HomeHypertensionVol. 66, No. 3Recommendations for Improving and Standardizing Vascular Research on Arterial Stiffness Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissionsDownload Articles + Supplements ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toSupplemental MaterialFree AccessResearch ArticlePDF/EPUBRecommendations for Improving and Standardizing Vascular Research on Arterial StiffnessA Scientific Statement From the American Heart Association Raymond R. Townsend, MD, FAHA, Chair, Ian B. Wilkinson, MD, DM, FRCP, FAHA, Vice Chair, Ernesto L. Schiffrin, MD, PhD, FAHA, Vice Chair, Alberto P. Avolio, BE, PhD, Julio A. Chirinos, MD, PhD, FAHA, John R. Cockcroft, FRCP, Kevin S. Heffernan, PhD, Edward G. Lakatta, MD, Carmel M. McEniery, PhD, Gary F. Mitchell, MD, Samer S. Najjar, MD, Wilmer W. Nichols, PhD, Elaine M. Urbina, MD, MS, FAHA, Thomas Weber, MD and Raymond R. TownsendRaymond R. Townsend Search for more papers by this author , Ian B. WilkinsonIan B. Wilkinson Search for more papers by this author , Ernesto L. SchiffrinErnesto L. Schiffrin Search for more papers by this author , Alberto P. AvolioAlberto P. Avolio Search for more papers by this author , Julio A. ChirinosJulio A. Chirinos Search for more papers by this author , John R. CockcroftJohn R. Cockcroft Search for more papers by this author , Kevin S. HeffernanKevin S. Heffernan Search for more papers by this author , Edward G. LakattaEdward G. Lakatta Search for more papers by this author , Carmel M. McEnieryCarmel M. McEniery Search for more papers by this author , Gary F. MitchellGary F. Mitchell Search for more papers by this author , Samer S. NajjarSamer S. Najjar Search for more papers by this author , Wilmer W. NicholsWilmer W. Nichols Search for more papers by this author , Elaine M. UrbinaElaine M. Urbina Search for more papers by this author , Thomas WeberThomas Weber Search for more papers by this author and Search for more papers by this author and on behalf of the American Heart Association Council on Hypertension Originally published9 Jul 2015https://doi.org/10.1161/HYP.0000000000000033Hypertension. 2015;66:698–722Other version(s) of this articleYou are viewing the most recent version of this article. Previous versions: January 1, 2015: Previous Version 1 Much has been published in the past 20 years on the use of measurements of arterial stiffness in animal and human research studies. This summary statement was commissioned by the American Heart Association to address issues concerning the nomenclature, methodologies, utility, limitations, and gaps in knowledge in this rapidly evolving field. The following represents an executive version of the larger online-only Data Supplement and is intended to give the reader a sense of why arterial stiffness is important, how it is measured, the situations in which it has been useful, its limitations, and questions that remain to be addressed in this field. Throughout the document, pulse-wave velocity (PWV; measured in meters per second) and variations such as carotid-femoral PWV (cfPWV; measured in meters per second) are used. PWV without modification is used in the general sense of arterial stiffness. The addition of lowercase modifiers such as “cf” is used when speaking of specific segments of the arterial circulation.The ability to measure arterial stiffness has been present for many years, but the measurement was invasive in the early times. The improvement in technologies to enable repeated, minimal-risk, reproducible measures of this aspect of circulatory physiology led to its incorporation into longitudinal cohort studies spanning a variety of clinical populations, including those at extreme cardiovascular risk (patients on dialysis), those with comorbidities such as diabetes mellitus (DM) and hypertension, healthy elders, and general populations.In the ≈3 decades of clinical use of PWV measures in humans, we have learned much about the importance of this parameter. PWV has proven to have independent predictive utility when evaluated in conjunction with standard risk factors for death and cardiovascular disease (CVD). However, the field of arterial stiffness investigation, which has exploded over the past 20 years, has proliferated without logistical guidance for clinical and translational research investigators. This summary statement, commissioned by the American Heart Association Council on Hypertension, represents an effort to provide such guidance, drawing on the expertise of experienced clinical and basic science investigators in Europe, Australia, and the United States. Recommendations made in this statement are assumed to refer to the research aspect of arterial stiffness investigations, unless accompanied by language that emphasizes clinical use as well, and are based on the grid shown in Table 1.Table 1. Applying Classification of Recommendations and Level of EvidenceTable 1. Applying Classification of Recommendations and Level of EvidenceSection 1. What Is Arterial Stiffness?Recommendation1.1.It is reasonable to measure arterial stiffness clinically by determining PWV (Class IIa; Levelof Evidence A).1Arterial stiffness is a concept that refers to the material properties of the arterial wall, which in turn has functional consequences for the artery because it affects the manner in which pressure, blood flow, and arterial diameter change with each heartbeat. In addition to the passive mechanical properties of the load-bearing structures, arterial stiffness can be modulated by functional components related to cellular processes in which wall stiffness can be affected by endothelial function through modulation of smooth muscle tone or by alterations in the integrity of the extracellular matrix. As developed in this summary statement, stiffness is measured in different kinds of arteries (muscular, elastic) and in cross section, longitudinally along the vessel, or in both directions. Often, arterial stiffness is assessed as the velocity of pulse-wave travel in a defined segment such as the aorta. However, the research questions addressed by investigations of arterial stiffness are not restricted to this use, and stiffness has been measured in most named large arteries in humans.2 Arterial stiffness is also estimated by measuring pressure or diameter in a vessel and applying 1 or several of the now extensive formulas to the data to derive a value that reflects this inherent property of all arteries.3Surrogate Measures of Arterial Stiffness and What Is Not Technically StiffnessArterial stiffness is often determined by measuring the velocity of pulse-wave travel in a segment of vessel.1 This is a valid measure, justified by equations such as the Moens-Korteweg and Bramwell Hill equations with which these measures agree.3 Other methods to measure arterial stiffness include the assessment of arterial compliance or distensibility or measures of characteristic impedance (relating pressure changes to flow changes). When arterial geometry (size and wall thickness) is known, it can be used to compute the arterial wall elastic modulus, a direct expression of the stiffness of the wall. Confusion arises when measures such as systolic pressure augmentation, which compares the first and second systolic peaks in the central aortic waveform and is sometimes reported as an augmentation index (AIx), are presented as “stiffness” parameters. Such measures are the result of several factors, including, but not limited to, arterial stiffness (described further in the Section 4).4The Arterial Wall and StiffnessArterial stiffness refers to the material properties of the arterial wall, which in turn affect the manner in which pressure, blood flow, and arterial diameter change with each heartbeat. The pressure load of each heartbeat in large conduit arteries is borne mainly by the elastin and collagen components, with less contribution from smooth muscle in the muscular arteries. Because of the anatomic arrangement of the elastin and collagen fibers, elastin engages at low distention (hence at low pressure) and collagen at higher distention (and pressure).5 The contribution of elastin and collagen to wall stiffness along the aorta varies as distance from the aortic valve increases to optimize the reservoir function of the aorta.Arterial stiffness is a major determinant of vascular impedance. Impedance relates the change in arterial pressure to the change in blood flow. Flow is determined by the presence of a pressure gradient. The relationships between time, pressure, and flow are such that local wave velocity becomes a determinant of the instantaneous relationship between pressure and flow. For elastic conduits, wave velocity is related to the stiffness of the wall, so changes in stiffness will modulate the pressure/flow relationships. The need to buffer each stroke volume and to adapt to changes in flow requires an optimal balance in the elastic and inelastic elements in the wall. Disease, aging, and other exposures typically reduce the elastic component and promote the inelastic (collagen) component such that arterial stiffness generally increases with age in most people.Changes in arterial stiffness fall into passive and active categories. Passive categories relate to arterial wall fiber elements that are stretched and recoil with each heartbeat and to heart rate (higher heart rates can be associated with increased arterial stiffness6). Active categories include endothelial function as it relates to nitric oxide and endothelin and vascular smooth muscle in which higher resting tone is associated with increased arterial stiffness.7 Inflammation, oxidative stress, and turnover in the extracellular matrix of the vessel wall are additional active contributors to arterial stiffness.8 In addition, sympathetic tone and genetic polymorphisms appear to regulate arterial stiffness in some vascular beds. The degree of the passive and active (functional) effects on wall stiffness depends on the type of artery: A greater degree of functional effects would be manifest in more muscular arteries (eg, carotid, iliac) compared with larger nonmuscular conduit arteries (eg, aorta).Section 2: Devices Used to Measure PWVRecommendations2.1.Arterial stiffness should be determined noninvasively by measurement of cfPWV (Class I; Level of Evidence A).9,102.2.PWVs measured in other vascular segments such as ankle-brachial or the determination of the cardiac-ankle vascular stiffness index is useful in cardiovascular outcome predictions in Asian populations, but longitudinal studies in the United States and Europe by these methods are lacking (Class I; Level of Evidence B).11,122.3.Single-point estimates of PWV are not recommended because there is a lack of evidence of cardiovascular outcome prediction in longitudinal studies. Measurement of PWV in other arterial segments such as carotid-radial is not recommended because it does not predict outcomes (Class III; Level of Evidence B).13Measurements of PWV are undertaken with several methodologies, some of which require sophisticated equipment (magnetic resonance imaging [MRI]) and software. These fall into 4 categories:Devices that use a probe or tonometer to record the pulse wave with a transducerDevices using cuffs placed around the limbs or the neck that record arrival of the pulse wave oscillometricallyUltrasonography approachesMRI-based approachesDevices Using a Probe or a Tonometer to Measure PWVA number of devices based on this technology are available and have been used extensively in published research. Tonometry-based techniques (eg, the SphygmoCor device, AtCor Medical, West Ryde, NSW, Australia) use a piezoelectric Millar tonometer that is placed at any 2 sites where a pulse is detectable. Only 1 tonometer is attached to the unit, so PWV measurements require 2 sequential 10- to 20-second readings, gated to the ECG, to be taken. The average transit time (TT) is then derived with the R wave of the ECG used as a reference point, and PWV is calculated from the inputted distance measurement. The SphygmoCor device has been used in the Anglo-Cardiff Collaborative Study of arterial stiffness14 and the Chronic Renal Insufficiency Cohort (CRIC) study of chronic kidney disease,15 as well as in other cohorts and intervention studies. Newer versions of this device use a cuff and tonometer system to record simultaneous pressure waves.16 Published reproducibility of the PWV with the SphygmoCor, as judged by Bland-Altman plot analysis, is good.17Mechanotransducer-based techniques (eg, Complior, ALAM Medical, Vincennes, France) use similar principles but allow simultaneous measurement between sites with distention sensors. The Complior software provides an online pulse-wave recording and automatic calculation of the PWV.18 This device has been used extensively in epidemiologic studies in Europe and has provided much of the early outcome data relating PWV to CVD risk. The published reproducibility of the PWV with the Complior, as judged by Bland-Altman plot analysis, is good.19Other tonometry-based devices (eg, PulsePen, DiaTecne, Milan, Italy) use an ECG signal and a handheld tonometer (similar to the SphygmoCor) to perform cfPWV measures. The PulsePen has been used in the Predictive Values of Blood Pressure and Arterial Stiffness in Institutionalized Very Aged Population (PARTAGE) study conducted in elderly patients in France and Italy.20 The reproducibility of the PulsePen, as judged by Bland-Altman plot analysis, is good.21Still other tonometry-based devices (eg, those used by Cardiovascular Engineering, Inc, Norwood, MA) use a custom device to measure PWV with tonometric methods. The system uses the foot-to-foot measure of carotid and femoral pressure waveforms, with distance measures to the carotid artery site and femoral artery site calculated from the sternal notch. The ECG QRS complex is used as the timing onset point, and the elapsed time to the carotid pressure waveform foot and the femoral pressure waveform foot is calculated and divided into the distance measurement. This system has been used in the Framingham22 and Reykjavik23 studies, as well as other cohorts and intervention trials. Reproducibility of the PWV by this method is reportedly good (Gary F. Mitchell, MD, Cardiovascular Engineering, Inc, Norwood, MA; personal communication, June 1, 2015).Devices Using Cuffs Placed Around the Limbs or the Neck That Record Pulse-Wave Arrival OscillometricallyOscillometry-based devices (eg, VP1000, Omron Healthcare, Kyoto, Japan) rely on 4 oscillometric cuffs placed on both arms (brachial) and ankles to calculate brachial-ankle PWV (baPWV; measured in meters per second). It also provides an ankle-brachial index (ratio of systolic pressure in the ankle to that of the brachial artery; a marker of peripheral arterial disease when this ratio is <0.9). Newer models (eg, VP2000) have additional probes that can be secured in place (with straps) that detect carotid and femoral pulses simultaneously (ie, both probes capture the same pulse wave) by tonometry. ECG leads are attached, as is a phonocardiographic microphone (whether the measurements are being done by oscillometry or tonometry). The subject’s age, height, and sex are entered into the software, and the distance estimate is calculated with the use of statistical norms (based on Japanese individuals). The Omron device has been used in prospective observational studies, mainly in Asia, and for independently predicting loss of kidney function,24 CVD,25 and all-cause death.26 Published reproducibility of the PWV with the VP1000, as judged by Bland-Altman plot analysis, is good.27Cuff-based devices (eg, Mobil-O-Graph, IEM, Stolberg, Germany) capture brachial blood pressure (BP) and brachial waveforms (casual and at 24 hours) to estimate central aortic pressures and to estimate cfPWV.28,29 The Mobil-O-Graph 24-hour pulse-wave analysis ambulatory BP measurement device uses a proprietary algorithm to obtain conventional brachial BP readings, after which the brachial cuff is inflated to the diastolic BP level and held constant for ≈10 seconds to record the pulse waves. Subsequently, central pressure curves are obtained with the use of a transfer function. To estimate aortic PWV, several parameters from pulse-wave analysis, along with wave separation analysis, are combined in a proprietary mathematical model incorporating age, systolic pressure, and aortic characteristic impedance.30 The Mobil-O-Graph aortic PWV values have been validated by direct intra-arterial measurement in the catheterization laboratory.31 Reproducibility of the Mobil-O-Graph, as judged by Bland-Altman plot analysis, is good.32Some cuff-based devices (eg, Vasera, Fukuda Denshi, Tokyo, Japan) use cuffs on all 4 limbs and gate the timing for the pulse-wave arrival at the ankle relative to the heart using phonocardiography through a small microphone taped onto the chest.33 In addition to the cardio-ankle vascular index, which is derived from the cardio-ankle PWV, it provides an ankle-brachial index. This device has been used mainly in Japan for longitudinal studies of dialysis patients11 and in community studies of cognitive decline.34 Reproducibility of the Vasera, as judged by Bland-Altman plot analysis, is good.35Ultrasonographic ApproachesUltrasonography can be used to assess vessel distention and derived stiffness indexes or flow waveforms to calculate PWV. Distention waveforms can be assessed with ultrasound transducers at a variety of locations, but often the carotid and femoral sites are used. Although some parts of the aorta itself are assessable, measurements in the thoracic aorta are technically challenging. An average change in cross-sectional area of a vessel can be derived from the distention waveform with dedicated software (eg, ARTLAB, ESAOTE, Genoa, Italy). Using a value for the pulse pressure (PP), the operator can determine distention and compliance. Brachial artery pressure often is used rather than local PP, which may introduce inaccuracies, as may any delay between distention and BP assessment. Pulse-wave speed (c) and other indexes of elasticity such as incremental elastic modulus can also be derived, as discussed earlier. It is worth noting that most ultrasonographic systems and software produce a time-averaged waveform, and mathematically, this will yield different values for stiffness indexes compared with calculating distention beat by beat and then averaging.In addition, ultrasonography is used to assess local (cross-sectional) distensibility of vessels such as the carotid artery. B-mode ultrasonography, video analysis, and echo-tracking methodologies are commonly used approaches.36,37 The online-only Data Supplement (Section 6) has more discussion of this aspect and device comparisons (Table 6.4 in the online-only Data Supplement).Doppler ultrasonography may be used to record flow waveforms from accessible sites from which PWV can be estimated in a manner similar to PWV based on pressure waveforms. Waveforms may be recorded either sequentially with ECG gating or simultaneously.38 Typically, 1 ultrasound transducer is clamped to the left side of the neck to insonate the site of the left subclavian artery or carotid artery, and the second transducer is secured on the abdomen, insonating the abdominal aorta above the bifurcation. Distance is measured from the suprasternal notch (SSN) to the location of the second transducer. This can be challenging because the angle of insonation makes it difficult to reliably determine where the abdominal aorta is being interrogated in most (obese) people. The foot of the flow wave from each of the recording sites is used, and the time elapsed in milliseconds is calculated. There is no set duration of recording, but averaging several beats (commonly 5–10 beats) is beneficial to increase the accuracy of the measurement.39 Identifying the foot of the flow wave can be more challenging than identifying the foot of a pressure wave. However, such techniques have shown independent predictive value for cardiovascular outcomes and death in longitudinal studies of diabetics,39 the healthy elderly,40 and a general population.41 Published reproducibility of ultrasonography-based PWV, as judged by Bland-Altman plot analyses, is good.42,43MRI-Based ApproachesMRI can be applied in much the same way as ultrasonography to determine distention-based indexes or PWV. It has the advantage of being able to assess almost any vessel and providing more accurate distance and area estimates (Vessels can always be “cut” in a perpendicular manner). However, these advantages are offset by poorer time and spatial resolution and cost.Phase-contrast MRI (PC-MRI) can be used to acquire blood flow velocity maps along any given anatomic plane. When the gradient direction is applied exactly perpendicular to the cross-sectional vessel plane (“through-plane” velocity encoding), flow can be measured through the vessel cross section. Such an approach can be used to compute the time delay between the onset of flow in the ascending and descending thoracic aorta, which can be simultaneously interrogated in cross section in a properly prescribed anatomic plane. Alternatively, the gradient direction can be prescribed in plane with the vessel flow axis, allowing the acquisition of a velocity map along the length of the vessel. This approach allows the measurement of the spatiotemporal flow profile along the length of the vessel, thus allowing the computation of PWV. This approach can be easily applied to the thoracic aorta in the “candy-cane” plane.PC-MRI sequences require a user-defined velocity-encoding sensitivity, which should be as low as possible to minimize noise during the acquisition yet higher than peak flow velocity in the region of interest to avoid aliasing. Although velocity-encoding sensitivity should be tailored to individual measurements, a velocity-encoding sensitivity of 130 to 150 cm/s allows adequate interrogation of thoracic aortic flow in most cases. PC-MRI data are acquired over several cardiac cycles, and consistent cardiac timing in each cycle is assumed. Adequate PC-MRI flow measurements require careful attention to technical details, including the recognition and minimization of sources of error such as phase-offset errors caused by inhomogeneities of the magnetic field environment (short-term eddy currents),44,45 signal loss resulting from turbulent flow, partial volume averaging resulting from limited spatial resolution, and signal misregistration caused by in-plane movement of the aorta and pulsatile flow artifacts. The temporal resolution of PC-MRI flow measurements should be maximized, which requires data collection over multiple cardiac cycles. This is usually achieved by prolonged (several minutes) acquisitions during free breathing. Various alternative techniques have been proposed for fast, real-time assessments of PWV.46–49 More research is needed into the optimal algorithm to measure the time delay between the foot of the flow waves with PC-MRI.A second approach to measure arterial stiffness with MRI involves the assessment of arterial distention, which can be paired with pressure measurements to obtain local arterial compliance and distensibility. Steady-state free-precession techniques provide high contrast between the arterial lumen and arterial wall and allow automatic segmentation of aortic lumen throughout the cardiac cycle. Such approaches can be used to assess ascending aortic properties as long as simultaneous (or quasi-simultaneous) central pressure recordings are performed. Unfortunately, tonometric arterial pressure recordings are difficult within the MRI suite because available tonometry systems are not MRI compatible. Good reproducibility of PWV by PC-MRI has been reported, with intraclass correlation coefficients of ≈0.90.50Many of the devices reviewed in this section can also be used to capture waveforms for central aortic pressure-wave analysis. Section 4 in this executive summary and Section 4 in the online-only Data Supplement provide greater detail.Regardless of the approach used, it is critical to include an accurate measurement of BP at the time of stiffness measurement because mean arterial pressure (MAP) is an important determinant of stiffness (Section 7 and Recommendation 7.1). Reproducibility is generally good, and most devices and approaches have been in use for at least a decade. Other approaches to measuring arterial stiffness are covered in Section 2 in the online-only Data Supplement.Section 3. Why Is Arterial Stiffness Important?Recommendation3.1.It is reasonable to measure arterial stiffness to provide incremental information beyond standard CVD risk factors in the prediction of future CVD events (Class IIa; Level of Evidence A).10Arterial Stiffness as a Predictor of Future Cardiovascular RiskStiffening of the central arteries has a number of adverse hemodynamic consequences, including a widening of PP, a decrease in shear stress rate, and an increase in the transmission of pulsatile flow into the microcirculation. These effects have a number of detrimental consequences that may, in part, explain mechanistically why stiffness is a predictor of risk. Numerous studies involving various disease-specific and community-based cohorts have demonstrated that higher cfPWV is associated with increased risk for a first or recurrent major CVD event.9,10 Consideration of cfPWV substantively reclassifies risk in individuals at intermediate risk for CVD, suggesting that consideration of cfPWV provides novel and clinically relevant information beyond that provided by standard risk factors.10,22 In addition, small studies have demonstrated that persistent elevation of cfPWV during treatment for hypertension or CVD is associated with high risk for an adverse outcome in those with established disease.51,52 The added benefit of cfPWV in risk prediction models may be a manifestation of the relatively modest relation between cfPWV and standard risk factors other than age and BP.53 In a recent individual-participant meta-analysis, higher cfPWV was shown to be associated with increased risk for coronary heart disease, stroke, and composite cardiovascular events. Importantly, relative risk was strongest in younger individuals, in whom an opportunity exists for early identification, lifestyle modification, and possible mitigation or prevention of further potentially irreversible deterioration of aortic structure and function.10HypertensionThe association between arterial stiffness and hypertension is well established.54–58 There is a widely held belief that increased aortic stiffness in hypertensive individuals is largely a manifestation of long-standing hypertension-related damage that stiffens the large arteries. A recent analysis from the Framingham Heart Study found that higher arterial stiffness, as assessed by cfPWV, was associated with BP progression and incident hypertension 7 years later.54 However, higher BP at an initial examination was not associated with progressive aortic stiffening, suggesting that aortic stiffness is a cause rather than a consequence of hypertension in middle-aged and older individuals. These results and several additional studies provide strong evidence in support of the hypothesis that arterial stiffness represents a cause rather than a consequence of hypertension and underscore the importance of better defining the pathogenesis of aortic stiffening.55–58High aortic stiffness is associated with increased BP lability.59–61 A stiffened vasculature is less able to buffer short-term alterations in flow. Increased aortic stiffness is also associated with impaired baroreceptor sensitivity.59,62–64 Together, these limitations may result in potentially marked alterations in BP as cardiac output changes during normal daily activities such as changes in posture and physical exertion.65Cardiac DiseaseExcessive arterial stiffness represents a compound insult on the heart. Aortic stiffening increases left ventricular (LV) systolic load, which contributes to ventricular remodeling and reduced mechanical efficiency. This leads to an increase in myocardial oxygen demand,66 compounded by a reduction in diastolic coronary perfusion as PP widens and diastolic BP decreases with aortic stiffening.67 Arterial stiffening may be associated with impaired measures of LV diastolic function,68,69 which may increase cardiac filling pressure and further limits coronary perfusion. Finally, arterial stiffness is associated with atherosclerosis,70–73 which may further impair ventricular perfusion, possibly leading to catastrophic reductions in ventricular function during ischemia.67Arterial stiffness is associated with diastolic dysfunction and diastolic heart failure resulting from direct effects of abnormal load and loading sequence on myocyte contraction and relaxation and indirectly through ventricular hypertrophy.69,74–78 Diastolic dysfunction increases filling pressures and thus may increase load on the atria, which will contribute to atrial hypertrophy and fibrosis and ultimately to atrial fibrillation.79 Arterial stiffness is independently associated with an increased risk of heart failure80 and is increased in patients with established heart failure regardless of whether LV function is preserved or impaired.81–83Peripheral Vascular FunctionArterial stiffness (arteriosclerosis) is associated with atherosclerosis, although the association is not strong and the 2 processes should be viewed as distinct pathophysiological entities. Aortic stiffening may increase the risk for development of atherosclerosis as a result of atherogenic hemodynamic stresses associated with a stiffened aorta, including increased pressure pulsatility and