Noninvasive (Input) Impedance, Pulse Wave Velocity, and Wave Reflection in Healthy Middle-Aged Men and Women
The relation between arterial function indices, such as pulse wave velocity and augmentation index with parameters derived from input impedance analysis, is still incompletely understood. Carotid pressure, central flow waveforms, and pulse wave velocity were noninvasively acquired in 2026 apparently healthy, middle-aged subjects (1052 women and 974 men) 35 to 55 years old at inclusion. Input and characteristic impedance, reflection coefficient, the ratio of backward-to-forward pressure amplitude (reflection magnitude), and augmentation index were derived. Pulse wave velocity increased by 15% (from 6.1 to 7.0 m/s) both in men and women. In qualitative terms, input impedance evolved from a pattern indicative of wave transmission and reflection to a pattern more compatible with a windkessel-like system. In women, a decrease in total arterial compliance led to an increased input impedance in the low frequency range, whereas few changes were observed in men. Characteristic impedance did not change with age in women and even decreased in men (P<0.001) and could not be identified as the primary determinant of central pulse pressure. Augmentation index increased with age, as was expected, and was systematically higher in women (P<0.001). Reflection coefficient and reflection magnitude increased with age (P<0.001) without gender differences. We conclude that, in healthy middle-aged subjects, the age-related increase in arterial stiffness (pulse wave velocity) is not fully paralleled by an increase in arterial impedance, suggesting a role for age-dependent modulation of aortic cross-sectional area. Wave reflection increases with age and is not higher in women than in men.
The analysis of arterial stiffness and function and of pressure wave reflection received increasing attention for the past 3 decades. Several methods to describe arterial stiffness in a clinical setting emerged,1 the most investigated being pulse wave velocity (PWV) and augmentation index (AIx). Some aspects related to AIx still need further investigation, such as the systematically higher AIx in women even after adjustment for body size and heart rate, and the observation that AIx tends to levels off to a plateau value above the age of 60 years.2 Although PWV and AIx are related, they seem to reflect different aspects of arterial function.2,3
A more global view on the arterial system can be obtained from impedance analysis,4–6 requiring measurement of central pressure and flow waveforms. In addition, pressure waveforms (Pwf) can be separated into their forward and backward components.7 In a recent study in hypertensive patients, Mitchell et al8 applied impedance analysis and found an increased aortic characteristic impedance (Zc) to play an important role in the elevated pulse pressure in patients with systolic hypertension. There are no integrated large-scale studies where both impedance analysis and newer indices like PWV and AIx were simultaneously acquired.
We have set up a broad prospective, longitudinal study where we aim to assess the development and progression of cardiovascular disease in the general population, referred to as the Asklepios Study.9 The first inclusion round, on which we report here, was finished in October 2004, and the baseline data were measured in middle-aged individuals (age range: 35 to 55 years; >2500 participants) free from overt cardiovascular disease. In addition to PWV and AIx, we estimated arterial Zin and direct estimates of wave reflection.
The aims of the present investigation in apparently healthy middle-aged men and women were as follows: (1) to provide reference data of arterial input impedance (Zin) and wave reflection parameters; (2) to investigate the relation between aortic PWV and Zc; (3) to assess the contribution of aortic Zc to central pulse pressure; and (4) to investigate the relation between AIx and wave reflection.
We refer to the online supplement for an extended version of the Methods section (available at http://hyper.ahajournals.org). Only a brief summary will be given here. A total number of 2026 subjects (1052 women) met the requirements to be included into this study. The ethical committee of the Ghent University Hospital approved the study protocol, and all of the subjects gave written informed consent.
Noninvasive Assessment of Aortic Flow Waveform and Carotid Pressure
The aortic flow waveform, stroke volume (SV), and cardiac output were assessed using ultrasound (Vivid7, GE Vingmed Ultrasound) from the cross-sectional area of and blood velocities in the left ventricular outflow tract (LVOT). Carotid Pwfs were obtained using applanation tonometry at the left common carotid artery following an earlier described calibration scheme based on brachial artery tonometry. The maximum of Pwf is carotid systolic pressure (SBPCA), the systolic-diastolic pressure difference is carotid pulse pressure (PPCA). In 417 subjects, brachial artery Pwf could not be recorded with satisfactory quality within a reasonable time frame. In these subjects, mean arterial pressure was calculated from a scaled radial artery waveform using a linear regression equation to account for the brachial-to-radial pressure amplification.10
Derived and Measured Parameters
Zin was derived from time-aligned pressure and flow as the ratio of the corresponding pressure and flow harmonics. Zc was assessed in the frequency domain (Zc-FD; average of harmonics 3 to 10 with exclusion of values >3 times the median value of Zin over that range of harmonics). Because this method may introduce a bias to lower values of Zc, we also calculated Zc in the time domain (Zc-TD) following an approach proposed by Mitchell et al.11 The modulus of Zin at 0 Hz equals systemic vascular resistance (SVR). Total arterial compliance was estimated using the pulse pressure method (CPPM)12 as well as the area method (Carea).13
We calculated the reflection coefficient (Γ) of the vascular bed and used the amplitude of Γ at the heart frequency (Γ1 ) to represent the reflection coefficient. We also separated Pwf into its forward (Pf) and backward (Pb) traveling component, with their ratio (Pb/Pf) giving the reflection magnitude. We further calculated the AIx and measured carotid-femoral PWV.
Determinants of Central Pulse Pressure
To assess the major determinants of central pulse pressure, we constructed forward linear regression models with PPCA as the dependent variable. In a first model, the included independent variables were age, sex, cardiac performance (SV, heart rate, and maximal aortic flow), and the parameters related to the arterial system: SVR, aortic Zc (Zc-TD and Zc-FD), total arterial compliance (CPPM and Carea), parameters following from wave reflection analysis (Γ1 and Pb/Pf), PWV, and AIx. In a second model, we excluded the total arterial compliance CPPM and Carea to assess the contribution of Zc and wave reflection to pulse pressure. To retain the predominant parameters in both models, only parameters entering the model improving the model R2 by >2% were retained.
In the text and tables, data are given as mean values (SD). In the figure, SEMs are displayed. Subjects were subdivided into 4 half-decades of age (Q1 to Q4), defined as Q1: 35 to 40 years; Q2: 41 to 45 years; Q3: 46 to 50 years; and Q4: 51 to 56 years. Effects of age and sex were assessed with ANOVA techniques. For parameters depending on body size, height and weight were included as covariants to adjust for these confounding factors (ANCOVA) and eventually for mean arterial pressure if appropriate. Differences between men and women within an age stratum were assessed by checking for overlap of the 95% CIs of the estimated marginal mean. Nonoverlapping intervals indicated differences with P<0.05. All of the analyses were done in SPSS 12.0 (SPSS Inc).
Population and general hemodynamic data are given in Tables 1 and 2⇓. There was a relation of body length with age (P<0.001), with the subjects in Q1 being on average 3 cm taller than in Q4. Body mass index increased with age, both in men and in women. As expected, heart rate was higher in women than in men, and SV and cardiac output were lower in women, but none of these hemodynamic variables varied with age over the studied range.
Input Impedance, Zc and Total Arterial Compliance
The modulus and phase angle of Zin are displayed in Figure 1. Because body size is a major determinant of Zin, the data were adjusted by including height and weight as covariants in the statistical analysis. In men, there was a significant increase in the modulus of Zin in the low-frequency range (harmonics 1 and 2, P<0.05 and P<0.001, respectively), but the statistical power was modest (F values 3.4 and 8.1, respectively). There were pronounced oscillations in Zin modulus at the higher harmonics in the youngest subjects (most pronounced at harmonic 7), which leveled off with age. In women, the impedance modulus at harmonics 1 and 2 increased from Q1 to Q4 (all P<0.001; F value >26), whereas the oscillations at the higher harmonics (at harmonic 6) also cancelled out with age. It can be noticed that between harmonics 2 and 4, the phase angle became more negative with increasing age both in men and women (P<0.001 for harmonics 2 and 3 with F value >17; P<0.01 for harmonic 3 with F value >4.7).
Parameters describing the impedance patterns are summarized in Figure 2, with SVR being the value at 0 Hz, CPPM reflecting the impedance in the low -frequency range (harmonics 1 and 2), and Zc representing the high-frequency values of Zin. SVR increased with age and was not different between men and women. CPPM and Carea were overall higher in men (P<0.001), whereas Zc-TD and Zc-FD were overall higher in women (P<0.01). All of the parameters were adjusted for weight and height; CPPM, Zc-TD, and Zc-FD were additionally adjusted for mean arterial blood pressure. For these 3 parameters, the age–sex interaction term was significant or borderline significant (P<0.05 for Zc-TD; P<0.07 for CPPM and Zc-FD), and data were further analyzed separately. In men, CPPM (P=0.182) did not change with age, whereas it decreased in women (P<0.001). Zc-FD decreased with age in men (P<0.001) and tended to remain at the same level in women (P=0.101). Similarly, Zc-TD decreased with age in men (P<0.001) and did not change in women (P=0.657). None of the 3 parameters was different between men and women over the age range 35 to 40 years (Figure 2). Comparing the 2 methods to assess Zc, we found that Zc-TD was slightly higher than Zc-FD (P<0.001, paired t test), but their values correlated well (correlation coefficient: 0.82; P<0.0001).
Reference Indices of Wave Reflection: Γ1 and Pb/Pf
Both Γ1 and Pb/Pf increased with age in men and women (P<0.001), but there was no gender difference for these parameters (Figure 3A and 3B). Averaged for men and women, Γ1 increased from 0.414 (0.080) in Q1 to 0.463 (0.106) in Q4, whereas Pb/Pf increased from 0.449 (0.080) to 0.506 (0.097) over the same range of age. The correlation coefficient between Γ1 and Pb/Pf was 0.78 (P<0.001).
AIx and PWV
AIx increased with age but was higher in women than in men (P<0.001), even after adjustment for differences in height and systolic duration (Figure 3C). Systolic duration was higher in women (P<0.001) and tended to increase with age both in men and women (see Table S1). The transit time of the reflected wave (time delay between foot of the wave and the characteristic point) decreased with age in men and women (P<0.001), with lower values in women at all of the ages. The augmented pressure increased with age in men and women and was higher in women than in men. PWV, adjusted for mean arterial pressure, increased with age in men and women (P<0.001; Table S1) but was not different between both. The correlation coefficient between AIx and Γ1 and Pb/Pf was 0.49 and 0.59, respectively (both P<0.001).
Determinants of Central Pulse Pressure
The first model tested explained 87% of the variance of PPca, the major contributors being total arterial compliance (CPPM), SV, and SVR (see Table 3). Excluding total arterial compliance from the model, the major determinants of PPCA were Zc-TD and maximal aortic flow, explaining 44% of total variance, and the AIx, which explained an additional 26%.
The most obvious changes in Zin patterns between the age of 35 and 55 years (Figure 1) were a (modest) progressive age-related increase in SVR (Zin at 0 Hz) and a progressive “smoothening” of the oscillations in the modulus of Zin in the higher harmonic range (>5 Hz). In the phase angle, progressively more negative values for harmonics 3 to 5 with age were found, consistent with the invasive data from Murgo et al.14 Nichols and O’Rourke4 explained this shift in pattern on the basis of an asymmetrical T-tube model, where, in the young, 2 distinct reflection sites (from the upper and lower body) appear to determine the impedance spectrum. With aging and with a more rapid increase in PWV to the lower body, the timing of wave reflections from the upper and lower body becomes more similar, with the system presenting itself as a single tube with a discrete reflection site.4
Approaching the Zin pattern from the perspective of a windkessel model, a more negative phase angle, as we observed with increasing age (Figure 1B and 1D), is characteristic for an increase in the total arterial compliance of the model. This is a counterintuitive finding, because aging is generally associated with a loss in arterial compliance. It is possible that this finding is because of the fact that we calculated impedance combining carotid pressure with central aortic flow and, therefore, an artifact. We speculate, however, that the explanation could also be found in the basic assumptions underlying the windkessel model concept. In these lumped parameter models, the changes in pressure and flow take place simultaneously throughout the arterial tree,15 which is an assumption not entirely fulfilled, because there is wave travel at finite wave speed in the arterial tree. The faster waves travel (the higher the PWV), the better this assumption is fulfilled, and the more the arterial tree will resemble a windkessel model. As such, it is possible that the more negative phase angles with increasing age are also a reflection of the fact that the arterial tree is better mimicked by a windkessel model in the older age range, whereas in younger subjects with slow wave travel, a model accounting for wave travel is probably more appropriate. We speculate that this “transitional behavior” of the arterial tree also explains why the AIx (an index intrinsically determined by pressure wave travel and reflection) is probably a more useful concept in young than in older subjects, where its value tends to reach a plateau.2
Although standard textbooks display age-related changes in impedance patterns,4,5,14 few data are available in larger populations for comparison with our data. Mazzaro et al16 recently measured Zin in 71 healthy men and women aged between 20 and 69 years and reported an increase in the minimum of the frequency spectrum (fmin) with age. In this study, we did not calculate fmin, because it is often not easy to identify the minimum in the impedance spectrum, which has a resolution limited to the fundamental frequency (≈1 Hz in humans). Nevertheless, when we calculated a theoretical fmin based on the theory of wave reflections and the so-called quarter-wavelength formula,17 we found an increase in fmin from ≈3.4 Hz in Q1 to 4.5 Hz in Q4 (values averaged for men and women). These values are higher than the numbers reported by Mazzaro et al (who locate fmin <2.6 Hz for subjects of ≈65 years old)16 but are in the range reported in standard textbooks using invasive high-fidelity technology.4,5
Although the overall patterns appear qualitatively similar in men and women, there are distinct quantitative differences in Zin between both sexes. The modulus of the first 2 harmonics of Zin systematically increases with age in men and women, but the evolution is most clear in women (Figure 1). This is consistent with the data on total arterial compliance (Figure 2): CPPM systematically decreases with age in women, whereas it remains relatively constant in men. Interestingly, Zc also demonstrated a different age-related evolution in men and women: whereas it remained constant in women, Zc actually decreased in men over the studied age range. Because Zc is a parameter that can be estimated in different ways18 and that generally exhibits considerable variance, we estimated this parameter using both a time (Zc-TD) and frequency domain (Zc-FD) approach to exclude the possibility that our findings were dependent on the method used to estimate Zc. Both values correlated well, and both demonstrated the different trend between men and women in the evolution of Zc with age, although the effect was somewhat more pronounced for Zc-TD.
These findings stand, at first glance, in contrast with the systematic age-related increase in PWV (+15% from Q1 to Q4), which is considered as a good marker of true arterial stiffness. Assuming that the aorta can be approximated as a reflectionless tube with an “effective cross sectional area” (Aeff) and that carotid–femoral PWV represents the PWV of this tube, one can relate PWV to changes in stiffness (distensibility coefficient), total compliance, and Zc. PWV is proportional to 1/√DC, and a 15% increase in PWV as observed both in men and women thus reflects a decrease in distensibility of ≈25%. Zc varies proportional to PWV and inversely proportional to Aeff. In women, Zc was found to vary little over the studied age range, which suggests that the increase in PWV is paralleled by equal changes in Aeff (+15%). Assuming that total compliance is proportional to Aeff·distensibility coefficient, one therefore expects a net drop of 14% in CPPM from Q1 to Q4, which is indeed the order of magnitude that we found in this study. In men, Zc actually decreased by ≈15%, suggesting that the increase in Aeff (+30%) outweighed the increase in PWV. This increase in Aeff, together with the 25% decrease in distensibility coefficient, roughly matches with the unchanged CPPM that we found in men. Mitchell et al8 already stressed the role of a reduced Aeff as an important mechanism contributing to elevated pulse pressure in hypertension, and they suggested a role for vessel tone in the modulation of aortic dimensions and Zc. Using MRI techniques, we found previously that an (age-related) increase in PWV can accompany aortic enlargement in patients with Marfan disease but also in control subjects. The sample size in that study (26 patients and 26 control subjects) was, however, too small to differentiate between men and women.19 Nevertheless, the interpretation of Aeff is not straightforward in the context of the aorta with its complex topology, and it is difficult to pinpoint Aeff to a specific anatomic location. We could not relate the cross-sectional area of the LV outflow tract (see Table 1 for the data) to the anticipated changes in Aeff (as did Mitchell et al in their response to a comment of O’Rourke et al). The question is, however, how well LVOT dimensions reflect (changes in) aortic dimensions. Although the latter have been shown to increase with age,19–21 there was no relation between LVOT cross-sectional area and age in our (middle-aged) population. In summary, although it is acknowledged that Zc is an estimated value (it is derived from a carotid tonometry curve combined with blood flow measured in the LVOT, and its value depends on appropriate realignment in time of this pressure and flow curve) and depends on body size and inherent assumptions on aortic physiology, the different evolution with time in men and women over this specific age range is at least remarkable and deserves our further attention.
One of the goals of the study was to assess the relation between wave reflection and AIx. This relation is relatively modest, with correlation coefficients <0.6. Also, unlike for AIx, there was no gender difference in Γ1 or Pb/Pf. As mentioned before, AIx is a composite measure, and although it is related to the magnitude of wave reflection, it is obvious that other factors codetermine its value.19 Given the fact that PWV was not different between men and women and that we statistically corrected for differences in ejection duration and subject height, the origin of the gender difference in AIx is likely to be sought in gender-related differences in the relative distance to the reflection sites, differences in LV ejection patterns or other factors that are overlooked here. Possibly an approach as recently published by Westerhof et al22 to assess reflection coefficients from Pwf alone (they assume a triangular approximation for the flow waveform and apply a wave decomposition similar to what we presented here) may be applicable in clinical practice.
There has been some recent discussion on the importance of wave reflection and its contribution to elevated pulse pressure.8 The prevailing view is that high pulse pressure (isolated systolic hypertension) is caused by a progressive degeneration and dilation of the aorta and increased stiffening, causing an early return of pressure waves and boosting systolic pressure.23,24 Our data (Table 3, model 2) show that in a representative middle aged population, peak aortic flow and Zc-TD (which largely account for the amplitude of the forward pressure wave as Pf≈Zc-TD·maximal aortic flow) account for ≈44% of the variance in central pulse pressure and that wave reflection (AIx) explains an additional 26%. This supports the findings of Mitchell et al,25 who stressed the importance of the amplitude of the forward pressure wave as the major factor determining pulse pressure. However, in a model that also included total arterial compliance (model 1), the parameters determining PPCA were total arterial compliance (CPPM), SV, and SVR, explaining 83% of the variance in PPCA. One might argue that this finding is mainly driven by the inherent high dependency of CPPM on PPCA. Excluding CPPM yielded a slightly different model with Carea, SV, and Zc-TD as determinants, with the first 2 parameters accounting for 65% and Zc-TD for an additional 10% of the variance of PPCA. These results thus match with the “windkessel” notion that pulse pressure is predominantly determined by the low frequency properties of the arterial system (total arterial compliance).26,27
Finally, we also wish to stress some limitations of our work. Although the noninvasive character of the methodology is a prerequisite for large-scale studies such as ours, it also implies some methodologic limitations. We used hand-held carotid applanation tonometry recordings as surrogates for central pressure curves and combined these with flow waveforms measured at the LVOT. Although both curves were carefully realigned in time, the contours of central and carotid pressure are not entirely equal, which may have an effect on our impedance data and calculated reflection coefficients. We could not complete the tonometer calibration protocol (which includes brachial tonometry) in all of the subjects. Although we accounted for the brachial-to-radial amplification using a regression equation, this approach has introduced some additional variability in the data. On the other hand, these data sets were distributed randomly in the population, and the impact of this alternative calibration procedure in the subgroup should be limited. Also, the age range in our study is rather narrow. This has the advantage of focused analysis over this specific age range but precludes analysis of effects of aging. We did not perform a subgroup analysis in women to assess the potential effect of menopause or hormonal therapy.
Over the age of 35 to 55 years, the Zin of the systemic arterial system evolves from a pattern indicative of wave transmission and reflection to a pattern more compatible with a windkessel-like system. Indices targeting the quantification of wave reflection may therefore have more importance in young and middle-aged subjects than in the elderly. In healthy, middle-aged subjects, the magnitude of wave reflection increases with age both in men and women without any gender difference. There is also a progressive stiffening of the arterial tree as evidenced by an increasing PWV, but this evolution is not paralleled by increases in aortic Zc. This is suggestive for adaptive mechanisms modulating the (“effective”) aortic cross-sectional area. Total aortic compliance is found to be the primary determinant of carotid pulse pressure in our population, and it outweighs the contributions of aortic Zc or wave reflection. As such, measurement of central pressure and flow for the assessment of global arterial parameters and arterial impedance remains most relevant and provides important mechanistic information that is, at least, complementary to the more frequently measured PWV and the AIx.
We thank Frida Brusselmans, Femke Van Hoeke, Linda Packet, Marcel Anteunis, Jurgen Deviche, and the master thesis students for their professional help and invaluable assistance in completing the study.
Source of Funding
This research was funded by Fonds voor Wetenschappelijk Onderzoek Vlaanderen research grant G.0427.03 (for the Asklepios Study).
The first 2 authors contributed equally to this work.
- Received December 2, 2006.
- Revision received December 24, 2006.
- Accepted March 6, 2007.
Van Bortel LM, Duprez D, Starmans-Kool MJ, Safar ME, Giannattasio C, Cockcroft J, Kaiser DR, Thuillez C. Clinical applications of arterial stiffness, Task Force III: recommendations for user procedures. Am J Hypertens. 2002; 15: 445–452.
Nichols W, O’Rourke M. Vascular impedance. In: McDonald’s Blood Flow in Arteries. Theoretical, Experimental and Clinical Principles. 5th ed. London: Hodder Arnold, Oxford University Press; 2005: 233–267.
Milnor WR. Vascular impedance. In: Hemodynamics. 2nd ed. Baltimore, MD: Williams & Wilkins; 1989: 167–203.
Segers P, Verdonck P. Principles of vascular physiology. In: Lanzer P, Topol EJ, eds. Pan Vascular Medicine. Integrated Clinical Management. Heidelberg, Germany: Springer-Verlag; 2002: 116–137.
Mitchell GF, Lacourciere Y, Ouellet JP, Izzo JL Jr, Neutel J, Kerwin LJ, Block AJ, Pfeffer MA. Determinants of elevated pulse pressure in middle-aged and older subjects with uncomplicated systolic hypertension: the role of proximal aortic diameter and the aortic pressure-flow relationship. Circulation. 2003; 108: 1592–1598.
Rietzschel ER, De Buyzere ML, Bekaert S, Segers P, De Bacquer D, Cooman L, Van Damme P, Cassiman P, Langlois M, Van Oostveldt P, Verdonck PR, De Backer G, Gillebert TC. Rationale, design, methods and baseline characteristics of the Asklepios Study. Eur J Cardiovasc Prev Rehabil. In press.
Verbeke F, Segers P, Heireman S, Vanholder R, Verdonck P, Van Bortel LM. Noninvasive assessment of local pulse pressure: importance of brachial-to-radial pressure amplification. Hypertension. 2005; 46: 244–248.
Mitchell GF, Tardif JC, Arnold JM, Marchiori G, O’Brien TX, Dunlap ME, Pfeffer MA. Pulsatile hemodynamics in congestive heart failure. Hypertension. 2001; 38: 1433–1439.
Murgo JP, Westerhof N, Giolma JP, Altobelli SA. Aortic input impedance in normal man: relationship to pressure wave forms. Circulation. 1980; 62: 105–116.
Westerhof N, Stergiopulos N, Noble M. Arterial windkessel. In: Snapshots of Hemodynamics. An Aid for Clinical Research and Graduate Education. New York, NY: Springer Science+Business Media; 2004: 121–130.
Mazzaro L, Almasi SJ, Shandas R, Seals DR, Gates PE. Aortic input impedance increases with age in healthy men and women. Hypertension. 2005; 45: 1101–1106.
Segers P, De Backer JF, Devos D, Rabben SI, Gillebert TC, Van Bortel L, De Sutter J, De Paepe AM, Verdonck P. Aortic reflection coefficients and their association with global indices of wave reflection in healthy controls and patients with Marfan disease. Am Journal of Physiol-Heart Circ Physiol. 2006; 290: H2385–H2392.
Vasan R, Larson M, Levy D. Determinants of echocardiographic aortic root size: the Framingham Heart Study. Circulation. 1995; 91: 734–740.
Westerhof BE, 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, Parise H, Benjamin EJ, Larson MG, Keyes MJ, Vita JA, Vasan RS, Levy D. Changes in arterial stiffness and wave reflection with advancing age in healthy men and women: the Framingham Heart Study. Hypertension. 2004; 43: 1239–1245.