Journal of Human (2000) 14, 533–535  2000 Macmillan Publishers Ltd All rights reserved 0950-9240/00 $15.00 www.nature.com/jhh COMMENTARY Large arterial stiffness: an important therapeutic target

JR Cockcroft1 and IB Wilkinson2 1Department of Cardiology, Wales Research Institute, University Hospital, Heath Park, Cardiff CF4 4XN, UK; 2Clinical Pharmacology Unit, University of Edinburgh, Western General Hospital, Edinburgh EH4 2XU, UK

Keywords: arterial stiffness; pulse wave analysis; drug therapy; ; wave form

In the current issue of the Journal of Human Hyper- closely related to diastolic pressure, is the product tension, Mahmud and Feely1 report on the use of of peripheral vascular resistance and cardiac output. pulse wave analysis (PWA) to assess the effect of the , however, is mainly determined by angiotensin II receptor antagonist, valsartan, on large compliance and the pattern of left ven- arterial stiffness in patients with resistant hyperten- tricular ejection. Therefore, in middle-aged and sion. This timely report emphasises the importance older subjects, the available evidence suggests that of arterial stiffness as a predictor of cardiovascular arterial stiffness is a more important predictor of car- risk, and the changing perception of the various diovascular risk than peripheral vascular resistance. components of blood pressure over the last 100 Direct assessment of aortic stiffness using measure- years. ment of the (PWV) in patients with end-stage renal failure,9 and hypertension10 Arterial stiffness and cardiovascular risk confirms this view. The importance of arterial blood pressure as a determinant of cardiovascular risk has been clearly Pulse wave analysis demonstrated by a number of observational and Arterial stiffness can be measured in a number of intervention studies conducted since the introduc- ways.11 Ultrasound and MRI can be used to assess tion of the mercury at the turn local distensibility and/or compliance, and a variety of the last century. However, the component of of methods exist for assessing PWV, which provides blood pressure that best predicts cardiovascular risk 2 accurate information concerning the distensibility of has been a matter of considerable recent debate. By the arterial segment under investigation.12 However, convention, diastolic pressure was initially con- such methods do not provide information about sidered to be the best predictor of risk, and a high systemic arterial compliance and in particular the systolic pressure was believed to reflect a strong left interaction between the vascular tree and the left ventricle. Despite data in the 1970s and 80s from the 3 ventricle. Pulse wave analysis, offers a non-invasive Framingham Cohort, demonstrating that systolic method for assessment, of systemic arterial stiffness, pressure was probably more important than dias- and also estimation of central arterial pressure.13 tolic values in defining risk, most physicians Intermittent left ventricular ejection creates press- remained heavily biased towards diastolic pressure ure waves that are reflected back, towards the heart for both risk assessment and treatment. However, from sites of impedance mis-match in the periph- overwhelming evidence from epidemiological ery.14 Therefore, the pressure waveform at any point observations now indicates that, at least in the over in the vascular tree is a composite of a forward-going 50s, pulse pressure is a more important determinant and reflected wave. Normally, the reflected wave of cardiovascular risk than mean arterial pressure in returns to the in diastole and boosts coronary normotensive,4 hypertensive5 and post-myocardial 6 artery filling. However, as stiffen, the ampli- infarction populations. Recent re-analysis of sev- tude and speed of the reflected wave both increase, eral blood pressure-lowering studies confirms these 7,8 so that a larger reflected wave returns to the aorta findings. Mean arterial pressure, which is most sooner, which changes the shape of the waveform augmenting central systolic pressure.14 Therefore, the shape of the arterial pressure wave provides use- Correspondence: Dr JR Cockcroft, Department of Cardiology, Wales Heart Research Institute, University Hospital, Heath Park, ful information concerning systemic arterial stiff- Cardiff CF4 4XN, UK. E-mail: cockcroftjrȰcf.ac.uk ness. Mahomed and others recognised this over 100 Received 10 May 2000; accepted 10 May 2000 years ago,15 and since then a variety of devices have Large arterial stiffness JR Cockcroft and IB Wilkinson 534 been employed to record and analyse the pressure large artery stiffness, it is also dependent upon dis- waveform.16,17 However, until recently, PWA was tending pressure, ie, mean arterial pressure.14 There- confined to the peripheral waveform. O’Rourke and fore, some of the beneficial effects observed with colleagues have further developed the technique of valsartan may simply be due to the passive effect of PWA, making non-invasive derivation of central lowering mean arterial pressure. In order to dis- pressure waveforms possible.18 Their technique uses tinguish between passive effects and those due to applanation tonometry to record accurately pressure direct drug actions on arterial structure and function waves from the radial artery,19 and a validated gen- it becomes important to include positive control eralised transfer factor20–24 is then used to generate agents in such studies that will reduce mean arterial the corresponding central arterial waveform. From pressure to a similar degree as the study drug. Pre- this, augmentation index (AIx) – a measure of sys- vious studies have, however, demonstrated differ- temic arterial stiffness, the timing of the reflected ences between antihypertensive agents in their wave – which provides an estimate of the aortic effect on large arterial stiffness and wave reflec- PWV,25 and central pressure can be determined non- tion.32–34 Indeed, ACE inhibitors decrease wave invasively and reproducibly.26 reflection and large artery stiffness,32,35 whereas at Due to differences in vessel compliance and wave least acutely, ␤-adrenoceptor antagonists have been reflection, the shape of the arterial waveform,27,28 shown to increase wave reflection.36 and systolic blood pressure in particular, varies Such observations have important implications throughout the arterial tree.29 Peripheral pulse for the management of patients with isolated sys- pressure is higher than brachial artery pressure in tolic hypertension (ISH), which is primarily a dis- most individuals – ie, there is amplification moving ease of increased arterial stiffness.37 Isolated systolic away from the aorta.29 However, with age, central hypertension currently affects around 50% of the pulse pressure rises and the amplification ratio over 60s, and this number is set to rise substantially falls,14 and similar changes might be expected with with increasing life expectancy over the next 20 other conditions, that induce premature vascular years. Most of the drugs currently used in patients stiffening such as hypertension, mellitus with ISH were developed to treat essential hyperten- and cigarette smoking. Therefore, simply measuring sion, a disease of increased peripheral vascular peripheral pulse pressure will not provide a reliable resistance, and thus were targeted at the resistance estimate of central haemodynamics. This is vessels. Improved treatment of ISH requires new important, because central systolic pressure, not drugs which selectively target the large arteries and brachial artery pressure, defines left ventricular improve arterial compliance both directly, by func- workload and, therefore, left ventricular mass – tional and structural modifications, and indirectly, an important and independent predictor of by reducing mean pressure. To some extent nitrates cardiovascular mortality.30 Moreover, carotid pulse may be useful in this respect, although nitrate toler- pressure correlates more closely with carotid intima- ance and ignorance of the different mechanisms media thickness than brachial pulse pressure.31 underlying and ISH have pre- However, whether measurement of central pulse vented a proper randomised controlled trial pressure will improve risk stratification remains to investigating the effectiveness of sustained-action be proven, although the recent inclusion of PWA, as oral nitrates. a sub study, in large randomised intervention trials The vascular endothelium produces a number of such as the ASCOT, SEARCH and FIELD studies vasoactive mediators such as nitric oxide and endo- will hopefully address this issue. thelin that regulate smooth muscle tone and, are thus likely to contribute to the functional control of Arterial stiffness and therapeutic large artery stiffness. Endothelial dysfunction, which is present in most patients with established intervention cardiovascular risk factors, may explain why con- Mahmud and Feely1 report that addition of the ditions such as diabetes and hypercholesterolaemia angiotensin receptor antagonist valsartan to stan- also associated with increased large artery stiffness, dard drug therapy, in patients with resistant hyper- at an early stage, before the development of manifest tension already receiving an angiotensin-converting .38 Therefore, drugs that improve endo- enzyme (ACE) inhibitor, results in a significant thelial function, such as ACE inhibitors, may also reduction in peripheral blood pressure. They did be useful in these conditions. As yet, little is known not observe any serious adverse events with this about the effects of angiotensin II receptor antagon- combination, although theirs was a small study and ism on endothelial function, although recently the data from larger studies are required to adequately angiotensin II receptor antagonist losartan has been address this important issue. Interestingly, therapy shown to prevent endothelial dysfunction induced with valsartan also resulted in a significant by triglyceride-rich lipoproteins in healthy volun- reduction in AIx, a measure of systemic arterial stiff- teers.39 Mahmud and Feely suggest that valsartan ness, and an increase in the amplification ratio of may exert its beneficial effect on arterial stiffness peripheral to central pulse pressure, again indicat- and wave reflection by improving endothelial func- ing reduced large artery stiffness. tion. This view is supported by our own recent data Although the observed changes in stiffness were demonstrating an inverse correlation between endo- highly significant, they must be interpreted with thelial dysfunction and AIx in a group of growth caution, because while both structural and func- hormone deficient subjects.40 However, further work tional components within the vessel wall define investigating the link between endothelial function

Journal of Human Hypertension Large arterial stiffness JR Cockcroft and IB Wilkinson 535 and arterial stiffness is required in this emerging 20 Karamanoglu M, O’Rourke MF, Avolio AP, Kelly RP. field. An analysis of the relationship between central aortic In summary, arterial stiffness is a key determinant and peripheral upper limb pressure waves in man. Eur of cardiovascular risk in older individuals, and Heart J 1993; 14: 160–167. drugs that target large artery stiffness will be increas- 21 Takazawa K, O’Rourke M, Fujita M. Estimation of ascending aortic pressure from radial arterial pressure ingly important in the treatment of ISH. A number using a generalized transfer function. Z Kardiol 1996; of methods exist for assessing stiffness. PWA is one 85: 137–139. such technique that has a number of advantages: it 22 Chen C-H et al. Estimation of central aortic pressure is simple, non-invasive and reproducible, and also waveform by mathematical transformation of radial provides an estimation of central blood pressure. tonometry pressure. Circulation 1997; 95: 1827–1836. 23 Soderstrom S et al. Substantial equivalence between References ascending aortic pressure waveforms and waveforms derived from the radial pulse using a generalized 1 Mahmud A, Feely J. Favourable effects on arterial transfer function? FASEB J 1998; A712: 4131. wave reflection and pulse pressure amplification of 24 Fetics B, Nevo E, Chen C-H, KAss DA. Parametric adding angiotensin II receptor blockade in resistant model derivation of transfer function for non-invasive hypertension. J Hum Hypertens 2000; 14: 541–546. estimation of aortic pressure by radial tonometry. IEEE 2 O’Rourke MF, Frohlich ED. Pulse pressure: is this a Trans Biomed Eng 1999; 46: 698–706. clinically useful risk factor? Hypertension 1999; 34: 25 Murgo JP, Westerhof N, Giolma JP, Altobelli SA. 372–374. Aortic input impedance in normal man: relationship 3 Kannel WB et al. Systolic blood pressure, arterial rigid- to pressure waveforms. Circulation 1980; 62: 105–116. ity, and risk of . JAMA 1981; 245: 1225–1229. 26 Wilkinson IB et al. The reproducibility of pulse wave 4 Benetos A et al. Pulse pressure: a predictor of long- velocity and augmentation index measured by pulse term cardiovascular mortality in a French male popu- wave analysis. J Hypertens 1998; 16: 2079–2084. lation. Hypertension 1997; 30: 1410–1415. 27 Remington JW, Wood EH. Formation of the peripheral 5 Benetos A, Rudnichi A, Safar M, Guize L. Pulse press- ure and cardiovascular mortality in normotensive and pulse contour in man. J Appl Physiol 1956; 9: 433– hypertensive subjects. Hypertension 1998; 32: 560– 442. 564. 28 Latham RD et al. Regional wave travel and reflections 6 Mitchell GF et al. Sphygmomanometrically determ- along the human aorta: a study with six simultaneous ined pulse pressure is a powerful independent predic- micromanometric pressures. Circulation 1985; 72: tor of recurrent events after in 1257–1269. patients with impaired left ventricular function. Circu- 29 Pauca AL, Wallenhaupt ST, Kon ND, Tucker WY. Does lation 1997; 96: 4254–4260. radial artery pressure accurately reflect aortic press- 7 Miller JA, Lever AF, Burke V. Pulse pressure as a risk ure? Chest 1992; 102: 1193–1198. factor for cardiovascular events in the MRC Mild 30 Levy et al. Prognostic implications of echocardio- Hypertension Trial. J Hypertens 1999; 17: 1065–1072. graphically determined left ventricular mass in the 8 Blacher J et al. Pulse pressure not mean pressure deter- Framingham heart study. N Engl J Med 1990; 332: mines cardiovascular risk in older hypertensive 1561–1566. patients. Arch Intern Med 2000; 160: 1085–1089. 31 Boutouyrie P et al. Association between local pulse 9 Blacher J et al. Impact of aortic stiffness on survival in pressure, mean blood pressure, and large-artery end-stage renal disease. Circulation 1999; 99: 2434– remodeling. Circulation 1999; 100: 1387–1393. 2439. 32 Chen C-H et al. Different effects of fosinipril and aten- 10 Blacher J et al. Aortic pulse wave velocity as a marker olol on wave reflection in hypertensive patients. of cardiovascular risk in hypertensive patients. Hyper- Hypertension 1995; 25: 1034–1041. tension 1999; 33: 1111–1117. 33 van Bortel LM, Kool MJ, Boudier AS. Effects of antihy- 11 Cockcroft JR, Wilkinson IB. Vessel wall properties and pertensive agents on local arterial distensibility and cardiovascular risk. J Hum Hypertens 1998; 12: 343– compliance. Hypertension 1995; 26: 531–534. 344. 34 Breithaupt-Grogler K et al. Influence of antihyperten- 12 Bramwell JC, Hill AV. Velocity of transmission of the sive therapy with Cilazapril and hydrochlorothiazide pulse-wave and elasticity of the arteries. Lancet 1922; on the stiffness of the aorta. Cardiovasc Drugs Therapy i: 891–892. 1996; 10: 19–37. 13 Wilkinson IB, Cockcroft JR, Webb DJ. Pulse wave 35 Ting C-T et al. Arterial hemodynamics in human analysis and arterial stiffness. J Cardiovasc Pharmacol hypertension: Effects of angiotensin converting 1999; 32: 33–37. enzyme inhibition. Hypertension 1993; 22: 839–846. 14 Nichols WW, O’Rourke MF. McDonald’s Blood Flow 36 Ting C-T et al. Arterial haemodynamics in human in Arteries: Theoretical, Experimental and Clinical hypertension effects of adrenergic blockade. Circu- Principles. Arnold: London, 1998. lation 1991; 84: 1049–1057. 15 Mahomed FA. The physiological and clinical use of 37 Staessen J, Amery A, Fagard R. Isolated systolic hyper- the sphygmograph. Medical Times Gazette 1872; 1: tension in the elderly. J Hypertens 1990; 8: 393–405. 62–64. 38 Cockcroft JR, Wilkinson IB, Webb DJ. The Trevor How- 16 Lax H, Feinberg AW, Cohen BM. Studies of the arterial ell Lecture: age, arterial stiffness and the endothelium. pulse wave. J Chron Dis 1956; 3: 618–631. Age Ageing 1997; 26: 53–60. 17 Freis ED, Heath WC, Lushsimger PC, Snell RE. 39 Wilmink HW et al. Effect of angiotensin-converting Changes in the carotid pulse which occur with age and enzyme inhibition and angiotensin II type 1 receptor hypertension. Am Heart J 1966; 71: 757–765. antagonism on post-prandial endothelial function. J 18 O’Rourke MF, Gallagher DE. Pulse wave analysis. J Am Coll Cardiol 1999; 34: 140–145. Hypertens 1996; 14: 147–157. 40 Smith JC et al. Increased arterial stiffening occurs in 19 Kelly RP et al. Non-invasive registration of the arterial growth hormone deficient (GHD) adults and is related pulse waveform using high fidelity applanation to endothelial dysfunction J Endocrinol 2000; 164: tonometry. J Vasc Med Biol 1989; 1: 142–149. S146.

Journal of Human Hypertension