The two faces of hypertension: role of aortic stiffness

Highlights

• Midlife hypertension is relatively distinct from systolic hypertension of the aged.

• Systolic hypertension is largely a result of aortic stiffening—arteriosclerosis.

• Cardiovascular complications of arteriosclerosis differ from those of atherosclerosis.

• Pulse wave velocity, a measure of aortic stiffness, is an independent cardiovascular risk factor.

• Aortic stiffening of aging may be amenable to new pharmacologic therapy.

Abstract

Adult hypertension can be divided into two relatively distinct forms—systolic/diastolic hypertension in midlife and systolic hypertension of the aged. The two types differ in prevalence, pathophysiology, and therapy. The prevalence of systolic hypertension in the elderly is twice that of midlife hypertension. The systolic pressure is elevated in both forms, but the high diastolic pressure in midlife is due to a raised total peripheral resistance, whereas the normal or low diastolic pressure in the elderly is due to aortic stiffening. Aortic stiffness, as measured by the carotid/femoral pulse wave velocity, has been found to be a cardiovascular risk marker independent of traditional risk factors for atherosclerosis. Instead, it is related to microcirculatory disease of the brain and kidney and to disorders of inflammation. Loss of aortic distensibility is an inevitable consequence of aging, but a review of its causes suggests that it may be amenable to future pharmacologic therapy.

Introduction

From a hemodynamic standpoint, essential hypertension has been conventionally divided into three groups—diastolic hypertension, systolic/diastolic hypertension, and systolic hypertension. It is the purpose of this article to offer an alternative separation into two main groups—the systolic/diastolic hypertension of midlife (40 to 59 years) and the systolic hypertension of the aged (>60 years). Although this new classification appears little changed from the original, it is alternatively based on an understanding of the age group differences in prevalence, pathophysiology, cardiovascular (CV) complications, and therapy rather than on the blood pressures (BPs) alone. Clearly, the two new groups overlap as the midlife population ages, but the multiple differences that develop with aging are sufficient to consider these two forms of hypertension as relatively distinct—a distinction that allows for better understanding of the entire disease across the entire population. Underscoring the differences between these two forms of hypertension, in addition, requires a review of the role played by increased aortic and arterial stiffness and the growing appreciation of aortic stiffness as an independent CV risk factor.

Hypertension becomes more and more prevalent as the American population ages, and the BP values themselves vary at different ages. In 1995, the National Health and Nutrition Examination Survey (NHANES) published a cross-sectional analysis of 9900 individuals representative of the US population that related their BP to their age.¹ This article showed the now well-known steady rise in systolic blood pressure (SBP) with aging. Perhaps, less well known was that after age 55, the diastolic blood pressures (DBPs) progressively fell. The rise in SBP and fall in DBP after midlife were confirmed in a now classic article from the Framingham study published in 1997.² This longitudinal study looked at the effect of aging in more than 2000 subjects with different entry BPs and showed in all BP groups after age 55, a steady rise in SBP and a fall in DBP. The rise in SBP and fall in DBP resulted in a marked increase in pulse pressure (PP) as that population aged. Although the SBP and PP rose and the DBP fell with advancing age, the mean blood pressure (MBP) rose slightly in midlife and then stabilized.

The MBP is often an enigma because it is seldom measured clinically having found limited use as a risk predictor, but it is important in understanding the differences between the two types of hypertension. The MBP is the time-weighted average of the arterial pressure pulse or the pressure that divides the area of the pressure pulse into equal upper and lower halves. Physiologically, it is the pressure that drives the blood steadily through the peripheral resistance where there is no arterial pulsation. Calculation of the brachial MBP from the SBP and DBP requires a bit of arithmetic (PP/3 + DBP) but is now displayed in the digital output of some automatic oscillometric BP devices. The importance of the MBP will become apparent below under pathophysiology.

In the most recent NHANES, the overall prevalence of hypertension in the United States was 29.1% with an equal distribution between men and women³ (Figure 1). The prevalence at midlife (ages 40–59 years) for both sexes was 32.4%, whereas that over the age of 60 years was doubled at 65%, a striking difference. Previous NHANES studies have shown the prevalence of hypertension to be 78% more than the age of 75 years⁴ and in the 1995 study¹ as high as 80% in black Americans. It is concerning to realize that nearly 4 of 5 of individuals of advanced age are hypertensive and that the already large number of Americans with hypertension will increase as the number of our elderly progressively rises.

Hypertension in the elderly is of a different types than that of midlife. Between ages 50 and 59 years, there is an elevated DBP in approximately half of the cases. By contrast, more than the age of 70, only 10% of the patients have a DBP elevation and the vast majority (90%) has systolic hypertension only.⁵ In a related aside, 50% of treatment failures in midlife were due to inadequately controlled DBP elevations, but in the aged, 90% of failures were due to uncontrolled SBP elevations.5, 6 By comparison, diastolic midlife hypertension is relatively easy to treat, whereas the systolic elevation of the aged is difficult.

Figure 2 depicts the differences among the arterial pulses in normotensive subjects (pulse A), patients with midlife hypertension (pulse B), and those with systolic hypertension (pulse C). In midlife hypertension, the SBP and DBP are both elevated, both drawn up by the high MBP. The MBP is somewhat higher than normal in systolic hypertension but far lower than that of the MBP in midlife hypertension. In systolic hypertension, the SBP is similarly elevated, the result of a normal stroke volume pumped into a stiff aorta. But the aortic systolic pressure may be influenced by other factors. With a stiff aorta, the SBP could be normal if LV function and/or stroke volume were reduced. Conversely, with normal aortic distensibility, the aortic SBP could be elevated in the presence of obesity, increased LV contractility, and/or stroke volume. However, in addition to raising the SBP, increased aortic stiffness lowers the DBP on which coronary flow is dependent.

The MBP is also important, and its determinants can be seen in the standard formula:

$$\text{MBP}=\text{Cardiac}\text{output}\times \text{total}\text{peripheral}\text{resistance}\text{.}$$

The cardiac output in essential hypertension has been often measured and found to be normal. Thus, the elevated MBP, about which the SBP and DBP oscillate, is high because of a raised total peripheral resistance (TPR), the circulatory abnormality believed most responsible for this form of hypertension. The primarily elevated MBP therefore raises both the SBP and DBP and at the same time stiffens the normal aorta. Stiffening of the aorta as the internal pressure rises represents a fundamental property of the viscoelastic materials that comprise the aortic wall. A plot of the stress/strain relationships for viscoelastic materials is displayed in Figure 3. Increasing stress on the vertical axis is associated with progressively decreasing deformation (strain), shown on the horizontal axis, producing the familiar curvilinear relationship. Stiffness is defined as the ratio of stress to strain or the slope of the line at any given stress. The shape of the plot shows that stiffness rises exponentially with increasing stress. Applied to the aorta, this principle explains progressive aortic stiffening at increasing intraluminal pressures. Therefore, the increased aortic stiffness in midlife hypertension is due to the hypertension itself without a necessary alteration of the aortic wall. This belief is supported by central arterial wall stiffness, as measured by Young's Modulus in the carotid artery, that is similar in subjects with and without hypertension when measured at common intraluminal stresses.⁷ A recent statement from the American Heart Association confirms “that aortic stiffness is a cause rather than a consequence of hypertension in middle-aged and older individuals.”⁸ The mildly elevated MBP in systolic hypertension (Figure 2C) when compared with the high MBP of midlife hypertension is due to the large rise in SBP and the smaller fall in DBP, both related to changes in the aortic wall composition but unrelated to the TPR.

To summarize, midlife hypertension induces a stiff aorta, whereas the independently stiffened aorta of aging, absent other factors that can raise the SBP, causes systolic hypertension. The pathophysiologic differences in the two forms of hypertension explain the high SBP, low DBP, and high prevalence of systolic hypertension in the aged and the elevated DBP and MBP in midlife hypertension. The differences also explain the relative success of current drugs (angiotensin-converting enzyme inhibitors [ACEI], angiotensin receptor blockers [ARB], and calcium channel blockers [CCB]) that lower the TPR in midlife hypertension and have lesser or controversial direct effects on the aortic wall (see below–Implications for Therapy).

The mechanisms that induce stiffened central aortic walls and systolic hypertension in the aged are complex. Elastin that governs the distensibility of the aortic wall in youth becomes progressively fragmented and degraded over time.9, 10 The fragmentation is related to the incessant pounding of the PP—its magnitude and its frequency (heart rate).¹¹ The behavior of the wall then becomes gradually dependent on the residual collagen content that is 100 to 1000 times stiffer than elastin.¹² With aging, aortic wall stiffness is further increased by an increase in collagen content, and the cross-linking of collagen and elastin fibers by advanced glycation end products.12, 13 The fragmented elastin also serves as a nidus for the microdeposition of calcium, an additional factor leading to wall stiffening.14, 15 Inflammation with aging has also been shown to be related to aortic stiffness by the association of inflammatory markers in healthy and hypertensive individuals with pulse wave velocity (PWV), a measure of aortic stiffness (see below–Aortic Distensibility and PWV).10, 16 The entire issue of aortic stiffness is complicated by other influences such as urban living¹⁷ and genetic backgrounds.10, 18, 19 The many and varied studies of the mechanisms responsible for increasing aortic stiffness with aging have led to discussions of whether the changes in the aortic wall, as measured by PWV, are pathologic8, 17, 19 or inevitable.²⁰ It would appear that they are both.

Decreased aortic distensibility, in addition to its role in hypertension, has become recognized as an important variable in the overall assessment of circulatory function. This variable can be easily assessed by measuring the carotid-femoral PWV, a method that has been widely applied in population studies and is now approaching clinical use. The underlying principle of the technique is that the velocity of pulse travel is determined by the stiffness of the pipe through which it passes—the stiffer the tube, the faster the pulse travels. Both carotid and femoral pulses can be readily applanated, recorded noninvasively, and the difference in their arrival times measured. The distance between the pulse sensors is measured over the body surface, and this distance, divided by the interval, represents the PWV. Some authors introduce a correction by subtracting the distance between the suprasternal notch and the carotid recording site or using a standard correction factor of 0.8. The method is noninvasive and convenient, and no BP measurements are needed, but there are limitations. In some patients, the pulses are difficult to record (ie, obesity). The distance measurement over the body surface as an estimate of aortic length is an approximation at best because aortic tortuosity and abdominal obesity cannot be accounted for. Nonetheless, the method is now recognized as the “gold standard” for the clinical assessment of aortic stiffness.21, 22

As an indicator of aortic stiffness, PWV has found many applications. Similar to SBP and PP, it rises inexorably with aging, but apart from BP itself, it is surprisingly little influenced by traditional risk factors such as hypercholesterolemia, diet, or smoking.23, 24 And yet alone, it is an excellent risk predictor for CV mortality, strokes, and myocardial infarctions.25, 26 It differs from standard circulating biomarkers that are variable and offer only risk “snapshots” at the moment of measurement.26, 27 PWV is a more stable parameter that becomes abnormal gradually over time and represents an overall, time-weighted, cumulative, composite picture of risk rather than a risk estimate made at a given moment.26, 27 It is therefore a measure of “vascular age” contrasted with chronological age.²⁸ Therapeutic reduction of traditional risk factors ameliorates the risks of atherosclerosis, but reduction of traditional risk factors offers little to change the circulatory risks of arteriosclerosis and its attendant aortic stiffness. PWV alone adds significantly to the traditional risk estimates of CV disease and a carotid-femoral PWV of >10 m/s is already included in the Practice Guidelines for the Management of Arterial Hypertension in the European Society of Hypertension and the European Society of Cardiology.²⁹ These Guidelines concluded, “The additive value of PWV above and beyond traditional risk factors, including SCORE and Framingham risk score has been quantified in a number of studies. In addition, a substantial proportion of patients at intermediate risk could be reclassified into a higher or lower CV risk when arterial stiffness is measured.” PWV has not yet appeared in the American guidelines, but a Scientific Statement from the American Heart Association⁸ has given PWV a “Class I; level of Evidence A” recommendation as a method for the measurement of arterial stiffness. Furthermore, it states that “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).”

The reduced cushioning function of a stiffened aorta causes the primary arterial pulse to travel more quickly to the periphery. There, some of the forward pulsatile energy is returned as a reflected wave back into the central aorta. The early aortic SBP, already elevated by aortic stiffness, is made even higher by the later superimposition of this reflected wave. Importantly, both raised early and late systolic waves have been related to the development of heart failure.30, 31, 32 Arterial stiffening has also been associated with impaired LV diastolic function,⁸ another determinant of heart failure. The increases in SBP add to the left ventricular stroke work of ejection and its need for additional coronary blood flow. But the DBP on which the coronary flow depends is lower also due to the increased aortic stiffness. Taken together, reduced aortic distensibility increases the need for increased cardiac stroke work but may reduce the heart's ability to deliver it.

There are two high-flow, low-resistance organs in the body: the brain and kidneys. These low vascular resistance organs permit the increased pulsatility induced by a stiffened aorta to penetrate further into their microcirculation. This pulse penetration does microcirculatory damage to the brain and kidney that is less true of higher resistance organs.³³ Figure 4 illustrates the effect of proximal aortic stiffness in the elderly and the resultant arterial impedance mismatch on the microcirculation. Silent small vessel disease of the brain is associated with lacunar infarcts, microbleeds, and white matter hyperintensities that are independently related to PWV in hypertensive elderly patients.34, 35 Increased PWV has also been associated with loss of cognitive function³⁶ and called pulse wave encephalopathy.³⁷ Separate studies in patients with chronic renal disease have shown an independent relationship of PWV with decline in renal function and all-cause mortality.38, 39, 40 Reduced glomerular filtration and increased albumin excretion have also been related to greater arterial stiffness.⁴¹ Furthermore, diastolic aortic flow reversal, induced by proximal aortic stiffness and aortic impedance mismatch, may be responsible for the reduction in renal function.⁴² Thus, stiffening of the aorta by arteriosclerosis leads to microcirculatory complications of the brain and kidney that differ from the occlusive arterial disease of atherosclerosis.

Increased arterial and/or aortic stiffness as manifested by increased PWV has appeared as a factor in several other apparently unrelated illnesses. The commonest causes of death in obstructive pulmonary disease are CV in nature. Even when corrected for other pertinent factors, PWV was elevated in obstructive pulmonary disease and related to its severity.⁴³ The raised PWV in this disease was associated with circulating inflammatory mediators, and the PWV was reduced by antioxidant therapy.⁴⁴ PWV is also associated with obesity and varies with changes in weight.45, 46 Another association with inflammation has been well demonstrated by the relationship of PWV to levels of C-reactive protein in both hypertensive subjects and in healthy men, ages 45 to 59 years.11, 16 PWV is also elevated in rheumatoid arthritis,⁴⁷ and the PWV level is reduced by antitumor necrosis factor alpha and by the presumed anti-inflammatory effects of ezetimibe and simvastatin.48, 49 Nuclear imaging has directly demonstrated the presence of inflammation in the stiffened aortic wall.⁵⁰ These CV risks of arteriosclerosis are separate from the atherosclerotic risks associated with standard risk factors.

The JNC 8 guidelines for the treatment of hypertension also recognized the differences between midlife hypertension and that of the aged by assigning higher SBP values as therapeutic targets for the latter (140 vs. 150 mm Hg) (Table 1).⁵¹ The evidence for these guidelines was drawn from randomized controlled trials available at that time (2014). These values, controversial when published, may now be supplanted by the recently released results of the SPRINT trial.⁵² This landmark study has shown a lower all-cause mortality and fatal and major nonfatal CV events by using a therapeutic SBP target of 120 mm Hg when compared with a target of 140 mm Hg. These results are the same for patients above or below the age of 75 and are also included in Table 1. The SPRINT conclusions may be the basis for revised guidelines in future.

The BP in midlife hypertension can effectively be lowered by now commonly used agents (thiazide diuretics, ACEIs, ARBs, and CCBs). Their primary mode of action is to reduce the TPR and, by doing so, lower the primarily raised MBP and thus the SBP and DBP. These drugs, especially CCBs, are widely and effectively used in systolic hypertension but have a separate, limited, and even controversial effect on aortic stiffness that is independent of the MBP.8, 53, 54, 55, 56 Their limited ability to reduce the aortic stiffness of aging accounts for the therapeutic difficulty encountered clinically in lowering the SBP and/or raising the DBP in the elderly. Nonetheless, there is unequivocal evidence that treatment of isolated systolic hypertension with standard antihypertensive therapy is beneficial.57, 58, 59

Yet, there still remains an unmet need for other agents that more directly attack aortic stiffness in the management of hypertension in the large and growing numbers of affected elderly patients. Aldosterone inhibition has been reported to have an effect on stiffness60, 61 but has not had a major clinical impact. Breakers of collagen cross-links, such as alagebrium, more directly attack aortic stiffness,⁶² but they too have had limited clinical success in hypertension therapy. Efforts to delay or reduce aortic elastin fragmentation¹⁰ using matrix metalloproteinase inhibitors⁶³ offer some hope in this direction but have not been clinically tested. Because of a suspected role for inflammation in aortic stiffness,⁶⁴ anti-inflammatory agents have been tried only in limited circumstances.48, 49 A promising recent therapeutic candidate has been a vasopeptidase inhibitor that inhibits both ACE and neprilysin—a neutral endopeptidase that inactivates natriuretic peptides such as atrial and brain peptides. This inhibition of an inactivator permits greater activity of atrial natriuretic peptide and brain natriuretic peptide, both of which have natriuretic and antihypertensive actions. Omapatrilat was the first drug of this type to be studied for its clinical value on aortic stiffness. It was more effective than the ARB valsartan in lowering the PP in patients with systolic hypertension⁶⁵ but was later withdrawn from further study because of the unwanted side effects of bradykininemia. A newer drug of the same class (LCZ 696—valsartan/sacubitril) has also been found more effective than valsartan in lowering the BP in patients with essential hypertension⁶⁶ but with fewer side effects. A clinical study of this drug is nearly complete comparing it with olmesartan and measuring its effects on arterial stiffness in the elderly (PARAMETER trial).⁶⁷ The results of all these and future trials should be closely watched as newer efforts unfold to address the currently unmet needs of the elderly with systolic hypertension.