- Email: [email protected]
Effects of age on ventricular-vascular coupling
Effects of Age on Ventricular-Vascular Coupling WILMER W. NICHOLS, PhD, MICHAELF. O’ROURKE,MD, ALBERT P. AVOLIO, PhD, TOSHIO YAGINUMA, MD, JOSEPHP. MURGO,MD, CARL J. PEPJNE,MD, and C. RICHARDCONTI, MD
The effects of age on the interrelation between the physical properties of the arterial tree (aortic input impedance) and left ventricular performance (cardiac output) were studied in 45 subjects, aged 19 to 62 years, without apparent cardiovascular disease. Ascending aortic pulsatile pressure and blood flow velocity were measured with a multisensor catheter and cardiac output by green dye or the Fick method. Heart rate and end-diastolic aortic pressure remained unchanged with age, whereas aortic systolic, mean and pulse pressures and aortic radius increased. In subjects younger than 30 years, early systolic pressure usually exceeded late systolic pressure (type C beat); in subjects older than 50 years, late systolic pressure usually exceeded early systolic pressure (type A beat). In 55 % of subjects aged 30 to 50 years, early and late systolic pressures were essentially equal (type B beat). The impedance spectra from all subjects showed fluc-
tuations about the characteristic impedance (index of elastance) that were greater in the older subjects. Peripheral resistance increased 37 % (r = 0.47, p
The effects of aging on cardiovascular function have concernedphysiciansand physiologistssince1809,when Thomas Young1 introduced his modulus of elasticity to study arterial function. However, remarkably little is known about the detailed effects of aging on cardiovascular hemodynamics. Vascular load increases through increases in arterial pressure,2 peripheral re-
tra.3~8~g~11~12 Aortic input impedance is a function of systemic vascular resistance, central aortic elastance (inverseof compliance)or stiffnessand wavereflections, and it selectively describesthe arterial system.13~14 The present study provides detailed information on the effects of ageon LV-vascular coupling in human subjects without apparent cardiovascular disease.
sistance3 and arterial stiffness.3-5Aging is associated with mild cardiac hypertrophy6 and prolonged or de-
Methods
layed left ventricular
(LV) relaxation.’
Increases in
The data reported here were obtained from studiesperformed at The JohnsHopkinsUniversity(n = lo), The University of Florida (n = 13), Brooke Army Medical Center (n = 18), and The University of New South Wales(n = 4) and have beenpublished in part elsewhere.15J6 The observations were made on 45 adults, aged 19 to 62 years, who were
vascular load are associatedwith decreasedstroke voland decreased capacity for increasing stroke ume3y8vg volume with exercise.lO LV external vascular load can
be characterized by the aortic input impedance spec-
undergoingcardiaccatheterizationfor variousclinicalindications,the mostcommonof whichwasa chestpainsyndrome.
From the Department of Medicine, University of Florida, and Veterans Administration Medical Center, Gainesville, Florida; Department of Medicine, St. Vincent’s Hospital, Sydney, Australia: Department of Medicine, Cardiology Service, Brooke Army Medical Center, Fort Sam Houston, Texas. Manuscript received July 20, 1984; revised manuscript received December 17, 1984, accepted December 3 1, 1984. Address for reprints: Wilmer W. Nichols, PhD, Department of Meditine, Box J-277, University of Florida, Gainesville, Florida 32610.
No cardiovasculardiseasewasfound by hemodynamicmeasurements, LV cineangiographyor coronary arteriography. During basalstate studies all patients wereeither unsedated or very lightly sedatedwith oral diazepam(10 mg 1 hour before the procedure)or sodium pentobarbital(100 mg orally). Multisensor catheters,custom-designedfor the left side of the 1179
1180
AGE AND AORTIC
TABLE I
IMPEDANCE
Effects of Age on Arterial Hemodynamics Variable
Slope
Intercept
Correlation Coefficient
Aortic radius (cm) Heart rate (beatsimin) Systolic pressure (mm Hg) Diastolic pressure (mm Hg) Mean pressure (mm Hg) Pulse pressure (mm Hg) Cardiac output (litersimin) Stroke volume (ml) Peak flow (ml/s) Characteristic impedance (dynes s cm+) Peripheral resistance (dynes s cmm5) Frequency of 1Z 1,,,I” (Hz)
0.0097 0.0486 0.6852 0.0702 0.1911 0.6150 -0.0365 -0.5605 -3.2638 1.2135 9.7409 0.0364
0.98 73.64 95.00 76.12 90.27 18.89 7.85 109.30 777.92 10.3 853.97 2.41
0.61 0.05 0.63 0.11 0.33 0.70 -0.37 -0.30 -0.31 0.66 0.47 0.65
T 5.08 0.30 5.36 0.73 2.26 6.38 -2.63 -2.076 -2.17 5.71 3.45 5.59
p Value 0.001 0% ‘NS 0.01 0.001 0.005 0.025 0.025 0.001 0.001 0.001
IZI,,,i, = first minimum of impedance moduli; NS = not significant.
heart, were used in 40 subjects. The catheters each contained 2 laterally mounted solid-state pressure sensors (Millar Instruments) and an electromagnetic blood flow velocity sensor (Carolina Medical Electronics or Millar Instruments). One pressure sensor was mounted at the same site as the velocity sensor, and the other was mounted at the catheter tip 5 to 7 cm away and was used for recording LV pressure.A fluid-filled : system was used to record aortic and LV pressures in the remaining 5 subjects. The frequency responseof this system was’ constant in amplitude (f5%) from 0 to 20 Hz with a damping ratio of 0.15. This response is adequate for accurate measurement of pulsatile pressure in the human aorta, because more than 98% of the variance of the pulsations is included in the first 7 harmonies.14Js The velocity-pressure catheter was advanced through a brachial arteriotomy to the aortis valve so that the tip was within the LV cavity. This arrangement placed the velocity probe and associatedpressure sensor 3 to 5 cm above the aortic valve at approximately the upper border of the sinuses of Valsalva. This arrangement tends to stabilize the transducer in the central axis of the ascending aorta where the velocity profile is relatively flat.17 The velocity sensors were used in conjunction with either a square wave (Carolina Medical Electronics) or a sine wave (Biotronex or SE Laboratories) blood flowmeter. Details of the technical characteristics and clinical use of these flowmeters and multisensor catheters, including frequency response, drift characteristics and calibration techniques, have been described previously.lsJg LV cineangiography and coronary arteriography were performed after hemodynamic data were collected. If the aortic root was not adequately visualized during the ventriculographic study, a selective aortic root angiogram was performed to measure the mean systolic radius of the ascending aorta at the site where the pressure and velocity were recorded. The mean output signal (velocity) of the flowmeter was calibrated in volumetric flow units (milliliters per second) by. reference to a simultaneous determination of cardiac output by the dye dilution or Fick method. The pressure and flow velocity signals were recorded on analog magnetic tape and later digitized at a sampling interval of 5 or 10 ms. The technical details of the analog and digital processing have been described elsewhere.ls*is Data were analyzed on a digital computer, which converted pressure and flow velocity signals to Fourier series, applied corrections for the measured dynamic responses of the transducer and computed aortic impedance (modulus and phase) as functions of frequency. Input impedance modulus at each harmonic frequency was computed by dividing pressure modulus by flow modulus, and the impedance phase was computed by sub-
tracting the phase angle of flow from that of pressure.14Details regarding signal averaging and techniques for removing spurious data caused by noise have been published.l5Js The impedance at 0 Hz or “input resistance” was calculated by dividing mean aortic pressure by mean flow. Charactaristic aortic impedance was estimated by averaging impedance moduli above 2 Hz. Characteristic impedance depends on the viscoelastic properties of the artery under study, whereas input impedance moduli oscillate around the characteristic value because of waves reflected from more distal
points.*3J4 The systolic configuration of the ascending aortic pressure wave form was classified according to the criteria of Murgo et all6 Type A: Peak systolic pressure (Ppk) occurs in late systole after a well-defined anacrotic notch or inflection point (Pi)
and (Ppk-Pi)/(pulse pressure)is greater than 0.12. Type B: Peak systolic pressurealso occurs in late systole following inflection point, but (Ppk-Pi)/(pulse pressure) is
between0.0 and 0.12. Type C: Peak systolic pressure precedes a well-defined
inflection point, and (Ppk-PiMpulse pressure) is less than 0.0. The relation between age and the different variables was obtained by means of least-squares linear regression analysis.
Results The subjects’ ages were fairly well distributed over 4 decades. However, there was a larger concentration of patients (18) in the fourth decade. Of all measurements collected, only heart rate (range 56 to 110 beats/min, average 76 f 3, difference not significant) and aortic diastolic pressure (range 64 to 94 mm Hg, average 76 f 2, difference not significant) were not age-related. This finding agreeswith observations made by others.10*20 The relation between age and other cardiovascular variables are given in Table I. Aortic radius: The mean systolic internal radius of the ascending aorta measured at approximately the same site as the pressure and blood flow velocity (3 to 5 cm above the aortic valve) ranged from 1.0 to 1.8 cm and showed a positive correlation with age (r = 0.61, p CO.001) (Fig. 1). This finding represents an increase of 9% in radius or 20% increase in cross-sectional area per decade. These results are similar to those reported previously.20*21
April 15. 1985
Aortic pressure waveform: Pressure wave forms recorded in subjects younger than 30 years were generally type C beats (75%), whereas those recorded in subjects older than 50 years were type A beats (90%). Subjects aged 30 to 50 years showed 11% type C, 55% type B and 33% type A beats. Murgo et al16 also noted a similar relation between pressure wave form and age. Examples of pressure waves recorded in a young subject and an old subject are shown in Figure 2. Over the age range of 20 to 60 years, aortic systolic pressure increased 25% (from 109 to 136 mm Hg), or 6% per decade (r = 0.63, p
AORTIC
RADIUS
. . . . -L-- __-. . . .-.-C.-L . . . . t---.. . . . . e,-i;-.. . .. ---:. . _-- . . .. * * .
r = 0.61 P < 0001
L 0
-.T--.I 10
20
30 AGE
._.-(..-.^.---. 40
50
60
JOURNAL
OF CARDIOLOGY
Volume 55
1161
94 to 102 mm Hg), or 2% per decade (r = 0.33, p
70
(YEARS1
1. Change in mean systolic internal ascending aortic radius with age. Radius was measured from contrast angiograms at approximately the same site as the pressure and velocity (3 to 5 cm above the aortic valve).
FIGURE
ASCENDING
THE AMERICAN
200
AORTIC
PRESSURE I = 0.63
AORTA rz011 P = NS (]w
----T--
30 AGE
I
40
50
60
70
circles)
and
(YEARS)
FIGURE 3. Change in ascending aortic systolic (closed diastolic (open circles) pressure with age.
154 CARDIAC
OUTPUT
r = -037 P < 0.005
L--m. 0 and pressure waveforms recorded in a young (type C pressure wave) and an old (type A pressure wave) subject. FIGURE
2. Ascending aortic blood flow velocity
I.. 10
20
30 AGE
FIGURE
, .1 40 (YEARS)
4. Change in cardiac output with age.
-..50
-I 60
70
1182
AGE AND AORTIC
IMPEDANCE
STROKE
The changes in arterial pressure are a little different from those of epidemiologic studies, in which pressure was recorded indirectly in the brachial artery with the subject sitting (rather than recumbent) and unsedated and uninstrumented.25 We find here a greater change in systolic pressure with age than was found in those studies and a smaller change in diastolic pressure. Although the aforementioned factors may be responsible and our data are minute compared with the mass of epidemiologic data on brachial artery pressure, we stress that recordings of brachial artery pressure may be quite different from recordings of central aortic pressure26 and that the epidemiologic studies take no account of changing amplification between ascending aorta and brachial artery with age. The changes reported here are similar to those predicted when change in brachial -amplification was taken into account (Fig. 10).1s*2s The studies reported here show a definite increase in both pulsatile (characteristic impedance and reflections) and nonpulsatile (peripheral resistance) components of LV vascular load. The increase in characteristic impedance was surprisingly high-137% between ages 20 and 60 years. This increase would have been even higher if allowance had been made for change in aortic diameter and this had been expressed in terms Merillon et a120calculated the of linear flow velocity. 13127 increase in characteristic impedance as approximately 80% between ages20 and 60 years, whereas other studies of aortic pulse wave velocity suggest change in characteristic impedance of approximately 60% over this age range.28 The cause of change in vascular load can be inferred from the change in peripheral resistance and in characteristic impedance. The former is determined by opposition to steady flow in the small peripheral vessels, whereas the latter is determined by arterial stiffness. A small change occurs in resistance and can be attributed to decreased caliber or decreased total cross-sectional area of small vessels, whereas the greater change is in arterial stiffness and is attributable to asymptomatic arterial degeneration and a possible increase in the collagen-elastin ratio.29
VOLUME
150 2 .5
Y 3
.
1
-----_____
loo
.
*. .. * .** . ;‘--.----;$. --r---z . . ---;.** -* . . *.
.
.
G
.
I
50
-
.
l ’
-____
I
--_
. .
*
.
r = -0.30 P < 0.025
0
10
20
30 AGE
40
50
60
70
[YEARS)
FIGURE 5. Change in stroke volume with age.
mum and then usually became positive for the higher harmonics. These impedance spectral patterns are similar to those reported previously by ~srsJs*~sand others.24 The nonpulsatile component of LV load, peripheral resistance, increased (from 1,049 to 1,438 dynes s cm-s) 37% (9% per decade) over the age range of 20 to 60 years (r = 0.47, p
Old
$!I
\/-
“ii 2
4
6
8
10
FIGURE 6. Aortic input impedance spectra (modulus and phase) and flow modulus (lop) vs frequency in a young (lefl) ancJanoldsubject(rlgM).
k 2
4
6
8
10
April 15, 1965
The relation between the left ventricle and its vascular load is best considered by comparing the modulus of ascending aortic impedance against the frequency components of LV ejection (flow wave).13J7A favorable relation is apparent from the low impedance values over the frequency range that normally contains most of the energy of the LV ejection wave (Fig. 6). Age disturbs this relation by displacing impedance spectra to the right, so that the impedance modulus is even greater at the frequency of the first flow harmonic than one would predict from consideration of change in characteristic impedance alone. This point illustrates one of the shortcomings in use of characteristic impedance and peripheral resistance as LV load. The LV load is affected not only by aortic distensibility and arteriolar caliber but also by reflected waves from arterial terminations.s13J4 Earlier wave reflection resulting from arterial stiffening causing increase in pulse wave velocity26r27would be expected to increase vascular load more than through increase in characteristic impedance alone.sJ3J4 Another feature was the increase in fluctuations of impedance modulus with increasing age (Fig. 6). This can be explained by differences in pulse wave
PERIPHERAL
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume 55
1193
velocity in the brachiocephalic and subclavian arterial system on the one hand and the descending aortic and femoral arterial system on the other.2s In young subjects, reflections return much earlier from the upper body than from the lower body, thus attenuating fluctuation in modulus.lsJs As age increases, wave velocity increases to a lesser degree in upper body arteries than in the descending aorta and lower body arteries, so that major reflections return at much the same time, with the systemic circulation appearing to present a single (rather than 2) discrete reflecting site, and the impedance modulus showing considerable fluctuation, as predicted in such a condition.23 Change in pressure pulse wave contour is largely responsible for the increase in aortic systolic pressure. This is also explicable on the basis of increased aortic
FREQUENCY
1Z lrn~n
84
RESISTANCE
2000 -I .
.
. ..
.
15001
.
.
.-se---,
.
6
_ -,-‘.
.
g
i
I--.,
0
10
20
30 AGE
.
.
2
__--
,
. *
__--
g lOOO*
.
- -‘yy-
P < 0.001 l
--r)--
l __-
.
?
r = 0.65
.
-7-p-y 50
40
60
70
(YEARS)
FIGURE 9. Change in frequency of the impedance modulus minimum d&d with we.
. r = 0.47
5ooj
P < 0.001
0
10
20
180 -
r-IT 40
30 AGE
50
60
70
(YEARS)
lso-
FfGURE 7. Change in nonpulsattle component of left ventricular external load (peripheral vascular resistance) with age. 140 a ?
CHARACTERISTIC
150
:
IMPEDANCE
120-
5 D i
loo-
f
so-
. . -*--
. ’ l .
f;--
. .
--.--*
: .,-
. .
. .-*-. . -
.-
(
W-
-+
l.
.-*.,
---.
---
_-
-
. :.
:
I
0’
r 7 0.66
.
P c 0001
I 10
I 10
I 30
AGE IN YEARS L-
0
r-
10
.--.
20
--.
. ..-
30
--..--.
60
70
FIGURE 9. Change in pulsatile component of left ventricular load (characteristic impedance) with age.
external
AGE
40
---
50
(YEARS)
I 40
I w
I so
I
I
70
so
AT LAST BIRTHDAY
FIGURE 10. Change in brachial artery systolic and diastolic pressures with age in American male subjects (solld Ilnes). From the U.S. National Center for Health Statistics National Survey (1977). Uotted lines indicate expected changes in ascending aortic systolic and diastolic pressures with age.
1104
AGE AND AORTIC
IMPEDANCE
stiffness and earlier wave reflection.26 Change in characteristic impedance alone (resulting from increase in aortic stiffness) could only cause increase in aortic pulse pressure (for a given stroke volume) but no change in wave contour. Change in wave contour is due to change in timing of wave reflection.16 The younger subjects in this study almost invariably exhibited a type C aortic pressure waveform with the foot of the diastolic wave corresponding with or following the incisura caused by aortic valve closure. This is the type of pattern previously noted in children’s and regularly seen in sheep, dogs, rabbits and other large experimental animals. l3 This pattern is explicable on the basis of wave reflection from the lower body returning at the conclusion of ventricular ejection. The type A pattern, almost invariably seen in the older subjects, is attributable to earlier return of reflected waves, with the reflected wave from the lower body returning during midsystole so that pressure in late systole is augmented. l6 The type B pattern in the middle-aged subjects is attributable to intermediate timing of wave reflection. The findings described here were from occidental subjects without apparent cardiovascular disease. As shown by coronary angiography, none had significant coronary atherosclerosis, and they had no direct evidence of atherosclerosis elsewhere. The changes seen here were most likely associated with aging, but the role of subclinical atherosclerosis cannot be excluded. Similar changes in pulse wave velocity and in arterial pressure with age have been found in occidental subjects, in whom atherosclerosis is prevalent, and in oriental subjects, in whom it is not.28 The major findings described here can be attributed to increased arterial stiffness with age because age affects stiffness of the ascending aorta itself (and so characteristic impedance) or because age affects the timing of wave reflection from peripheral sites. Both increased aortic stiffness and early wave reflection clearly affect cardiac performance adversely26and may account for the mild LV hypertrophys and prolonged relaxationr7 with advancing age. We currently are studying ways by which aortic and arterial stiffness may be altered directly or indirectly by pharmacologic means,30and how pulse wave reflection may be reduced or delayed pharmacologically to benefit the heart. We believe that such interventions are feasible and logical and follow directly from the studies on the pathophysiology of aging on the cardiovascular system.
References 1. Young T. On the functions of the heart and arteries. The Croonian lecture. Philos Trans R Sot 1809;99: l-3 1. 2. G&tones MA, Gaasch WH, Alexander TK. Influence of acute changes in preload. afterload, ccntractile state and heart rate on ejectll and isovofumic indices of myocardial contractility in man. Circulation 1976;53:293-302. 3. Elzlnga 0, Westerhof N. Pressure and flow generated by the left ventricle against different impedances. Circ Res 1973;32:178-186. 4. SaIlsbury PF, Cross CE, Rieben PA. Ventricular performance modified by elastic properties of outflow system. Circ Res 1962;11:319-326. 5. O’Rwrke MF. Steady and pulsatite energy losses in the systemic circulation under normal conditions and simulated arterial disease. Cardiovasc Res 1967;1:313-326. 6. Shock NW. Physiological aspects of aging in man. Annu Rev Physiol 1961;23:97-122. 7. Lakatta EG, Gerstenblllh G, Angel1 CS, Shock NW, Welsfekll ML. Prolonged contraction duration in aged myocardium. J Clin Invest 197555: Rl-RR 8. Nichols WW, Peplne CJ. Left ventricular afterload and aortic input impedance. Implications of pulsatile blood flow. Prog Cardiovasc Dis 1962:34:293-306. 9. Nichdls WW, Peplne CJ, Gelser EA, Contl CR. Vascular load defined by the aortic input impedance spectrum. Fed Proc 1960;39:196-201. IO. Gerstenblllh G, Lakatla EG, Welsfeklt ML. Age changes in myocardial function and exercise response. Prog Cardiovasc Dis 1976;19:1~21. 11. O’Flourlce MF, Taylor MO. Input impedance of the systemic ctrculattkrn. Circ Res 1967;20:365-360. 12. Mllnor WR. Arterial imoedance as ventricular afterload. Circ Res 1975: 36565-570. ’ 13. O’Rourke MF. Arterial Function in Health and Disease. London: Churchill Livingstone, 1962;133-152. 14. McDonald DA. Blood Flow in Arteries. London: Edward Arnold, 1974: 351-366. 15. Nichols WW, Contl CR, Walker WE, Mllnor WR. Input impedance of the systemic circulation in man. Circ Res 1977;40:451-456. 16. Murgo JP, Westerhoi N, Glolma JP, Altobelll SA. Aortic input impedance in normal man: relationship to pressure wave shape. Circulation 1960; 62:105-l 16. 17. Schultz DL. Pressure and flow in large arteries. In: Bergel DH. ed. Cardio vascular Fluid Dynamics. London: Academic Press, 1972:287-314. 19. Nichols WW, Peplna CJ, Contl CR, Ckrtstle LG Evaluation of a new catheter mounted electromagnetic velocity sensor during cardiac catheterization. Cathet Cardiovasc Diagn 1960;6:97-113. 19. Murgo JP, Glolma JP, Allobelll SA. Signal acquisition and processing for human hemcdynamic research. Proc IEEE 1977;65:696-702. -.
““.
20. Merlllon JP, Yotte 0, Masquet C, Azancol I, Gulomard A, Gourgon R. Relationship between physical properties of the arterial system and left ventricular performance in the course of aging and arterial hypertension. Eur Heart J 1962;3:suppl A:95-102. 21. Learoyd EM, Taykr MO. Alterations with age in the visco-elastic prop&& of human arterial walls. Circ Res 1966;16:278-292. 22. Brandfonbrener M, Landowne M, Shook NW. Changes in cardiac output with age. Circulation 1955;12:557-566. 23. D’Rourke MF. Avollo AP. Pulsatile flow and oressure in human svstemic arteries: studies in man and in a multi-branched model of the human systemic arterial tree. Circ Res 1960;46:363-372. 24. Mills CJ, Gabe lT, Gault JH, Mason DT, Ross J Jr, Braunwakl E, Shltthgkwd JP. Pressur~flow relationships and vascular impedance in man. Cardiovasc Res 1970;4:405-417. 25. U.S. Government National Center for Health Statistics. In: Roberts J. ed. Blood Pressure Levels of Persons 6-74 Years. United States 1971-1974; 1977; DHEW Publ. No. (HRA) 76-1646. 20. Schatz RA, Paslpoularldes A, Murgo JP. The effect of arterial pressure reflections on myocardial supply-demand dynamics (abstr). Circulation 1961;64:suppl Iv:IV-324. 27. O’Rourke MF. Vascular impedance in studies of arterial and cardiac function. Physiol Rev 1962;62:570-623. 20. Avollo AP, Chen S, Wang R, Zhang C, LI M, D’Rourke MF. Effects of aging on changing arterial compliance and left ventricular load in a northern Chinese urban community. Circulation 1963;68:50-56. 29. Cox RH. Effects of age on the mechanical properties of rat carotid artery. Am J Physiol 1977;233:H256-H263. 30. Yaglnuma T, Morgan J, Roy P, Baron D, Feneley M, Branson J, Avollo AP, O’Rourke MF. Mechanism of nitroglycerin action in humam adults without heart failure: a new hypothesis (abstr). Aust NZ J Med 1963:13:87.
The effects of age on the interrelation between the physical properties of the arterial tree (aortic input impedance) and left ventricular performance (cardiac output) were studied in 45 subjects, aged 19 to 62 years, without apparent cardiovascular disease. Ascending aortic pulsatile pressure and blood flow velocity were measured with a multisensor catheter and cardiac output by green dye or the Fick method. Heart rate and end-diastolic aortic pressure remained unchanged with age, whereas aortic systolic, mean and pulse pressures and aortic radius increased. In subjects younger than 30 years, early systolic pressure usually exceeded late systolic pressure (type C beat); in subjects older than 50 years, late systolic pressure usually exceeded early systolic pressure (type A beat). In 55 % of subjects aged 30 to 50 years, early and late systolic pressures were essentially equal (type B beat). The impedance spectra from all subjects showed fluc-
tuations about the characteristic impedance (index of elastance) that were greater in the older subjects. Peripheral resistance increased 37 % (r = 0.47, p
The effects of aging on cardiovascular function have concernedphysiciansand physiologistssince1809,when Thomas Young1 introduced his modulus of elasticity to study arterial function. However, remarkably little is known about the detailed effects of aging on cardiovascular hemodynamics. Vascular load increases through increases in arterial pressure,2 peripheral re-
tra.3~8~g~11~12 Aortic input impedance is a function of systemic vascular resistance, central aortic elastance (inverseof compliance)or stiffnessand wavereflections, and it selectively describesthe arterial system.13~14 The present study provides detailed information on the effects of ageon LV-vascular coupling in human subjects without apparent cardiovascular disease.
sistance3 and arterial stiffness.3-5Aging is associated with mild cardiac hypertrophy6 and prolonged or de-
Methods
layed left ventricular
(LV) relaxation.’
Increases in
The data reported here were obtained from studiesperformed at The JohnsHopkinsUniversity(n = lo), The University of Florida (n = 13), Brooke Army Medical Center (n = 18), and The University of New South Wales(n = 4) and have beenpublished in part elsewhere.15J6 The observations were made on 45 adults, aged 19 to 62 years, who were
vascular load are associatedwith decreasedstroke voland decreased capacity for increasing stroke ume3y8vg volume with exercise.lO LV external vascular load can
be characterized by the aortic input impedance spec-
undergoingcardiaccatheterizationfor variousclinicalindications,the mostcommonof whichwasa chestpainsyndrome.
From the Department of Medicine, University of Florida, and Veterans Administration Medical Center, Gainesville, Florida; Department of Medicine, St. Vincent’s Hospital, Sydney, Australia: Department of Medicine, Cardiology Service, Brooke Army Medical Center, Fort Sam Houston, Texas. Manuscript received July 20, 1984; revised manuscript received December 17, 1984, accepted December 3 1, 1984. Address for reprints: Wilmer W. Nichols, PhD, Department of Meditine, Box J-277, University of Florida, Gainesville, Florida 32610.
No cardiovasculardiseasewasfound by hemodynamicmeasurements, LV cineangiographyor coronary arteriography. During basalstate studies all patients wereeither unsedated or very lightly sedatedwith oral diazepam(10 mg 1 hour before the procedure)or sodium pentobarbital(100 mg orally). Multisensor catheters,custom-designedfor the left side of the 1179
1180
AGE AND AORTIC
TABLE I
IMPEDANCE
Effects of Age on Arterial Hemodynamics Variable
Slope
Intercept
Correlation Coefficient
Aortic radius (cm) Heart rate (beatsimin) Systolic pressure (mm Hg) Diastolic pressure (mm Hg) Mean pressure (mm Hg) Pulse pressure (mm Hg) Cardiac output (litersimin) Stroke volume (ml) Peak flow (ml/s) Characteristic impedance (dynes s cm+) Peripheral resistance (dynes s cmm5) Frequency of 1Z 1,,,I” (Hz)
0.0097 0.0486 0.6852 0.0702 0.1911 0.6150 -0.0365 -0.5605 -3.2638 1.2135 9.7409 0.0364
0.98 73.64 95.00 76.12 90.27 18.89 7.85 109.30 777.92 10.3 853.97 2.41
0.61 0.05 0.63 0.11 0.33 0.70 -0.37 -0.30 -0.31 0.66 0.47 0.65
T 5.08 0.30 5.36 0.73 2.26 6.38 -2.63 -2.076 -2.17 5.71 3.45 5.59
p Value 0.001 0% ‘NS 0.01 0.001 0.005 0.025 0.025 0.001 0.001 0.001
IZI,,,i, = first minimum of impedance moduli; NS = not significant.
heart, were used in 40 subjects. The catheters each contained 2 laterally mounted solid-state pressure sensors (Millar Instruments) and an electromagnetic blood flow velocity sensor (Carolina Medical Electronics or Millar Instruments). One pressure sensor was mounted at the same site as the velocity sensor, and the other was mounted at the catheter tip 5 to 7 cm away and was used for recording LV pressure.A fluid-filled : system was used to record aortic and LV pressures in the remaining 5 subjects. The frequency responseof this system was’ constant in amplitude (f5%) from 0 to 20 Hz with a damping ratio of 0.15. This response is adequate for accurate measurement of pulsatile pressure in the human aorta, because more than 98% of the variance of the pulsations is included in the first 7 harmonies.14Js The velocity-pressure catheter was advanced through a brachial arteriotomy to the aortis valve so that the tip was within the LV cavity. This arrangement placed the velocity probe and associatedpressure sensor 3 to 5 cm above the aortic valve at approximately the upper border of the sinuses of Valsalva. This arrangement tends to stabilize the transducer in the central axis of the ascending aorta where the velocity profile is relatively flat.17 The velocity sensors were used in conjunction with either a square wave (Carolina Medical Electronics) or a sine wave (Biotronex or SE Laboratories) blood flowmeter. Details of the technical characteristics and clinical use of these flowmeters and multisensor catheters, including frequency response, drift characteristics and calibration techniques, have been described previously.lsJg LV cineangiography and coronary arteriography were performed after hemodynamic data were collected. If the aortic root was not adequately visualized during the ventriculographic study, a selective aortic root angiogram was performed to measure the mean systolic radius of the ascending aorta at the site where the pressure and velocity were recorded. The mean output signal (velocity) of the flowmeter was calibrated in volumetric flow units (milliliters per second) by. reference to a simultaneous determination of cardiac output by the dye dilution or Fick method. The pressure and flow velocity signals were recorded on analog magnetic tape and later digitized at a sampling interval of 5 or 10 ms. The technical details of the analog and digital processing have been described elsewhere.ls*is Data were analyzed on a digital computer, which converted pressure and flow velocity signals to Fourier series, applied corrections for the measured dynamic responses of the transducer and computed aortic impedance (modulus and phase) as functions of frequency. Input impedance modulus at each harmonic frequency was computed by dividing pressure modulus by flow modulus, and the impedance phase was computed by sub-
tracting the phase angle of flow from that of pressure.14Details regarding signal averaging and techniques for removing spurious data caused by noise have been published.l5Js The impedance at 0 Hz or “input resistance” was calculated by dividing mean aortic pressure by mean flow. Charactaristic aortic impedance was estimated by averaging impedance moduli above 2 Hz. Characteristic impedance depends on the viscoelastic properties of the artery under study, whereas input impedance moduli oscillate around the characteristic value because of waves reflected from more distal
points.*3J4 The systolic configuration of the ascending aortic pressure wave form was classified according to the criteria of Murgo et all6 Type A: Peak systolic pressure (Ppk) occurs in late systole after a well-defined anacrotic notch or inflection point (Pi)
and (Ppk-Pi)/(pulse pressure)is greater than 0.12. Type B: Peak systolic pressurealso occurs in late systole following inflection point, but (Ppk-Pi)/(pulse pressure) is
between0.0 and 0.12. Type C: Peak systolic pressure precedes a well-defined
inflection point, and (Ppk-PiMpulse pressure) is less than 0.0. The relation between age and the different variables was obtained by means of least-squares linear regression analysis.
Results The subjects’ ages were fairly well distributed over 4 decades. However, there was a larger concentration of patients (18) in the fourth decade. Of all measurements collected, only heart rate (range 56 to 110 beats/min, average 76 f 3, difference not significant) and aortic diastolic pressure (range 64 to 94 mm Hg, average 76 f 2, difference not significant) were not age-related. This finding agreeswith observations made by others.10*20 The relation between age and other cardiovascular variables are given in Table I. Aortic radius: The mean systolic internal radius of the ascending aorta measured at approximately the same site as the pressure and blood flow velocity (3 to 5 cm above the aortic valve) ranged from 1.0 to 1.8 cm and showed a positive correlation with age (r = 0.61, p CO.001) (Fig. 1). This finding represents an increase of 9% in radius or 20% increase in cross-sectional area per decade. These results are similar to those reported previously.20*21
April 15. 1985
Aortic pressure waveform: Pressure wave forms recorded in subjects younger than 30 years were generally type C beats (75%), whereas those recorded in subjects older than 50 years were type A beats (90%). Subjects aged 30 to 50 years showed 11% type C, 55% type B and 33% type A beats. Murgo et al16 also noted a similar relation between pressure wave form and age. Examples of pressure waves recorded in a young subject and an old subject are shown in Figure 2. Over the age range of 20 to 60 years, aortic systolic pressure increased 25% (from 109 to 136 mm Hg), or 6% per decade (r = 0.63, p
AORTIC
RADIUS
. . . . -L-- __-. . . .-.-C.-L . . . . t---.. . . . . e,-i;-.. . .. ---:. . _-- . . .. * * .
r = 0.61 P < 0001
L 0
-.T--.I 10
20
30 AGE
._.-(..-.^.---. 40
50
60
JOURNAL
OF CARDIOLOGY
Volume 55
1161
94 to 102 mm Hg), or 2% per decade (r = 0.33, p
70
(YEARS1
1. Change in mean systolic internal ascending aortic radius with age. Radius was measured from contrast angiograms at approximately the same site as the pressure and velocity (3 to 5 cm above the aortic valve).
FIGURE
ASCENDING
THE AMERICAN
200
AORTIC
PRESSURE I = 0.63
AORTA rz011 P = NS (]w
----T--
30 AGE
I
40
50
60
70
circles)
and
(YEARS)
FIGURE 3. Change in ascending aortic systolic (closed diastolic (open circles) pressure with age.
154 CARDIAC
OUTPUT
r = -037 P < 0.005
L--m. 0 and pressure waveforms recorded in a young (type C pressure wave) and an old (type A pressure wave) subject. FIGURE
2. Ascending aortic blood flow velocity
I.. 10
20
30 AGE
FIGURE
, .1 40 (YEARS)
4. Change in cardiac output with age.
-..50
-I 60
70
1182
AGE AND AORTIC
IMPEDANCE
STROKE
The changes in arterial pressure are a little different from those of epidemiologic studies, in which pressure was recorded indirectly in the brachial artery with the subject sitting (rather than recumbent) and unsedated and uninstrumented.25 We find here a greater change in systolic pressure with age than was found in those studies and a smaller change in diastolic pressure. Although the aforementioned factors may be responsible and our data are minute compared with the mass of epidemiologic data on brachial artery pressure, we stress that recordings of brachial artery pressure may be quite different from recordings of central aortic pressure26 and that the epidemiologic studies take no account of changing amplification between ascending aorta and brachial artery with age. The changes reported here are similar to those predicted when change in brachial -amplification was taken into account (Fig. 10).1s*2s The studies reported here show a definite increase in both pulsatile (characteristic impedance and reflections) and nonpulsatile (peripheral resistance) components of LV vascular load. The increase in characteristic impedance was surprisingly high-137% between ages 20 and 60 years. This increase would have been even higher if allowance had been made for change in aortic diameter and this had been expressed in terms Merillon et a120calculated the of linear flow velocity. 13127 increase in characteristic impedance as approximately 80% between ages20 and 60 years, whereas other studies of aortic pulse wave velocity suggest change in characteristic impedance of approximately 60% over this age range.28 The cause of change in vascular load can be inferred from the change in peripheral resistance and in characteristic impedance. The former is determined by opposition to steady flow in the small peripheral vessels, whereas the latter is determined by arterial stiffness. A small change occurs in resistance and can be attributed to decreased caliber or decreased total cross-sectional area of small vessels, whereas the greater change is in arterial stiffness and is attributable to asymptomatic arterial degeneration and a possible increase in the collagen-elastin ratio.29
VOLUME
150 2 .5
Y 3
.
1
-----_____
loo
.
*. .. * .** . ;‘--.----;$. --r---z . . ---;.** -* . . *.
.
.
G
.
I
50
-
.
l ’
-____
I
--_
. .
*
.
r = -0.30 P < 0.025
0
10
20
30 AGE
40
50
60
70
[YEARS)
FIGURE 5. Change in stroke volume with age.
mum and then usually became positive for the higher harmonics. These impedance spectral patterns are similar to those reported previously by ~srsJs*~sand others.24 The nonpulsatile component of LV load, peripheral resistance, increased (from 1,049 to 1,438 dynes s cm-s) 37% (9% per decade) over the age range of 20 to 60 years (r = 0.47, p
Old
$!I
\/-
“ii 2
4
6
8
10
FIGURE 6. Aortic input impedance spectra (modulus and phase) and flow modulus (lop) vs frequency in a young (lefl) ancJanoldsubject(rlgM).
k 2
4
6
8
10
April 15, 1965
The relation between the left ventricle and its vascular load is best considered by comparing the modulus of ascending aortic impedance against the frequency components of LV ejection (flow wave).13J7A favorable relation is apparent from the low impedance values over the frequency range that normally contains most of the energy of the LV ejection wave (Fig. 6). Age disturbs this relation by displacing impedance spectra to the right, so that the impedance modulus is even greater at the frequency of the first flow harmonic than one would predict from consideration of change in characteristic impedance alone. This point illustrates one of the shortcomings in use of characteristic impedance and peripheral resistance as LV load. The LV load is affected not only by aortic distensibility and arteriolar caliber but also by reflected waves from arterial terminations.s13J4 Earlier wave reflection resulting from arterial stiffening causing increase in pulse wave velocity26r27would be expected to increase vascular load more than through increase in characteristic impedance alone.sJ3J4 Another feature was the increase in fluctuations of impedance modulus with increasing age (Fig. 6). This can be explained by differences in pulse wave
PERIPHERAL
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume 55
1193
velocity in the brachiocephalic and subclavian arterial system on the one hand and the descending aortic and femoral arterial system on the other.2s In young subjects, reflections return much earlier from the upper body than from the lower body, thus attenuating fluctuation in modulus.lsJs As age increases, wave velocity increases to a lesser degree in upper body arteries than in the descending aorta and lower body arteries, so that major reflections return at much the same time, with the systemic circulation appearing to present a single (rather than 2) discrete reflecting site, and the impedance modulus showing considerable fluctuation, as predicted in such a condition.23 Change in pressure pulse wave contour is largely responsible for the increase in aortic systolic pressure. This is also explicable on the basis of increased aortic
FREQUENCY
1Z lrn~n
84
RESISTANCE
2000 -I .
.
. ..
.
15001
.
.
.-se---,
.
6
_ -,-‘.
.
g
i
I--.,
0
10
20
30 AGE
.
.
2
__--
,
. *
__--
g lOOO*
.
- -‘yy-
P < 0.001 l
--r)--
l __-
.
?
r = 0.65
.
-7-p-y 50
40
60
70
(YEARS)
FIGURE 9. Change in frequency of the impedance modulus minimum d&d with we.
. r = 0.47
5ooj
P < 0.001
0
10
20
180 -
r-IT 40
30 AGE
50
60
70
(YEARS)
lso-
FfGURE 7. Change in nonpulsattle component of left ventricular external load (peripheral vascular resistance) with age. 140 a ?
CHARACTERISTIC
150
:
IMPEDANCE
120-
5 D i
loo-
f
so-
. . -*--
. ’ l .
f;--
. .
--.--*
: .,-
. .
. .-*-. . -
.-
(
W-
-+
l.
.-*.,
---.
---
_-
-
. :.
:
I
0’
r 7 0.66
.
P c 0001
I 10
I 10
I 30
AGE IN YEARS L-
0
r-
10
.--.
20
--.
. ..-
30
--..--.
60
70
FIGURE 9. Change in pulsatile component of left ventricular load (characteristic impedance) with age.
external
AGE
40
---
50
(YEARS)
I 40
I w
I so
I
I
70
so
AT LAST BIRTHDAY
FIGURE 10. Change in brachial artery systolic and diastolic pressures with age in American male subjects (solld Ilnes). From the U.S. National Center for Health Statistics National Survey (1977). Uotted lines indicate expected changes in ascending aortic systolic and diastolic pressures with age.
1104
AGE AND AORTIC
IMPEDANCE
stiffness and earlier wave reflection.26 Change in characteristic impedance alone (resulting from increase in aortic stiffness) could only cause increase in aortic pulse pressure (for a given stroke volume) but no change in wave contour. Change in wave contour is due to change in timing of wave reflection.16 The younger subjects in this study almost invariably exhibited a type C aortic pressure waveform with the foot of the diastolic wave corresponding with or following the incisura caused by aortic valve closure. This is the type of pattern previously noted in children’s and regularly seen in sheep, dogs, rabbits and other large experimental animals. l3 This pattern is explicable on the basis of wave reflection from the lower body returning at the conclusion of ventricular ejection. The type A pattern, almost invariably seen in the older subjects, is attributable to earlier return of reflected waves, with the reflected wave from the lower body returning during midsystole so that pressure in late systole is augmented. l6 The type B pattern in the middle-aged subjects is attributable to intermediate timing of wave reflection. The findings described here were from occidental subjects without apparent cardiovascular disease. As shown by coronary angiography, none had significant coronary atherosclerosis, and they had no direct evidence of atherosclerosis elsewhere. The changes seen here were most likely associated with aging, but the role of subclinical atherosclerosis cannot be excluded. Similar changes in pulse wave velocity and in arterial pressure with age have been found in occidental subjects, in whom atherosclerosis is prevalent, and in oriental subjects, in whom it is not.28 The major findings described here can be attributed to increased arterial stiffness with age because age affects stiffness of the ascending aorta itself (and so characteristic impedance) or because age affects the timing of wave reflection from peripheral sites. Both increased aortic stiffness and early wave reflection clearly affect cardiac performance adversely26and may account for the mild LV hypertrophys and prolonged relaxationr7 with advancing age. We currently are studying ways by which aortic and arterial stiffness may be altered directly or indirectly by pharmacologic means,30and how pulse wave reflection may be reduced or delayed pharmacologically to benefit the heart. We believe that such interventions are feasible and logical and follow directly from the studies on the pathophysiology of aging on the cardiovascular system.
References 1. Young T. On the functions of the heart and arteries. The Croonian lecture. Philos Trans R Sot 1809;99: l-3 1. 2. G&tones MA, Gaasch WH, Alexander TK. Influence of acute changes in preload. afterload, ccntractile state and heart rate on ejectll and isovofumic indices of myocardial contractility in man. Circulation 1976;53:293-302. 3. Elzlnga 0, Westerhof N. Pressure and flow generated by the left ventricle against different impedances. Circ Res 1973;32:178-186. 4. SaIlsbury PF, Cross CE, Rieben PA. Ventricular performance modified by elastic properties of outflow system. Circ Res 1962;11:319-326. 5. O’Rwrke MF. Steady and pulsatite energy losses in the systemic circulation under normal conditions and simulated arterial disease. Cardiovasc Res 1967;1:313-326. 6. Shock NW. Physiological aspects of aging in man. Annu Rev Physiol 1961;23:97-122. 7. Lakatta EG, Gerstenblllh G, Angel1 CS, Shock NW, Welsfekll ML. Prolonged contraction duration in aged myocardium. J Clin Invest 197555: Rl-RR 8. Nichols WW, Peplne CJ. Left ventricular afterload and aortic input impedance. Implications of pulsatile blood flow. Prog Cardiovasc Dis 1962:34:293-306. 9. Nichdls WW, Peplne CJ, Gelser EA, Contl CR. Vascular load defined by the aortic input impedance spectrum. Fed Proc 1960;39:196-201. IO. Gerstenblllh G, Lakatla EG, Welsfeklt ML. Age changes in myocardial function and exercise response. Prog Cardiovasc Dis 1976;19:1~21. 11. O’Flourlce MF, Taylor MO. Input impedance of the systemic ctrculattkrn. Circ Res 1967;20:365-360. 12. Mllnor WR. Arterial imoedance as ventricular afterload. Circ Res 1975: 36565-570. ’ 13. O’Rourke MF. Arterial Function in Health and Disease. London: Churchill Livingstone, 1962;133-152. 14. McDonald DA. Blood Flow in Arteries. London: Edward Arnold, 1974: 351-366. 15. Nichols WW, Contl CR, Walker WE, Mllnor WR. Input impedance of the systemic circulation in man. Circ Res 1977;40:451-456. 16. Murgo JP, Westerhoi N, Glolma JP, Altobelll SA. Aortic input impedance in normal man: relationship to pressure wave shape. Circulation 1960; 62:105-l 16. 17. Schultz DL. Pressure and flow in large arteries. In: Bergel DH. ed. Cardio vascular Fluid Dynamics. London: Academic Press, 1972:287-314. 19. Nichols WW, Peplna CJ, Contl CR, Ckrtstle LG Evaluation of a new catheter mounted electromagnetic velocity sensor during cardiac catheterization. Cathet Cardiovasc Diagn 1960;6:97-113. 19. Murgo JP, Glolma JP, Allobelll SA. Signal acquisition and processing for human hemcdynamic research. Proc IEEE 1977;65:696-702. -.
““.
20. Merlllon JP, Yotte 0, Masquet C, Azancol I, Gulomard A, Gourgon R. Relationship between physical properties of the arterial system and left ventricular performance in the course of aging and arterial hypertension. Eur Heart J 1962;3:suppl A:95-102. 21. Learoyd EM, Taykr MO. Alterations with age in the visco-elastic prop&& of human arterial walls. Circ Res 1966;16:278-292. 22. Brandfonbrener M, Landowne M, Shook NW. Changes in cardiac output with age. Circulation 1955;12:557-566. 23. D’Rourke MF. Avollo AP. Pulsatile flow and oressure in human svstemic arteries: studies in man and in a multi-branched model of the human systemic arterial tree. Circ Res 1960;46:363-372. 24. Mills CJ, Gabe lT, Gault JH, Mason DT, Ross J Jr, Braunwakl E, Shltthgkwd JP. Pressur~flow relationships and vascular impedance in man. Cardiovasc Res 1970;4:405-417. 25. U.S. Government National Center for Health Statistics. In: Roberts J. ed. Blood Pressure Levels of Persons 6-74 Years. United States 1971-1974; 1977; DHEW Publ. No. (HRA) 76-1646. 20. Schatz RA, Paslpoularldes A, Murgo JP. The effect of arterial pressure reflections on myocardial supply-demand dynamics (abstr). Circulation 1961;64:suppl Iv:IV-324. 27. O’Rourke MF. Vascular impedance in studies of arterial and cardiac function. Physiol Rev 1962;62:570-623. 20. Avollo AP, Chen S, Wang R, Zhang C, LI M, D’Rourke MF. Effects of aging on changing arterial compliance and left ventricular load in a northern Chinese urban community. Circulation 1963;68:50-56. 29. Cox RH. Effects of age on the mechanical properties of rat carotid artery. Am J Physiol 1977;233:H256-H263. 30. Yaglnuma T, Morgan J, Roy P, Baron D, Feneley M, Branson J, Avollo AP, O’Rourke MF. Mechanism of nitroglycerin action in humam adults without heart failure: a new hypothesis (abstr). Aust NZ J Med 1963:13:87.