An abnormally high stiffness of large conductance arteries is a marker of increased cardiovascular risk and is typically found in patients with diabetes or hypertension. These vessels also become stiffer with advancing age [1,2], explaining the high prevalence of systolic hypertension in the elderly. The stiffening of large conductance arteries acts on blood pressure (BP) in several ways (reviewed in ). The windkessel function is hampered, leading to elevation of systolic and pulse pressure, lowering of diastolic pressure, and thus increased left ventricular afterload possibly associated with reduced coronary perfusion pressure. In addition, the propagation of pressure waves along the arterial tree is accelerated, so that the reflections generated at sites of impedance mismatch reach the ascending aorta sooner relative to the onset of ventricular ejection, resulting in a greater augmentation of central (aortic) systolic pressure, than would occur with more compliant vessels. In this latter case by contrast, forward and reflected pressure waves travel more slowly, so that reflections tend to reach the central aorta after ejection termination, thereby augmenting aortic diastolic pressure and thus contributing to coronary perfusion pressure.
Applanation tonometry is an easily applicable noninvasive method for the assessment of the pressure waveform in the ascending aorta [4,5], based on the recording of the peripheral pulse by means of a piezoelectric high-fidelity pressure sensor applied on the skin overlaying a peripheral vessel, usually the radial artery. From the raw radial pulse, a well validated algorithm [6,7] estimates the detailed shape of the pressure wave within the ascending aorta, allowing to recognize if, and in which phase of the cardiac cycle (systole, diastole or both) aortic pressure is augmented by reflected waves. With this method, it is possible to derive a systolic augmentation index (sAix), which expresses the percentage of central pulse pressure caused by reflections. A string of studies have shown that the sAix correlates with other measures of arterial stiffness such as pulse wave velocity (PWV) , increases as does the latter with age  and with cardiovascular risk factors, and has an ability to predict cardiovascular outcome . In contrast with this abundance of data centered on systolic events, the augmentation of diastolic central pressure by reflected waves has received very little attention.
We have recently developed an analysis of the diastolic part of central aortic pressure, as reconstructed from the radial pulse recorded with applanation tonometry [11–13]. In this analysis, the contribution of reflected waves to central diastolic pressure is quantified by a diastolic augmentation index (dAix), computed based on the deviation of the diastolic pressure profile from the monexponential decay that would be expected in absence of reflections. The dAix is expressed in the same units used for the sAix (i.e. percentage of central pulse pressure). We reasoned that, because changes in arterial stiffness modulate PWV and thus the travel times of pressure waves to the reflection sites and back to the central aorta, any increase in the amount of reflected energy reaching the aorta in systole should be associated with less reflected energy arriving at the same point in diastole. On the basis of this consideration, and if the sAix and dAix are true quantification of reflection phenomena, one would expect an inverse relationship between both indices.
The purpose of the present study was to verify this prediction in a sample of healthy controls of both sexes, with a wide range of ages. Specifically, because the sAix is known to increase with age and to be larger in women than in men of the same age [9,14], we tested whether differences in the opposite direction would be observed for the dAix. In addition, the effects on both indices of a change in body position was also examined, because the sAix has been shown to be larger when evaluated with the patient sitting rather than lying supine . A further incentive for the present experimental design consisted in recent results from our laboratory, which showed major effects of body position on the dAix in young women , prompting us to evaluate whether this specific finding could be extended to men and too other age groups.
Healthy men and women were enrolled by public advertisement. The recruitment was stratified by age (young: 18–39 years, middle-aged: 40–55 years, aged: 56–70 years) and sex, thus defining six experimental groups. Exclusion criteria were current smoking, any history of chronic or cardiovascular disease, office BP more than 140/90 mmHg, BMI more than 30 kg/m2, regular intake of any drug, and intake of any drug in the week preceding the examination.
The study was approved by the local Ethics Committee. All included participants had signed a consent form after receiving detailed explanations.
The study size was chosen so as to obtain 80% power to detect a two-fold difference in mean dAix between any pair of experimental groups, keeping the overall rate of type I errors below 5%. Information on interindividual variance of the dAix was available from previous studies by our group [11,15], leading to a planned inclusion of 16 patients per group.
Measurement of blood pressure
BP was measured with an oscillometric device (Stabil-O-graph; IEM Gmbh, Stolberg, Germany), in compliance with the recommendations of the European Society of Hypertension . In each of the studied positions (seated and supine), three determinations were made, and the average of the last two was recorded.
Pulse wave analysis
Pulse wave analysis (PWA) was performed in both the supine and seated position, using applanation tonometry of the radial artery carried out with the SphygmoCor device (AtCor Medical; Sydney, Australia), as previously described [11–13,15]. In short, the central element in this system is a pen-shaped piezoelectric pressure transducer, with which the radial artery is flattened against the radial bone, thus allowing recording the radial pulse. Pulse waves obtained during 10 s are then processed in order to reconstruct a central waveform using a Generalized Transfer Function, as abundantly described elsewhere [4,7].
The SphygmoCor software, provided by the manufacturer, implements a range of criteria to determine whether the quality of the recorded waveform is sufficient for subsequent analysis. These criteria include a mean pulsatile amplitude of the raw tonometric signal higher than 80 mV and a beat-to-beat variability of pulse pressure and diastolic pressure lower than 5% of average pulse pressure. We rejected any recording, which did not fulfill these criteria. All recordings were carried out in triplicate.
Pulse wave velocity
Carotid-femoral PWV was evaluated in the supine position using simultaneous noninvasive recordings of the carotid and femoral pulse with the Complior device and version 1.3.0 of the Complior SP software (Alam Medical, 94300 Vincennes, France) . All recordings were carried out in triplicate. The carotid to femoral distance was obtained with a compass, so as to avoid any influence body surface shape on the measurement.
The study took place in the morning, in the fasting state, in a quiet room. The patients were weighed, their height measured, and they were run through a short questionnaire to evaluate whether or not they had regular physical activity . After that, examinations were sequentially carried out in the seated and in the supine position, in randomized order. In each position, 15 min were allowed for stabilization. Then, brachial BP was measured, after which PWA was carried out. As carotid-femoral PWV is technically difficult to obtain in seated patients, this measurement was restricted to the supine position, where it took place in randomized order with respect to PWA.
To obtain a graphical overview of the influence of experimental conditions, the central arterial pulse waveforms were ensemble-averaged across patients and expressed in percentage of the corresponding pulse pressure, as previously described [11–13,15].
From the radial pressure waveform, the Sphygmocor software estimates ejection duration as the time from the foot of the pressure wave to the incisura. On the reconstructed central pressure waveforms, it then positions the systolic points P1 and P2 (Fig. 1). P1 is the inflection point marking the onset of the reflected systolic wave, the peak of which is indicated by P2, as shown in Figure 1. The time lag sT1r from the initial systolic upstroke to point P1 gives an estimate of twice the travel time of the systolic reflected wave. The relative importance of the forward and reflected pressure waves is expressed as a sAix according to the formula:
Equation (Uncited)Image Tools
The sAix may be positive or negative, with a positive value indicating augmentation of peak systolic pressure by reflections. Because the sAix is dependent on HR, the software also corrects for this influence by estimating the value that would be observed with a beating frequency of 75 per min (sAix@75).
Equation (Uncited)Image Tools
We complemented the above calculations by computing the amplitude of diastolic reflection waves (dAix), as previously described [11–13,15]. Briefly, the dAix corresponds to the maximal distance between the upward convexity seen on the diastolic part of the aortic or radial pressure waveform and a straight line passing through the onset of the incisura and tangent to the last diastolic part of the waveform (i.e. the vertical distance between points P3 and P4 on Figure 1). This distance is interpreted as an approximate deviation from the monoexponential decay that would be expected if there were no reflected waves in diastole. As in the case of the sAix, the dAix is expressed in percentage of aortic pulse pressure, that is
Finally, information regarding the timing of the diastolic reflected wave was captured by computing its mean transit time (dMTT), according to the formula shown in the legend to Figure 1, using the trapezoidal rule to carry out numerical integration.
Equation (Uncited)Image Tools
As subjected to statistical analysis, indices derived from PWA were mean values calculated from recordings made in triplicate. We used analysis of variance for repeated measures, with a model including age group and sex as nonrepeated fixed factors, body position as a repeated fixed factor, and patient as a random factor nested under age group and sex. In the first step, all two-way and three-way interactions between the fixed factors were included. Only significant interactions were retained in the final model. Relationships of each PWA index with HR and PWV was evaluated by simultaneously adding both variables as covariates to the final model. When omnibus F-tests were significant, further pairwise comparisons were made using Tukey's honestly significant difference (HSD, also designated as Tukey-a test), a conservative approach . The alpha level of all tests was set at 0.05, except for interaction terms (in their case, F-tests yielding a P value less than 0.1 were deemed to justify separate tests of each involved factor at each level of the other factor, such tests being then carried out with HSD and alpha = 0.05). Group summary statistics are means and SD.
Overall, 93 patients were recruited. Their demographic and hemodynamic characteristics are shown in Tables 1 and 2. To avoid cluttering with statistical symbols, the results of statistical analyses relevant to the data in Table 1 are displayed separately in Table 2. Within each age class, groups of women and men were well balanced for age, and only minor imbalances were noted between sexes, if at all, regarding body size, heart rate (HR) and BP. However, there was an uneven distribution of sedentary versus physically active patients between experimental groups. The impact of changing body position from seated to supine was to lower HR by approximately 10% and BP (whether peripheral or central) by a few mmHg.
PWV, which could only be measured in the supine position, was systematically higher in men than in women. The difference between age groups did not reach statistical significance (omnibus F: P = 0.07). However, with a model including sex, age as a continuous variable, and their interaction, the association of PWV with age was highly significant (P = 0.006).
The ensemble-averaged profiles of central aortic pressure, as recorded in the two body positions, are shown for the six study groups in Figure 2, allowing qualitative comparisons to be made between waveform morphologies. In that respect, there were no obvious differences between men and women. In both sexes, the waveforms of the middle-aged and aged had, in comparison of the young groups, a much more prominent late systolic followed by a much smaller diastolic augmentation by reflected waves. The effect of body position on the profiles was especially noticeable in young patients, and especially on diastolic augmentation, which seemed delayed and of larger amplitude when recorded in the supine, in comparison with the sitting condition.
The changes noted in the amplitudes of sAix and dAix as a function of age, sex, and body position are shown in Figure 3 (with numerical counterpart in Table 3). So as not to clutter this figure with symbols, complete results of the relevant statistical comparisons are displayed in Table 4. In comparison with values in men of the same age, the sAix of women was larger, and their dAix smaller, although this latter difference did not reach statistical significance. With advancing age, the sAix increased and the dAix decreased, an effect especially marked in the transition from the young to the middle-aged groups. The impact of changing body position was age-dependent but not sex-dependent (age-position interaction: P < 0.0001; sex-position interaction P > 0.3 for both the sAix and the dAix). In young patients, switching from sitting to supine markedly enhanced the dAix, with no statistically significant effect noted on the sAix. In the middle-aged groups, position had no statistically significant effect. In aged patients, taking on the supine position increased the sAix, but did not affect the dAix.
These observations suggest the general pattern of an inverse relationship between the sAix and the dAix, which also becomes apparent when plotting these indices against each other, as done in Figure 4a. In further corroboration of this pattern, there was a highly significant negative correlation between the respective changes induced in the dAix and in the sAix by changes in body position (Fig. 4b).
Timing of reflected waves
Detailed data on the timing of reflected waves appear in Table 3, with statistical analysis in Table 4. In short, the transit times of reflected waves reaching the aorta in both systole and diastole were larger in men than in women, decreased with age, and lengthened upon shifting from the sitting to the supine position.
Relationships to heart rate and pulse wave velocity
Standard analysis of covariance was used to evaluate the independent relationship of the sAix and the dAix with HR and PWV, after adjustment for age, sex, and position. Table 5 shows the regression coefficients thus obtained (β), which represent estimated changes of the corresponding index (expressed in % of pulse pressure) for each increase of 10 b/min for HR, or 1 m/s for PWV.
As expected, the sAix increased with PWV (positive β). In contrast, a higher PWV was associated with a lower dAix (negative β). Although both indices were related to PWV, and thus to arterial stiffness, only the dAix was clearly independent of HR. Analyses carried out separately on each age group gave entirely similar results, indicating that the lack of relationship between dAix and HR was not driven by the clustering of low dAix values in aged volunteers (Table S1 in Supplemental Digital Content 1, http://links.lww.com/HJH/A208).
Several differences noted between women and men regarding systolic and diastolic indices (Table 3) were statistically significant (Table 4). Therefore, the analysis of covariance mentioned in the previous paragraph was repeated with additional adjustment for the main covariates, which differed between sexes, namely height, weight, HR, mean central BP and PWV (Table S2 in Supplemental Digital Content 1, http://links.lww.com/HJH/A208). In these adjusted analyses, sex remained a significant predictor of sAIx, sAIx@75, sT1r, and dAix, although not of dMTT. This apparent independent effect of sex on the central pulse waveform might be explained by unmeasured factors possibly differing between men and women, such as spatial distribution of reflexion sites or regional PWV, for example in arteries of the lower limbs.
Reconstruction of the central aortic pressure wave from the tonometric recording of the radial pulse is a widely used noninvasive tool for the evaluation of arterial stiffness, the main relevant information having been so far extracted from systolic augmentation of pressure by reflected waves, via computation of the sAix or an analogous index. The present study supports that, at least in healthy adults, a similar information can be retrieved by analyzing the diastolic part of the reconstructed waveform (done here by calculating the dAix). This statement is based on two lines of evidence: the reciprocal relationship found between the systolic and the diastolic indices, both at the interindividual (i.e. age and sex impacting on the dAix in a direction opposite to that observed for the sAix) and intraindividual level (impact in opposite directions of changes in body position); the statistically strongly significant relationship of both indices with PWV, a standard measure of arterial stiffness made independently from radial tonometry. In contrast to the sAix, the dAix seems immune to the confounding influence of HR, suggesting that it might have some advantage over the former for the evaluation of arterial stiffness.
Effects of age and sex on standard indices of aortic stiffness
The variation of PWV and sAix with age and sex were as expected (see Supplemental Digital Content 1, http://links.lww.com/HJH/A208).
Reciprocal relationship of diastolic and systolic augmentation
The overall shape of the aortic pressure pulse depends on many factors, including the pattern of ventricular ejection, the elastic behaviour of the great vessels , hydrodynamic phenomena related to flow deceleration , waveform dispersion  and wave reflections. The present study has no ambition to discriminate between these. Nevertheless, it remains a prevalent concept that wave reflections play a fundamental role , and that, depending on their time of arrival, the reflected pressure waves may augment aortic pressure predominantly in systole, predominantly in diastole, or in both phases of the cardiac cycle [24–27]. The present study extends the available knowledge in two ways. First, the simple calculation of the dAix allows the relationship between systolic and diastolic augmentation to be expressed quantitatively. Second, at least in this healthy population, the reciprocal relationship between the sAix and the dAix appears quite general, both at the interindividual, (impact of advancing age and sex, Figure 4a, Figure 3, Tables 3 and 4), and intraindividiual levels (impact of changing the body position, Figure 4b, Figure 3, Tables 3 and 4). With respect to the effects of age, it is particularly noteworthy that the largest changes in the dAix occurred in the transition from the young to the middle-aged groups, in a mirror image of those observed for the sAix in this study as in previous ones cited above [1,14]. The impact of body position was age-dependent. Indeed, in young patients, shifting from sitting to supine had the effects of increasing the dAix without affecting the sAix, while changes opposite to these were noted in the aged, and none reached statistical significance in the middle-aged groups (Figure 3, Tables 3 and 4). This result agrees with findings of our recent study in young women ; it extends them to young men and shows how the picture becomes modified with advancing age in both sexes. This picture is consistent with data reported by McEniery et al. in elderly normotensive patients .
Timing of reflected waves
The values found in the present study for the transit time of systolic reflected waves (Table 3, sT1r) are in agreement with those in the literature [1,8,14,28–30]. Regarding the transit time of diastolic reflected waves, the data presented here are entirely new. The overall very similar behavior of dMTT and sT1r with respect to changes in age, sex, and body position suggests that the same factors affect the timing of both systolic and diastolic reflected waves. A simplified view would hold that these factors comprise travelling velocity and distance between proximal aorta and sites of reflection. The lower transit times in women compared with men could thus be seen as the resultant of lower velocity (as suggested by the measurement of aortic-femoral PWV, Tables 1 and 2) – and shorter travelling distance, as documented for systolic reflected waves by Mitchell et al.. The association of advancing age with reduced transit times is likely to be driven by increases in travelling velocity due to arterial stiffening, keeping in mind that this effect might be mitigated by the distal migration of reflection sites in older people .
One might be surprised by the value of dMTT (about 350 ms), which together with the measured values of PWV (8–10 m/s, Table 1) would suggest a distance of reflection sites hardly compatible with body size. However, as discussed by Westerhof et al., reflection sites in the arterial tree are not closed ends, but loci of impedance mismatch, introducing a phase delay between the Fourier components of the reflected and the incident waves. This state of affairs translates into bidirectional transit times larger than would be expected on the mere basis of PWV and vessel length.
Regardless of sex, the transition from the sitting to the supine position resulted in a striking increase of dMTT in the young and middle-aged groups. We have already observed this striking effect of posture in young women, and refer the reader to our previous publication for a discussion of its possible mechanisms .
Of interest, whenever diastolic reflections reach the aorta sooner, as indicated by a shorter dMTT (i.e. old vs. young age, female versus male sex, sitting vs. supine position), the sAix tends to increase and the dAix to decrease. This observation suggests that the reciprocal relationship between systolic and diastolic augmentation (Fig. 4) is largely driven by shifts in the timing of reflected waves, with shorter transit times resulting in more of the reflected energy reaching the aorta in systole.
The inverse relationship between sAix and dAix points to a common dependence on fundamental characteristics of arterial hemodynamics. Nevertheless, the interpretation of the former may be confounded by the influence of HR, which is not the case for the latter. Another potential difference between these indices, the calculated sAix is very sensitive to the location of the inflection point, which is not always well recognizable. In contrast, the dAix can be determined unambiguously whenever the diastolic augmentation is evident, and otherwise may be assumed to be zero. For these reasons, the quantification of diastolic aortic pressure augmentation by reflected waves could be a useful adjunct to PWA.
The authors wish to thank the staff at the Division de Pharmacologie clinique for housing the study.
Source of funding: Departmental funds.
Conflicts of interest
None of the authors has any conflict of interest in relation to the present work.
1. Mitchell GF, Parise H, Benjamin EJ, Larson MG, Keyes MJ, Vita JA, et al. Changes in arterial stiffness and wave reflection with advancing age in healthy men and women: the Framingham Heart Study. Hypertension 2004; 43:1239–1245.
2. Mitchell GF. Effects of central arterial aging on the structure and function of the peripheral vasculature: implications for end-organ damage. J Appl Physiol 2008; 105:1652–1660.
3. Feihl F, Liaudet L, Waeber B. The macrocirculation and microcirculation of hypertension. Curr Hypertens Rep 2009; 11:182–189.
4. O’Rourke MF, Pauca A, Jiang XJ. Pulse wave analysis. Br J Clin Pharmacol 2001; 51:507–522.
5. Nelson MR, Stepanek J, Cevette M, Covalciuc M, Hurst RT, Tajik AJ. Noninvasive measurement of central vascular pressures with arterial tonometry: clinical revival of the pulse pressure waveform? Mayo Clin Proc 2010; 85:460–472.
6. O’Rourke MF, Pauca A, Kon N, Tong W, Yung W, Qaasem A, Avolio A. Calibrated ascending aortic pressure wave can be derived from the radial pulse using a generalized transfer function. Am J Hypertens 1999; 12:166A.
7. Gallagher D, Adji A, O’Rourke MF. Validation of the transfer function technique for generating central from peripheral upper limb pressure waveform. Am J Hypertens 2004; 17:1059–1067.
8. Jatoi NA, Mahmud A, Bennett K, Feely J. Assessment of arterial stiffness in hypertension: comparison of oscillometric (Arteriograph), piezoelectronic (Complior) and tonometric (SphygmoCor) techniques. J Hypertens 2009; 27:2186–2191.
9. Segers P, Rietschel ER, De Buyzere ML, Vermeersch SJ, De Bacquer D, Van Bortel LM, et al. Noninvasive (input) impedance, pulse wave velocity, and wave reflection in healthy middle-aged men and women. Hypertension 2007; 49:1248–1255.
10. London GM, Blacher J, Pannier B, Guerin AP, Marchais SJ, Safar ME. Arterial wave reflections and survival in end-stage renal failure. Hypertension 2001; 38:434–438.
11. Delachaux A, Waeber B, Liaudet L, Hohlfeld P, Feihl F. Profound impact of uncomplicated pregnancy on diastolic, but not systolic pulse contour of aortic pressure. J Hypertens 2006; 24:1641–1648.
12. Dischl B, Engelberger RP, Gojanovic B, Liaudet L, Gremion G, Waeber B, Feihl F. Enhanced diastolic reflections on arterial pressure pulse during exercise recovery. Scand J Med Sci Sports 2011; 21:e325–e333.
13. Jaccoud L, Rotaru C, Heim A, Liaudet L, Waeber B, Hohlfeld P, Feihl F. Major impact of body position on arterial stiffness indices derived from radial applanation tonometry in pregnant and non pregnant women. J Hypertens 2012; 30:1161–1168.
14. McEniery CM, Yasmin, Hall IR, Qasem A, Wilkinson IB, Cockcroft JR. Normal vascular aging: differential effects on wave reflection and aortic pulse wave velocity: the Anglo-Cardiff Collaborative Trial (ACCT). J Am Coll Cardiol 2005; 46:1753–1760.
15. Gojanovic B, Waeber B, Gremion G, Liaudet L, Feihl F. Bilateral symmetry of radial pulse in high-level tennis players: implications for the validity of central aortic pulse wave analysis. J Hypertens 2009; 27:1617–1623.
16. O’Brien E, Asmar R, Beilin L, Imai Y, Mancia G, Mengden T, et al. Practice guidelines of the European Society of Hypertension for clinic, ambulatory and self blood pressure measurement. J Hypertens 2005; 23:697–701.
17. Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, et al. Expert consensus document on arterial stiffness: methodological issues and clinical applications. Eur Heart J 2006; 27:2588–2605.
18. Mader U, Martin BW, Schutz Y, Marti B. Validity of four short physical activity questionnaires in middle-aged persons. Med Sci Sports Exerc 2006; 38:1255–1266.
19. Winer BJ. Statistical principles in experimental design. 3rd ed.New York:McGraw-Hill; 1971.
20. Davies JE, Baksi J, Francis DP, Hadjiloizou N, Whinnett ZI, Manisty CH, et al. The arterial reservoir pressure increases with aging and is the major determinant of the aortic augmentation index. Am J Physiol 2010; 298:H580–H586.
21. Wang JJ, Parker KH. Wave propagation in a model of the arterial circulation. J Biomech 2004; 37:457–470.
22. Hope SA, Tay DB, Meredith IT, Cameron JD. Waveform dispersion, not reflection, may be the major determinant of aortic pressure wave morphology. Am J Physiol 2005; 289:H2497–2502.
23. Chowienczyk P. Pulse wave analysis: what do the numbers mean? Hypertension 2011; 57:1051–1052.
24. Murgo JP, Westerhof N, Giolma JP, Altobelli SA. Manipulation of ascending aortic pressure and flow wave reflections with the Valsalva maneuver: relationship to input impedance. Circulation 1981; 63:122–132.
25. Kelly R, Hayward C, Avolio A, O’Rourke M. Noninvasive determination of age-related changes in the human arterial pulse. Circulation 1989; 80:1652–1659.
26. Nichols WW, O’Rourke MF, Vlachopoulos C. Chap. 10: Contours of pressure and flow waves in arteries. In: Nichols WW, O’Rourke MF, Vlachopoulos C, editors. McDonald's blood flow in arteries. London: Hodder Arnold; 2011. pp. 225–253.
27. Nichols WW, O’Rourke MF, Vlachopoulos C. McDonald's blood flow in arteries. 6th ed.London:Hodder Arnold; 2011.
28. Mahmud A, Feely J. Effects of passive smoking on blood pressure and aortic pressure waveform in healthy young adults--influence of gender. Br J Clin Pharmacol 2004; 57:37–43.
29. Casey DP, Nichols WW, Braith RW. Impact of aging on central pressure wave reflection characteristics during exercise. Am J Hypertens 2008; 21:419–424.
30. Weber T, O’Rourke MF, Lassnig E, Porodko M, Ammer M, Rammer M, Eber B. Pulse waveform characteristics predict cardiovascular events and mortality in patients undergoing coronary angiography. J Hypertens 2010; 28:797–805.
31. Westerhof BE, van den Wijngaard JP, Murgo JP, Westerhof N. Location of a reflection site is elusive: consequences for the calculation of aortic pulse wave velocity. Hypertension 2008; 52:478–483.