Beat-to-beat variability of aortic pulse wave velocity: implications for aortic stiffness measurements

Figures

• 

  FIGURE 1. Method of calculation of the transit time (TT) for simultaneously (panel a) and sequentially acquired (panels b and c) proximal and distal pressure waveforms. (Panel a) Because waveforms are acquired simultaneously, each heartbeat's TT can be calculated as the difference between the foot times of the distal and proximal heartbeat waveforms. This yields a number of aortic pulse wave velocity (aPWV) estimates that corresponds to the total number of heartbeats in the recorded signals. (Panels b and c) Because waveforms are acquired sequentially, there is no correspondence between heartbeats in the proximal (panel b) and distal waveforms (panel c). Therefore, a TT value can be obtained for every possible combination of arrival times of the foot of the distal and proximal waveforms, as indicated in this figure's example by nine grey arrows connecting the three tproximal values in panel b with the three tdistal values in panel c. Hence, the total number of aPWV estimates for sequentially acquired pressure waveforms corresponds to the product of total number of heartbeats in the proximal and distal signals (3 × 3 = 9 estimates in this figure's example).

• 

  FIGURE 2. Effect of averaging on the beat-to-beat variability of invasive aortic pulse wave velocity (aPWV). Averaging reduces the variability of aPWV from sequential pressure acquisitions at a much faster rate than that of aPWV from simultaneous acquisitions. (Panels a–c) The effect of averaging is shown in terms of median within-subject standard deviation (note that outliers are not shown to improve figure readability. See Supplementary material, Figure S5 for complete graphs). (Panels d–f) The within-subject standard deviation of the nbeats-averaged aPWV is normalized with respect to that of the nonaveraged aPWV (i.e., nbeats = 1) to allow for an easier estimation of α, which is shown as best fit value [95% confidence interval].

• 

  FIGURE 3. Relationship between blood pressure and invasive aortic pulse wave velocity (aPWV). Each datapoint represents a single heartbeat from a patient of our cohort. Both aPWV and blood pressure are detrended (i.e., the mean over the entire acquisition is subtracted from individual heartbeat values) to aid comparability between subjects. Nondetrended plots are presented in Supplementary material, Figure S6. Note that for aPWVIT and aPWV2ndD, single heartbeat aPWV readings significantly deviated from the within-subject mean (panels a, c, d, and f), compromising the relationship with blood pressure.