technical implications. Though current device validation guidelines suggest averaging over at least n = 3 heartbeats to cope with variability, quantitative data on PWV's beat-to-beat variability is lacking. We aimed to quantify this variability and its impact on the confidence of the PWV estimate. Design and method: Pressure waveforms were acquired in n = 82 individuals in the ascending aorta and iliac bifurcation using a dual-tip intra-aortic catheter. For each individual, two datasets were recorded: 1) simultaneous ascending and iliac pressure acquisitions (with ECG), and 2) an additional ascending aorta pressure acquisition (with ECG). Pressure wave foot detection was performed using the intersecting interpolant method. By combining the iliac pressure recording from the first acquisition with the ascending aorta pressure recording from the second acquisition, we evaluated how a sequential acquisition protocol (e.g., catheter pull-back) affects the variability. Furthermore, we estimated how averaging over n consecutive heartbeats affects the standard deviation (SD) of such n-beats-averaged estimate of PWV. Results: PWV was 10.93 ± 4.24 and 11.12 ± 4.42 m/s (mean ± between-subject SD) for simultaneous (PWVsim) and sequential acquisitions (PWVseq; p = 0.002), respectively. The average within-subject SD ( = variability) was 0.48 ± 0.28 m/s for PWVsim and 1.47 ± 1.20 m/s for PWVseq (p < 0.001). This increase in variability was mainly due to the beat-to-beat variability of the ECG R-wave-to-pressure foot time interval, which was ∼2 times higher than that of the simultaneous foot-to-foot transit time (Figure A). Averaging over n = 3 reduced the variability of PWVsim to 0.37 ± 0.25 m/s, while n = 10.24 (median) is required to reach a comparable variability for PWVseq (Figure B). Conclusions: Variability of an invasive foot-to-foot PWV estimate increases considerably when using sequential rather than simultaneous acquisitions. This increase is caused by the high beat-to-beat variability of the duration of the pre-ejection period. To reduce the effective variability of a sequential acquisition to the same level as that of a simultaneous acquisition, averaging over an approximately threefold higher number of heartbeats is required.