Role of Pulse Pressure Amplification in Arterial Hypertension: Experts’ Opinion and Review of the Data

The left ventricle consumes energy by ejecting blood into the arterial system, thereby creating arterial blood flow and pressure. This phenomenon is easily conceived as a propagating pulse along the arterial bed. In daily clinical practice the arterial pulse, at a distinct site of the arterial tree (eg, at the brachial artery), is quantified as the variation in BP, ie, the systolic and diastolic BP. The most obvious feature of BP in the arterial system, generally neglected for decades in both clinical practice and research, is that it is a periodically oscillating wave (pulse wave [PW]) which travels from the heart toward the peripheral arteries.

The properties of the pressure PW can be readily understood when compared to a well-known wave, eg, the acoustic wave. As such, the pressure PW: (1) can be reflected and amplified, (2) has amplitude and frequency, and (3) these parameters can be analyzed in time (time-domain analysis, Figure 1a) and as sinusoidal harmonic components³ (frequency-domain analysis, Figure 1b), both approach providing analogous results.⁴

Reflections of the forward propagating pressure PW occur at multiple sites of the arterial bed because of changes in arterial properties (eg, elasticity/stiffness gradient and vasomotor tone) or in the architecture of the arterial tree (eg, branching points, calcifications). The “reflection site” is more of a statistical notion than a physically discrete site at which reflection occurs. In the recordings of the pressure PW, the multiple reflected waves are integrated as a single combined reflected wave, which sums with the forward ejected wave and thus forms the final pattern of the pressure PW (Figures 1a and 2).

The fundamental frequency (ω) of the pressure PW corresponds to the periodicity of the ejecting source, ie, to the heart rate (Figure 1a). At distinct sites of the arterial bed, the amplitude of the pressure PW corresponds to the difference between the nadir (end diastolic BP) and the zenith (peak systolic BP), ie, to the PP (Figure 1a). The augmentation pressure (AP) corresponds to that part of the PP which is attributed to the summation of the reflected wave. The augmenting effect of the reflected pressure wave is conventionally provided by the ratio of the AP over the PP, defined as the dimensionless augmentation index (AI). This quantification procedure is carried out by PW analysis (PWA) techniques and requires the identification of the inflection point (point of return of the reflected wave, Figure 1a). Ideally, the identification of the inflection point requires both flow and pressure measurement, as performed in wave separation⁵ or wave intensity analysis.⁶ The most widely used techniques of PWA use only pressure data and thus provide only approximations of the inflection point.⁵

The amplification of the pressure PW in the large elastic and smaller conduit arteries is associated with changes in the magnitude (amplitude) of each harmonic component of the wave. However, it can be conventionally understood in clinical practice as an increase in the whole amplitude of the pulse pressure as it travels distally, ie, a “gradual widening” of the PP between two sites of the arterial bed (eg, aorta and brachial artery, Figure 2). Amplification of PP with distance (spatial amplification) does not require additional energy input in the arterial system, and so, by definition, is more a distortion than true amplification, translated in an alteration in the morphology of the waveform. Mean BP remains almost unchanged (or slightly decreases because of viscous dissipation) between the two sites, reflecting the absence of any physiologically meaningful peripheral arterial resistance at the level of the large and middle sized conduit arteries.

The amplification (A) of the PW can be quantified as the ratio of the amplitude of the PP between a proximal (PP₁) and a distal (PP₂) location: A=PP₂/PP₁.

The assessment of PP amplification is based on the sequential, or ideally simultaneous, assessment of both peripheral and central BP. This need arises primarily because: (1) PP amplification is highly variable within⁷ and between subjects,⁸ and (2) the applied linear regression models for the prediction of PP amplification explain less than 70% of the variability.^8,9

The reference standard remains the invasive recording of both peripheral and central BP to detect the actual magnitude of amplification. However, this has obvious limitations and is not applicable to routine clinical practice. Noninvasive assessment of central BP has become widely available and officially recognized in recent years.¹ Although the available techniques have limitations,¹⁰ it has been accepted that they may provide further complementary data to peripheral BP, regarding the management of arterial hypertension and CV risk.^1,2

Two methods exist for noninvasive central BP assessment. The first method uses a radial-to-aorta mathematical transformation (generalized transfer function: GTF) of the pressure waveform, which has been derived and validated during invasive monitoring using frequency¹¹ or time¹² domain analyses. The GTF has been repeatedly tested and found reproducible in healthy^13,14 and in diseased^15,16 subjects, and after exercise.¹⁷ Yet, debate continues as to whether the population-based radial-to-aorta GTF is valid in all patients and conditions. This 2-step method requires: (1) recording of the radial PW by applanation tonometry, and (2) its calibration by diastolic and mean or diastolic and systolic BPs. The GTF yields accurate results when invasive peripheral BP is used¹⁸ for the calibration of the radial PW, even during physical exercise.¹⁹ However, the major problem of this technique is that noninvasive BP recordings in the upper limb (usually the brachial artery) are used to calibrate the radial PW.²⁰ The output error (at the aorta) is associated with the input errors at the radial artery attributable to: (1) the under- or overestimation of actual (intraarterial) systolic and diastolic pressures, depending on which noninvasive method is applied,²¹ and (2) the presence of brachial-to-radial pressure amplification.²² Moreover, this output error may be magnified by the GTF in the setting of slower heart rates and higher BPs.²³ As a consequence, the assessed central PP is usually underestimated.^20,22,24,25

The second method assesses systolic BP in the common carotid artery as a surrogate for the pressure in the ascending aorta. Because of the amplification phenomenon, the systolic BP is ≈2 mm Hg higher at the carotid artery than in the aorta.²⁶ This method is based on the observation that the mean and diastolic BPs, as well as the difference between them, are almost constant in the elastic/conduit arteries.^27,28 The carotid pressure PW is obtained directly at the common carotid artery (usually by applanation tonometry) and is calibrated by the noninvasively measured mean and diastolic BP. Diastolic BP can be obtained directly from readings at the brachial artery. Mean BP is optimally obtained by the integration of the brachial artery pressure curve after calibrating them with the brachial systolic and diastolic BP. If the radial pressure waveform is used, an additional error is introduced because of the presence of the brachial/radial systolic amplification.²² If no brachial or radial pressure curves are available, the mean BP can be estimated as MAP=DBP+0.4×PP.²⁹ Alternatively, the mean BP derived directly from automatic oscillometric devices seems to correlate well with mean aortic BP.²¹ This second method negates the need for a GTF. However, because of anatomic reasons, reproducible PW recordings are more difficult to obtain at the carotid rather than radial site. If brachial or carotid artery PWs are not available, one alternative may be to obtain distension waves by echo-tracking, at least up to a pulse pressure of 70 mm Hg.²⁶

Recently, it has been proposed that the second systolic peak on the radial PW curve correlates well with aortic systolic BP.³⁰ However, the application of this principle needs further validation.

Limited evidence is available regarding the normal values for PP amplification between the aorta and the arteries of the arm. Population-based data from the Anglo-Cardiff Collaborative Trial (ACCT),⁸ derived by the application of the GTF and by calibration of the radial PW by the brachial systolic and diastolic BP, obtained by a validated oscillometric device, suggest that the PP amplification values, expressed as the ratio of brachial/aortic PP, vary from 1.7 at less than 20 years of age to 1.2 at more than 80 years of age (Figure 3a). When expressed as the absolute change in mm Hg (brachial – aortic systolic AP) the amplification varies from: 20 mm Hg to 7 mm Hg (Figure 3b). Whether the relative and absolute expression of amplification have similar clinical relevance remains to be investigated. The absolute expression of amplification in mm Hg provides a more comprehensive approach of the phenomenon in terms of clinical practice; however, it might be modulated more by the level of BP.

The values of PP amplification provided by the ACCT⁸ may be potentially overestimated because of the applied methodology. Lower values are reported in the Asklepios Study,³¹ in which the method proposed by Kelly and Fitchett²⁷ was applied. Clearly, despite its demonstrated relevance for population studies, the calibration procedure still needs to be improved for individual subjects.

The widening of PP between the central and peripheral arteries was originally described in invasive studies,^3,32,33 in both healthy volunteers and patients undergoing diagnostic catheterizations. These invasive recordings established the qualitative morphological differences of PW between the central and the peripheral arteries in young healthy subjects (Figure 2). It was also suggested that the modification of the pressure waveform was, in part, associated with the arrival of the reflected wave at the various sites of the arterial bed.³³ In addition, the clear disparity in the height of the PP as the wave traveled from the central aorta to the periphery was noted, because of the gradual increase of systolic BP.

The phenomenon of amplification of the PW is not restricted to the central arteries and the arteries of the upper limb (brachial, radial). It is observed between any two, more or less remote, conduit arteries in the systemic circulation. In the field of arterial hypertension, the amplification of pressure between the central arteries (to which the major organs are exposed) and the upper limb (where the BP is conventionally measured) is of great interest and has been extensively investigated. The data presented below apply to this segment of the arterial bed and should be extrapolated to other segments of the arterial bed cautiously. Moreover, the technical limitations of noninvasive measurements may limit interpretation.

The properties of each periodically oscillating wave are determined by the characteristics of (1) the generating source, (2) the means of travel, and (3) peripheral resistance. Therefore, in theory, the characteristics of the pressure PW are attributable to left ventricular contractility and chronotropy, properties of the large (diameter, length, wall visco-elasticticity) and small (peripheral resistance, remodeling) arteries, and the blood (viscosity). Because viscosity remains constant, it is of limited importance in variations in PP amplification.

Cross-sectional observational data in healthy subjects from the ACCT^34,35 and the Asklepios Study,³¹ as well as data in subjects with hypertension³⁶ and metabolic syndrome,³⁷ have provided supportive evidence of the foregoing hypotheses. Amplification of the central-to-brachial PP is inversely related to: (1) large artery stiffness, as assessed by carotid-femoral pulse wave velocity (PWV) or total arterial compliance, (2) peripheral arterial resistance, and (3) characteristics of the reflected waves, such as the AI, the time to reflection (Tr; Figure 2), and the reflection coefficient (an index which provides a measure of the magnitude of reflection at a distinct site). The role of aortic geometry is still investigated.³⁸ Interventional studies in subjects undergoing atrial pacing during cardiac catheterization,^39,40 as well as observational studies in healthy subjects^31,35 and those with hypertension³⁶ and metabolic syndrome,³⁷ showed that the higher heart rate is associated with higher PP amplification.

The mechanism by which these parameters affect the PP amplification is largely related to the “timing–synchronization” of the forward and reflected waves. Theoretically, for the same amplitude of the forward and reflected wave, a prolongation of the cardiac ejection phase attributable to slower heart rate is associated with higher augmentation of the peak systolic BP (Figure 2). This is attributed to earlier “timing” of the two waves within the systolic phase of the ejected wave. For example, for a given brachial systolic BP of 130 mm Hg, aortic systolic BP may increase from 110 to 120 mm Hg, depending on heart rate, and the PP amplification may decrease from 20 to 10 mm Hg (Figure 2). Of note, the “timing” – arrival time of the reflected wave (Tr) depends also on the PWV (thus arterial stiffness) and the distance covered by the reflected wave. In this respect another important modulator of the PP amplification is the “arterial length,” ie, the actual distance between the heart and peripheral arteries, largely determined by height. This explains the lower amplification in shorter persons.

The relative contribution of each of the aforementioned parameters is extremely difficult to delineate in clinical practice and in experimental models, because these factors are closely interrelated. The AI measured in the aorta^31,41 or the radial artery⁴² seems to be the factor most closely associated with PP amplification. This may be because this index synthesizes the combined impact of arterial stiffness, forward ejected wave from the left ventricle, reflected wave magnitude, heart rate, and timing of the waves. Even so, all the above parameters explain less than half of the variability of PP amplification,³¹ and further investigation of the phenomenon is needed.

Aging is the main nonmodifiable factor associated with decreased PP amplification, as shown by cross sectional data from healthy subjects (Figure 3).^31,34,35 This is associated with the fact that “normal vascular aging” is the main modulator of large artery stiffening and increased wave reflections, although not in an age-dependent linear manner.³⁵

Gender is the second important nonmodifiable determinant. Females exhibit lower PP amplification than males^8,31,35 (Figure 3). This observation is in line with the following: (1) females have higher AI, (2) height is consistently and positively associated with PP amplification, and (3) height is consistently and inversely associated with aortic AI. These observations may be explained only partly by delayed timing of the reflected wave in diastole, due to greater distance from the “major” site of reflections (effective reflecting distance) in males; other gender-related factors may play important roles as well.⁴³

Accumulating evidence from cross-sectional studies suggests that subjects with traditional CV risk factors, such as hypertension,⁸ diabetes mellitus,⁸ hypercholesterolemia,^8,44 smoking,^8,45 or established CV disease⁸ have lower PP amplification, which is independent of age, gender, height, and heart rate. Aside from physical alterations in the arterial wall attributable to these CV risk factors, an acute increase in mean BP may increase arterial stiffness and pressure wave reflections, thus leading to acutely reduced amplification. Indeed, there is an inverse relationship between mean BP and PP amplification in untreated hypertensive subjects,^31,35,37 although this relationship may be altered in treated hypertensive subjects.⁴¹

Hypertensive subjects with concomitant obesity and metabolic syndrome have exaggerated PP amplification.^37,46 However, these results are mainly driven by the presence of higher heart rates and blunted pressure wave reflections in such patients.

Together with heart rate, the nonmodifiable factors (aging and gender) appear to be more important predictors of PP amplification than traditional modifiable CV risk factors (hypertension, diabetes, smoking, hypercholesterolemia).⁸ These data must be verified by longitudinal studies.

There is now mounting evidence from cross-sectional and longitudinal, as well as invasive and noninvasive studies, that central, more than peripheral, BP is associated with target organ damage and potentially with CV risk.^2,47–52

Indeed, for a given mean BP, lower PP amplification might be associated with unfavorable hemodynamic effects for the central arteries and the heart. For example (Figure 2), for a given brachial PP of 60 mm Hg, a subject with lower PP amplification is subjected to higher left ventricular afterload because of higher central PP.

Confirmatory evidence has been provided by prospective studies. In subjects with end-stage renal disease, low PP amplification, but not brachial PP, was an independent predictor of all-cause and CV mortality.⁵³ Moreover, the predictive power of PP amplification was superior to carotid PP. In untreated subjects with essential hypertension,⁵⁴ regression of left ventricular mass index after 1 year of drug treatment was independently associated with the increase of PP amplification, but not with the reduction in brachial PP. Further evidence is still awaited.

In clinical practice, arterial hypertension and BP-associated CV risk are stratified on the basis of brachial BP.¹ Because central BP tends to be more closely associated with intermediate markers of CV risk than brachial BP, classification of risk on the basis of central BP would appear to be important. This is not possible without direct assessment of the central BP, because of the marked variation in PP amplification within and between subjects.^7,8 Indeed, there is considerable overlap in aortic systolic BP between discrete categories of brachial systolic BP classified according to the international guidelines. Cross-sectional data⁸ show that 32% of men and 10% of women with normal brachial BP have aortic systolic BP similar to the aortic systolic BP of individuals with brachial BP within the limits of stage 1 hypertension. Conversely, individuals labeled as hypertensive based on brachial BP might actually have lower CV risk attributable to higher PP amplification and thus lower central BP. This apparent misclassification is likely to be clinically meaningful; confirmation by specifically designed longitudinal studies is awaited.

Isolated systolic hypertension (ISH), the most common phenotype of elevated BP in both adolescent/young and elderly hypertensives,^55,56 is primarily characterized by increased pulsatility of BP, as initially described 30 years ago,⁵⁷ rather than increased peripheral resistance. However, ISH in younger and older individuals can be distinguished clearly on the basis of PP amplification.^58,59 In the elderly (>60 years old), ISH is associated with increased aortic stiffness, early return of reflected waves, thus smaller PP amplification.⁵⁸ Nonetheless, aged individuals with ISH should be treated.⁶⁰ The underlying mechanism of ISH in adolescent/young adults is different⁵⁷ and still under investigation. PP amplification tends to be higher and wave reflection tends to be lower⁵⁹ in young subjects with ISH when compared with age matched normotensives and systolic/diastolic hypertensives. Increased stroke volume (hyperdynamic circulation) or relatively increased aortic stiffness may be the underlying cause.^55,57,59,61 Whether young subjects with ISH should be treated is not clear yet, but they should not be considered a “normal” population because diastolic BP and central systolic BP are higher in most cases^57,62 in comparison with normotensive subjects. The terms “spurious hypertension” or “pseudohypertension,”^62–64 initially proposed to describe young tall male subjects with elastic aortas who had ISH and normal central systolic BP,⁶⁴ may be applicable in special cases of extreme pulse pressure amplification, eg, >30 mm Hg. Because the normal levels of central systolic BP and PP are not yet defined, this term might be also misleading. Until the natural history of each potential subtype of ISH in adolescents and young adults is clarified, all subjects with ISH should be regarded as intermediate BP phenotypes and followed up to determine whether these individuals develop essential hypertension and target organ damage or regress to normotension.

Finally, the age-dependent decrease in PP amplification is likely to explain the low predictive value of brachial PP for CV events in young adults.³⁴

It is now accepted that the effect of antihypertensive drug treatment cannot be correctly evaluated solely by considering changes in brachial BP.^52,65 The available antihypertensive drugs reduce BP mainly by decreasing total peripheral resistance and cardiac output. However, there are additional effects of some antihypertensive drugs on central hemodynamics, beyond BP, which are associated with their ability to modify elasticity-associated arterial properties and wave reflections.⁶⁶ Two potential mechanisms may be responsible for the modification of wave reflections: (1) the reduction of the intensity of the wave (reduction of the reflection coefficient) or (2) the resynchronization of the timing of the reflected wave within systole, resulting in reduced systolic pressure augmentation. This latter mechanism may result from either a delay of the arrival of the wave (Tr), attributable to decreased PWV or a distal shift of the origin (effective reflecting distance) of the reflected wave, or of the shortening of the left ventricular ejection time (associated with tachycardia).

All classes of antihypertensive drugs might potentially decrease arterial stiffness passively, and thus decrease PWV, via the reduction of the distending mean BP. In this respect, all drugs might potentially increase PP amplification because of the resynchronization of the reflected pressure wave. However, there are important differences between the various classes of antihypertensive drugs regarding their effect on arterial stiffness and wave reflections.^67,68 As a consequence there are significant differences between the classes of antihypertensive drugs regarding their effect on PP amplification.

Most of the available evidence⁶⁹ derives from small invasive and noninvasive heterogeneous studies, which used various methodological designs, applied various methods for the assessment of PP amplification, and used various drug doses and for relatively short durations. Because modifications in arterial wall properties take place after long-term drug treatment,⁷⁰ there are considerable limitations regarding the extrapolation of the available data to daily clinical practice and it is not possible to draw definite conclusions regarding all classes of antihypertensive drugs.

The newer antihypertensive drugs (angiotensin converting enzymes inhibitors [ACEIs], angiotensin receptors blockers [ARBs], calcium channel blockers [CCBs]), as well as nitrates may increase PP amplification (Table). The common feature between these drug classes appears to be their arterial dilating capacity and their ability to reduce pressure wave reflections, as expressed by AI. Data concerning ACEIs are particularly convincing. ACEIs seem to act by decreasing wave reflections potentially attributable to chronic inverse remodelling of the small arteries leading to reduced reflection coefficients. There is also compelling evidence regarding the detrimental effect of β blockers (BBs; mainly atenolol) on PP amplification. This is largely attributed to the associated bradycardia (longer systolic duration) and the consequent resynchronization of the reflected pressure wave relatively earlier in the systolic phase (thus increasing AI).⁷¹ Indeed, a decrease in heart rate by 10 bpm is associated with an increase of aortic AI by 4%.⁷¹ For the moment there are no convincing data that newer BBs with vasodilating properties are devoid of these effects.⁷¹

Most importantly, these class-specific effects on central hemodynamics (especially of BBs) may explain in part the difference in mortality and left ventricular structure which were observed in the ASCOT (as suggested by the CAFE substudy)⁵² and REASON⁷² trials, respectively.

Theoretically, other drugs with pleiotropic effects (antiinflammatory, antioxidants) eg, statins may have potentially favorable effects on arterial properties and increase PP amplification. However, this hypothesis remains to be proven.⁷³

None.