Response to Modeled Decomposition of Aortic Pressure Waveforms Does Not Provide Estimates for Pulse Wave Velocity
We thank Hermeling et al1 for their interest in our recent study on the use of a single pressure pulse wave for the estimation of aortic pulse wave velocity (PWV).2 The main aim of the study was to propose a noninvasive and also a nonintrusive means of measuring aortic PWV. Aortic PWV is conventionally estimated from transit time (TT) between the carotid and femoral arteries, and this requires exposure of the groin. In addition, in many subjects with excessive adiposity in the groin area, it is difficult to obtain reliable femoral pulses. We agree that any noninvasive method has a certain number of assumptions and compromises, but we believe that we have obtained acceptable agreement between aortic PWV estimated from the single pulse method and aortic PWV measured directly using noninvasive external tonometric sensors. This warrants consideration of the methodology for situations where the femoral pulse cannot be obtained for whatever reason. It also offers the possibility of estimation of aortic PWV in retrospective analysis of large databases where central aortic pressure has been estimated from the tonometric radial pulse but with no direct measurements of PWV.
The methodology was based on the triangulation concept of the aortic flow wave published previously.3,4 We would like to clarify some of the observations made by Hermeling et al1 regarding the waveform decomposition and associated procedures. The methodology described in the study relates to the TT estimated from the maximum lag of cross-correlation of the decomposed forward and backward waves. The correlation of 0.73 for the whole cohort (n=90) was much higher than the correlation found for all of the other methods of estimating TT, including the raw measure of the inflection point (T1) on the aortic pressure wave, the cross-correlation of the original forward and backward waves, and the delay of the processed forward and backward waves (correlation range: 0.37 to 0.55). This was outlined in the Discussion section of the article.2 In addition, the mean difference for the proposed method was found to be 3±8 ms, whereas differences for other methods ranged from 8±7 ms to 25±15 ms. We also agree with Hermeling et al1 that our assumption was essentially that of a uniform elastic tube for the aorta. In fact, it is this assumption that suggests a factor of 2 as a divisor for the lag between the forward and backward wave. This gave the results shown in Figure 6,2 where, indeed, there was a bias in the Bland-Altman plot. This prompted us to reconsider the factor of 2 as a divisor, and when we separated the cohort into age tertiles, we found that this does not apply across all ages. The factor was estimated as 1.88, 1.93, and 2.10 for each age tertile for the measured carotid-femoral path length. Although relatively close to a value of 2, when this age correction was applied, the mean difference between estimated and measured PWVs was reduced by a factor of 10 (from 0.6±1.9 m/s to 0.06±1.4 m/s), and the bias was also markedly reduced (Figure 82).
The observation regarding the distance measurement is true for any noninvasive measurement of PWV, and we agree that it is critical to obtain true values of PWV. Our purpose was to compare the estimated values of aortic PWV with those conventionally obtained using the superficial carotid-femoral distance. We understand that our crucial measure is the estimated TT and any calculation of PWV will include the distance measurement as a common factor. However, although the distances are similar for the test and estimated PWV values, the (slightly) different correlations calculated for Figures 6 and 7 are because of the reciprocal relation between TT comparisons (Figure 6) and PWV (Figure 7).
When corrected for age, the relatively high correlation of 0.83 and low mean difference (0.06 m/s) would indicate that the whole procedure could be considered a viable nonintrusive method for the estimation of aortic PWV when the femoral pulse is not available for direct noninvasive measurements.
A.Q. is an employee of AtCor Medical, Sydney.
Hermeling E, Reesink KD, Hoeks APG, Reneman RS. Modeled decomposition of aortic pressure waveforms does not provide estimates for pulse wave velocity. Hypertension. 2008; 51: e60.
Qasem A, Avolio A. Determination of aortic pulse wave velocity from waveform decomposition of the central aortic pressure pulse. Hypertension. 2008; 51: 188–195.
Westerhof B, Guelen I, Westerhof N, Karemaker JM, Avolio A. Quantification of wave reflection in the human aorta from pressure alone: a proof of principle. Hypertension. 2006; 48: 595–601.
Mitchell GF. Triangulating the peaks of arterial pressure. Hypertension. 2006; 48: 543–545.