Vol. 13, No, 1 Printed in U.S.A.
PsYfHOPHYSIOl.OGY
Copyrighf '' i76 by The Sociely for Psychciphysioldgical Research
Methodology Pulse Wave Velocity as a Measure of Blood Pressure Change BRIAN GRIBBIN, Department of Cardiovascular Medicine, Radclife Infirmary, Oxford ANDREW STEPTOE, Department of Psychiatry, Oxford University AND PETER S L E I G H T
Department of Cardiovascular Medicine, Radclife Infirmary, Oxford
ABSTRACT The use of arterial pulse wave velocity (1*WV) as a continuous measure of blood pressure changes is outiined. Theoretical considerations indicate that changes in PWV reflect changes in blood pressure, and an experiment was carried out to assess this relationship. PWV along an arm artery was monitored in 26 subjects at a time when the arterial distending pressure of the limb was altered over a wide range by means of externally applied positive and negative pressures. The results show that changes in PWV reliably follow changes in blood pressure. This method can be considered suitable for studies requiring changes rather than absolute values of blood pressure. DESCRIPTORS: Pulse wave velocity, Blood pressure. Methodology. A major obstacle to the fuller understanding of cardiovascular psychophysiology has been the difficulty of continuous blood pressure measurement. Direct pressure monitoring by arterial cannulation is seldom used, even though the procedure may not be as stressful as is frequently supposed (Beamer & Shapiro, 1973). In recent years, hovs'ever, a number of indirect methods have been developed using sophisticated adaptations of the sphygmomanometric technique (Tursky, 1974). For example one system uses an arm cutF which is automatically inflated to a preset level near the systolic or diastolic pressure (Tursfcy, Shapiro, & Schwartz, 1972). The presence or absence of Korotkoff sounds is detected by a microphone in the cuff, and indicates whether pressure is above or below the preset level. This system thus provides information on every heart beat, but this is of a binary nature only since there is no indication of the size of any deviation from inflation pressure. Another disadvantage is that continuous monitoring Research was partially supported by the Medical Research Council. U.K., and the work carried out by Dr. Gribbin took place during his tenure of a Beit Memorial Junior Research Fellowship. Address requests for reprints to: Brian Gribbin, Department of Cardiovascular Medicine, Radcliffe Infirmary, Oxford, U.K.
is only possible for short periods (50 heart beats in Shapiro, Schwartz, & Tursky, 1972), since pressure must be released to allow blood flow into the arm, and so readings are available for only about 70% of the time. Another approach to this problem is reported by Lategola, Harrison, & Barnard (1966) and Brener (1974) who used a technique in ; which an arm cuif is rapidly inflated to near previous levels of pressure and then the system is switched to: a slow inflation rate until a Korotkoff sound is detected. This enables the diastolic blood pressure to be tracked, but of course readings are limited to every third or fourth heart beat. Neither of these techniques is thus ideal since they do not provide immediate analogue information on every heart beat. A number of problems common to occlusion methods can also be mentioned. For instance, there are inherent inaccuracies in auscultatory blood pressure recordings when compared with direct measurements (Geddes, 1970); systolic pressures tend to be low while diastolic values are commonly too high, although the latter depend on whether muffling or disappearance of Korotkoff sounds is used as the criterion (Ragan & Bardley, 1941; Roberts, Smiley, & Manning, 1953). The problem of accurate detection of the diastolic pressure appears to be more pronounced with automatic de86
January, 1976
PULSE WAVE VELOCITY
tection techniques, while incorrect cuff size and length can distort pressure readings and increase variability (Greatorex, 1971; Khosla & Lowe, 1965; Karvonen, Telivvo, & Jarvinen, 1964). The measurement of Pulse Wave Velocity (PWV) may provide an attractive alternative to the automated pneumatic methods, since immediate analogue readings can be made on every beat. PWV is the rate of propagation of pressure pulse waves along arteries, and short term changes in this parameter are due primarily to alterations in blood pressure. The velocity with which a pressure wave travels along an artery is determined by the mechanical properties of that artery, and the relationship between the two has been described by equations derived from bench experiments in which thin-walled elastic tubing was substituted for artery. One such equation was reported by Bramwell and Hill (1922): '^
Av
p
87
Coghlan, Gosling, Pickup. Newman, & Woodcock, 1972). Pressure waves used in the calculation of PWV have been recorded through intra-arterial needles or catheters and also by the use of external transducers accurately placed over the artery (Landowne, 1957; Gribbin, 1974). A modification of this method is to record the time delay between the R wave of the EKG and the upstroke of a single peripherally recorded pulse wave (Weltman, Sullivan, & Brendon, 1964). PWV values derived from such experiments have been found to increase during mental arithmetic and are considered likely to parallel changes in arterial pressure (Brod, Fend, Hejl, & Jirka, 1959; Williams & Williams, 1965). It is the purpose of this paper to examine the relationship of PWV to changes in arterial pressure within individuals, as opposed to the betweensubject comparisons more commonly made. The application of PWV monitoring to psychophysio^ logical experiments will also be considered. Methods
where c = PWV metres/sec, Ap = change in pressure, Av == change in volume, V = initial Volume, and p = density of fluid. One can observe from this equation that PWV is inversely related to true compliance or distensibility (Av/Ap) although also influenced by the dimensions of the vessel at the time of study. Because of the latter, PWV may be used only as an index of distensibility. Animal tissues obviously differ greatly in structure from elastic tubing but the relevance of PWV to arterial distensibility studies was confirmed by experimental work in animals which ''evealed a very close agreement between observed PWVs and values computed from vessel elastic "moduli and dimensions (Gow & Taylor, 1968; Patel, Janicki, & Carew, 1969). The fact that arteries and arterial segments become progressively more resistant to stretch, that is less distensible, at higher distending pressures, has been recognized for almost a century and subsequently confirmed in many studies (Roy, 1880; Bergel, 1961). Arterial PWV, as would be expected, also rises with distending pressure and a near linear relationship over a wide pressure range for excised artery was reported by Bramwell & Hill (1922) and more recently by Yoshimura, Sugai, Hashimoto, Okamura, Yamagishi, Hasegawa, Hayashi, Otsuka, Ishikawa, Yamashita, & Kozakai (1968). Measurements of PWV have also been "^ade in intact humans, the dependence of velocity on pressure being similar to that found with in vitro preparations, although PWV is substantially influenced by pathological cardiovascular conditions (Hemingway, McSwiney, & Allison, 1928; King,
Twenty-six patients were studied. Twelve were patients referred to the hospital because of high blood pressure, 3 of whotn were taking antihypertensive drugs at the time of study; and 14 were volunteers, members of the medical staffer students. Their ages ranged from 8 to 80 yfs and their resting mean arterial pressure (MAP) from 75 to 150 mm Hg. MAP was taken as diastolie pressure +'A pulse pressure. PWV was measured between two 'Pixie' strain gauge transducers (Endevco Ltd.) whieh were applied to the skin over the most prominent parts of the braehial and radial arteries of one arm; in the antecubital fossa and at the wrist. The eharaeteristics of these small transducers, which accurately reflect the intra-arterial pressure pulse contour, have been described by Van der Hoeven, Monchv. & Beneken (1973). In order to allow measurement of PWV in each subject over a range of arterial distending pressure.s, the arm under study and with the transducers in position was placed in an airtight perspex box within which controlled steady pressures of between - 8 0 and +80 mm Hg were achieved by a vacuum cleaner attached at one end. Arterial pressure was monitored in the opposite arm either by using an intra-arterial eannula connected to a strain gauge transducer (Consolidated Electrodynamics Ltd) in 9 subjects, or by repeated sphygmomanomctric readings in the others. Cuff blood pressures when eompared in the two arms were found to be virtually identical in every subject. On the basis of previous work, both positive and negative box pressures were assutned to be transmitted through the arm to the depth of the arterial wali (Ludbrook & Collins, 1967; Caro. Foley, & Sudlow, 1968). The arterial transmural pre.ssure (TMP) in the limb being studied could therefore be altered at will over a wide range Thus if MAP was 80 mm Hg and the box pressure - 4 0 mm Hg then the mean TMP would be 120 mm Hg. Each .subject was studied resting and supine, and after suitable amplification the pujse waves and if available intra-arterial pressure were recorded on magnetic tape (Thermionic Recorder). Records were made at rest and during induced changes in TMP which were made in increments and decrements of 10 mm
88
Vol. 13, No, 1
GRIBBIN, STEPTOE, AND SLEIGHT Subjects
Hg. In most subjects the TMP range attained was greater than that which occurs physiologically. PWy was later calculated for each level of TMP by selecting artifact free pairs of pulses and measuring the time interval between corresponding points 10% up the wave fronts. This point was chosen because of its high frequency content which renders it free of distorting influences in its passage along the artery (MacDonaU, I960; Gribdin, 1974). This measurement of transit time was made in at least 8 pairs of pulses at each pressure level and measurement of the distance between transducers then allowed calculation of the PWV. The relationship of PWV to mean TMP was examined using linear regression analysis. Three subjects were retested after a number of months so that retest validity of the correlations could be assessed.
IW,
JE,
HA
and
MP
Results The coefficients of correlation relating PWV to TMP are all .9 or greater and are significant at the .001 level, indicating that PWV increases in a linear fashion with TMP. There is no evidence of any decline in the validity of the correlations with increasing age or resting pressure, but the slopes of the relationship, represented by the regression coefficients, vary considerably between individuals (range 0.04 to 0.25). The data from 4 subjects with very diiferent levels of blood pressure and of different ages have been plotted in Fig. 1. The test/retest data for the subjects examined twice are shown in Table 1. Although the range of pressures differs on the 2 occasions for each subject, a comparison of the regression coefficients indicates that these are not significantly different on retest (Snedecor & Cochran, 1967). Discussion It is important to realize that in the experiment reported here, the arterial distending pressure was changed; the changes in arterial distending pressure (TMP) being limited to the brachial-radial artery of the limb studied. The data presented strongly support the hypothesis that changes in PWV can be used to detect and to follow changes in blood
IW
•
•
JE
= 0.95 * 0, 90
0.1!
HA
» *—
MP
"
''
0.09
0.97
0. 96
T M P i n m m H g
pressure. The wide range of regression coefficients makes clear that in any one subject an absolute change in arterial pressure cannot be deduced from a given change in PWV alone, although the linearity is such that monitoring PWV at two pressure levels only may be sufficient to establish the slope. Firm conclusions about the test/retest reliability of the measure cannot be drawn, however, from the limited amount of data available here, for despite the lack of a significant statistical difference in test/ retest coefficients the disparity between the actual values of these coefficients is substantial. Further study of this problem in a larger number of subjects is in progress. Other possible reasons for changes in propaga-
Comparison of regression coefficients of PWV on TMP in 3 suhjects tested twice
Subject AW
8 months
DJ
3 months
TS
4 months
Resting MAP 83 82 150 126 76 82
0.22
Fig. 1. PWV plotted against TMP for 4 subjects; r is the linear correlation coefficient andm is the slope of each regression line. Subject IW: age 27, resting MAP 75 mm Hg. Subject JE: age 44, resting MAP 150 mm Hg. Subject HA: age 62, resting MAP 87 mm Hg. Subject MP: age 80, resting MAP 107 min Hg.
TABLE 1
Intervai Between Tests
0.09
TMP range
Correlation Coefficient
Regression Coefficient
32-120 40-152 95-180 53-147 28- 95 76-153
.984 .939 .922 .974 .994 .946
.074 .094 .087 .102 .082 .125
F Value of Difference Between Slopes
, Fiin6) = 1.90 FiUlS) = 1.31 f(l/14) = 3./I4
January, 1976
PULSE WAVE VELOCITY
tion velocities during these experiments should also be considered, and in particular the possibility that some of the changes in arterial distensibility and hence PWV are caused by factors other than an increase in distending pressure, for example autonomic influences on vessel smooth muscle activities, perhaps initiated by fright or pain, or perhaps changes in wall properties brought about by the high positive and negative pressures achieved in the pressure box. Consideration of the smooth muscle tension over the measurement pathway is particularly pertinent to studies involving stress or biofeedback where alterations in autonomic tone cannot be neglected. The marked influence of sympathetic tone on peripheral resistance vessels is well recognized, but the magnitude of any such effect on the major vessels or conduit arteries of man is unknown, although animal experiments suggest that any change in arterial distensibility would be minimal (Aars, 1971; Gow, 1972). Changes in smooth muscle tension which alter the stiffness of arterial walls can be produced pharmacologically but this requires topical application of agents in doses far exceeding any natural autonomic influence (Dobrin & Rovick, 1969). As far as the present work is concerned, the results are unlikely to have been influenced by changes in autonomic tone, because the heart rates and blood pressures of all subjects studied did not differ significantly during the procedure. The effect of autonomic tone together with any intrinsic myogenic response to the box-induced pressure change was also excluded by rapid 'cut-off' studies reported elsewhere (Gribbin 1974). A further point for consideration is the effect of external pressure changes on intra-arterial pressure, ^fone have been reported during negative pressure application (Caro et al., 1968) but Ludbrook & Collins (1967) reported that positive pressure applied to a limb could influence intra-arterial pressure if allowed to approximate the diastolie value. Because of this finding box pressures were never raised to more than 20 mm Hg below the diastolie blood pressure.
89
It might be expected that PWV is frequency dependent due to the viscoelastic properties of arterial walls. If this were the case, changes in heart rate could affect PWV independently of blood pressure. Leroyd & Taylor (1966) and Gow (1970) have shown that the elastic and viscous moduli of arterial wall tissues are markedly frequencydependent, but any such effect would be negligible in the present study since the heart rate changes were extremely small. Moreover Yoshimura et al. (1968) demonstrated that pulse rates of 40 to 90 bpm did not affect pulse wave velocity, and manipulation of heart rates with atropine was similarly shown to be of no consequence (Lyon & Sands, 1925). An alternative method of monitoring PWV has been used and may be more suitable for psychophysiological experiments than the system employed here. The braehial transducer is dispensed with and the R wave of the EKG complex is taken as the central trigger for measurement. The time interval from the electrical depolarization of the heart to the radial pulse is thus measured (Weltman et al., 1964; Williams & Williams, 1965). Although lack of knowledge about the distance travelled by the pulse from the heart to the peripheral recording site precludes calculation of PWV, this method does have the advantage of requiring only a single peripheral transducer, so is more stable and less prone to movement artifact than the two transducer system (Thomas, 1965). Conclusions Our experiment suggests that the dependence of changes in PWV on alterations in blood pressure is a reliable phenomenon. As a technique for the continuous monitoring of pressure it has some advantages over other methods, in that it can be used to detect both fast pressure changes (from one beat to the next), and slow alterations occurring over several hours. It must be emphasized that its application is limited to situations where changes in blood pressure are of interest. PWV may have considerable value in psychophysiology for the monitoring of short term pressure changes.
REFERENCES s, H. Effects of altered stnooth muscle tone on aortic diameter and aortic baroreeeptor activity in anaesthetised rabbits. Circulation Research, \91\,28, 254-262. Beamer, V., & Shapiro, A. P. Response in cardiovascular testing. Pjj-c/ioOTmaric A/^djc/fif, 1973, i 5 , 112-120. Bergel, D. H. The static elastic properties of the arterial wall. Journal of Physiology, 1961, 156, 44.5-457. Bramweli,J. C..&Hill,A. V. The velocity of the puise wave in man. Proceedings of the Royal Societv. London, 1922, 9? 298-306. Brener, i. A general model of voluntary control applied to the phenomena of learned cardiovascular change. InP. A. Obrist.
A. H. Blaek, J. Brener, & L, V. DiCara (Eds.), Cardiovascular psychophysiology: Current issues in response mechanisms, biofefdhack and methodology. Chicago: Aldine Atherton, 1974. Pp. 365-391. Brod. J., Fencl, V., Hejl, Z., & Jirka, J. Circulatory changesunderlying blood pressure elevation during acute emotional stress in normotensive and hypertensive subjects. Clinieal Science, 1959,/S, 269-279. Caro, C. Ci., Foley, T. H., & Sudlow, M. F. Early effects of abrupt reduction of local pressure on the forearm and its circulation. yoHr«a/o/P/,v.Ho,'ogy, 1968, 194, 645-658. Dobrin, P. B., & Rovick, A. A. Influence of va.scular smooth
90
GRIBBIN, STEPTOE, AND SLEIGHT
muscle on contractile mechanics and elasticity of arteries. American Journal of Physiology, 1969.2/7, 1644-1651. Geddes, L. A. The direet and indirect measurement of blood pressure. Chicago: Year Book Medical Publi.shers, Inc. 1970. Gow, B. S. Viscoelastic properties of conduit arteries. In R. Reader (Ed.), Hypertensive mechanisms, American Heart Association Monograph. 1970, 32, 113-123. New York: American Heart Association. Gow, B . S . Influence of vascular smooth muscle on the viscoelastic properties of blood vessels. In D. H. Bergel (Ed.), Cardiovascular fluid dynamics. Vol. 2. London: Academic Press, 1972. Pp. 66-110. Gow, B. S., & & Taylor, M. G. Measurement of viscoelastic properties of arteries in the living dog. Circulation Research, 1968,25,111-122. Greatorex. C. A. Indirect methods of blood pressure measurement. In B. W. Watson (Ed.), IEEE Medical Electronics Monograph. LonAotw Peter Peregrinus. 1971. Pp. 1-29. Gribbin, B. A study of baroreflex function and arterial distensibility in normal and hypertensive man. M.D. Thesis, University of Dundee, 1974. Hemingway, A., McSwiney, B. A., & Allison, F. R. The extensibility of human arteries. Quarterly Journal of Medicine, 1928, 2/. 489-498. Karvonen, M. J., Telivvo, L. J., & Jarvinen, E. J. K. Sphygmoinanometer cuff size and the accuracy of indirect measurement of blood pressure. American Journal of Cardiology, 1964, / i . 688-693. Kho.s!a, T., & Lowe, C. R. Arterial pressure and arm circumference. British Journal of Preventive & Social Medicine, 196.5. 79, 159-163. King, D.,Ceghlan, B., Gosling, R., Pickup, A., Newman, D., & Woodcock, J. Transcutaneous measurement of pulse wave velocity and mean blood pressure in man. inC. Roberts (Ed.), Blood flow measurement. London: Sector, 1972. Pp. 40-43. Landowne, M. A method using induced waves to study pressure propagation in human arteries. Circulation Research, 1957,5, 594-601. Lategola, M. T., Harrison, H., & Barnard, C. Continuous measurement of human blood pressure transients without puncture. Aerospace Medicine. 1966, J 7 . 228-233. Leroyd, B. M., & Taylor, M. G. Alterations with age in the viscoeleastic properties of human arterial walls. Circulation Research, 1966, 18, 278-292. Ludbrook, J., & Collins, G. M. Venous occlusion pressure plethysmography in the human upper limb. Circulation Research, 1967, 2 / . 139-147.
Vol. 13, No. I
Lyon, W., & Sands, J. Studies in pulse wave velocity. 4: Effects of adrenaline on pulse wave velocity. American Journal of Physiology, 1925. 71, 534-542. MacDonald, D. A. Blood flow in arteries. London: Arnold, 1960. Patel, D. J., Janicki, J. S., & Carew, T. E. Static anisotropic elastic properties of the aorta in living dogs. Circulation Research, 1969. 25, 765-779. Ragan, C., & Bardley, J. The accuracy of clinical measurements of arterial blood pressure. Bulletin ofJohns Hopkins Hospital. 1941,69, 504-528. Roberts, L. N., Smiley, R. A.. & Manning, G. W. A comparison of direct and indirect blood pressure determinations.
Circulation, 1953,5, 232. Roy, C. S. The elastic properties of the arterial wall. Journal of Physiology, 1880, i , 125. Shapiro, D., Schwartz, G. E., & Tursky, B. Control of diastolic blood pressure in man by feedback and reinforcement. Psychophysiotogy, 1972, 9. 296. Snedecor, G. W., &Cochran, W. G. Statistical methods. Ames: Iowa State University Press, 1967. Thomas, J. G. Continuous pulse wave velocity recording for indirectly monitoring blood pressure in man. Medical Electronics & Biological Engineering, 1965, 3, 331-332. Tursky, B. The indirect recording of human blood pressure. In P. A. Obrist, A. H. Black, J. Brener, & L. DiCara (Eds.), Cardiovascular psychophysiology: Current issues in response mechanisms, hiofeedback and methodology. Chicago: Aidine Atherton. 1974. Pp. 93-105. Tursky, B., Shapiro, D.. & Schwartz, G. E. Automated constant cuff-pressure system to measure average systolic and diastolic blood pressure in man. IEEE Transactions on Biomedical Engineering, 1972, 19, 271-276. Van der Hoeven. G. M. A., de Monchy, C , & Beneken, J. E. W. Systolic time intervals and pulse wave transit time in normal children. Bm(.vft//€arr/o«™a/, 1973. i 5 , 669-678. Weltman, G., Sullivan, G., & Brendon, D. The continuous measurement of arterial pulse wave velocity. Medical Electronics & Biological Engineering, 1964, 2, 145-154. Williams, J. G. L., & Williams, B. Arterial pulse wave velocity as a psychophysiological measure. Psychosomatic Medicine, 1965.27. 408-^14. Yoshimura. S., Sugai, J., Hashimoto, H., Okamura, T., Yamagishi, N., Hasegawa, M., Hayashi, T., Otsuka, E., Ishikawa, E., Yamashita, H., & Kozakai, M. The estimation of arteriosclerosis by the measurement of pulse wave velocity. CorVasa, 1968, 10, 173-182.
(Manuscript received April 10, 1974; revision received May 15. 1975; accepted for publication September 9, 1975)
PsYfHOPHYSIOl.OGY
Copyrighf '' i76 by The Sociely for Psychciphysioldgical Research
Methodology Pulse Wave Velocity as a Measure of Blood Pressure Change BRIAN GRIBBIN, Department of Cardiovascular Medicine, Radclife Infirmary, Oxford ANDREW STEPTOE, Department of Psychiatry, Oxford University AND PETER S L E I G H T
Department of Cardiovascular Medicine, Radclife Infirmary, Oxford
ABSTRACT The use of arterial pulse wave velocity (1*WV) as a continuous measure of blood pressure changes is outiined. Theoretical considerations indicate that changes in PWV reflect changes in blood pressure, and an experiment was carried out to assess this relationship. PWV along an arm artery was monitored in 26 subjects at a time when the arterial distending pressure of the limb was altered over a wide range by means of externally applied positive and negative pressures. The results show that changes in PWV reliably follow changes in blood pressure. This method can be considered suitable for studies requiring changes rather than absolute values of blood pressure. DESCRIPTORS: Pulse wave velocity, Blood pressure. Methodology. A major obstacle to the fuller understanding of cardiovascular psychophysiology has been the difficulty of continuous blood pressure measurement. Direct pressure monitoring by arterial cannulation is seldom used, even though the procedure may not be as stressful as is frequently supposed (Beamer & Shapiro, 1973). In recent years, hovs'ever, a number of indirect methods have been developed using sophisticated adaptations of the sphygmomanometric technique (Tursky, 1974). For example one system uses an arm cutF which is automatically inflated to a preset level near the systolic or diastolic pressure (Tursfcy, Shapiro, & Schwartz, 1972). The presence or absence of Korotkoff sounds is detected by a microphone in the cuff, and indicates whether pressure is above or below the preset level. This system thus provides information on every heart beat, but this is of a binary nature only since there is no indication of the size of any deviation from inflation pressure. Another disadvantage is that continuous monitoring Research was partially supported by the Medical Research Council. U.K., and the work carried out by Dr. Gribbin took place during his tenure of a Beit Memorial Junior Research Fellowship. Address requests for reprints to: Brian Gribbin, Department of Cardiovascular Medicine, Radcliffe Infirmary, Oxford, U.K.
is only possible for short periods (50 heart beats in Shapiro, Schwartz, & Tursky, 1972), since pressure must be released to allow blood flow into the arm, and so readings are available for only about 70% of the time. Another approach to this problem is reported by Lategola, Harrison, & Barnard (1966) and Brener (1974) who used a technique in ; which an arm cuif is rapidly inflated to near previous levels of pressure and then the system is switched to: a slow inflation rate until a Korotkoff sound is detected. This enables the diastolic blood pressure to be tracked, but of course readings are limited to every third or fourth heart beat. Neither of these techniques is thus ideal since they do not provide immediate analogue information on every heart beat. A number of problems common to occlusion methods can also be mentioned. For instance, there are inherent inaccuracies in auscultatory blood pressure recordings when compared with direct measurements (Geddes, 1970); systolic pressures tend to be low while diastolic values are commonly too high, although the latter depend on whether muffling or disappearance of Korotkoff sounds is used as the criterion (Ragan & Bardley, 1941; Roberts, Smiley, & Manning, 1953). The problem of accurate detection of the diastolic pressure appears to be more pronounced with automatic de86
January, 1976
PULSE WAVE VELOCITY
tection techniques, while incorrect cuff size and length can distort pressure readings and increase variability (Greatorex, 1971; Khosla & Lowe, 1965; Karvonen, Telivvo, & Jarvinen, 1964). The measurement of Pulse Wave Velocity (PWV) may provide an attractive alternative to the automated pneumatic methods, since immediate analogue readings can be made on every beat. PWV is the rate of propagation of pressure pulse waves along arteries, and short term changes in this parameter are due primarily to alterations in blood pressure. The velocity with which a pressure wave travels along an artery is determined by the mechanical properties of that artery, and the relationship between the two has been described by equations derived from bench experiments in which thin-walled elastic tubing was substituted for artery. One such equation was reported by Bramwell and Hill (1922): '^
Av
p
87
Coghlan, Gosling, Pickup. Newman, & Woodcock, 1972). Pressure waves used in the calculation of PWV have been recorded through intra-arterial needles or catheters and also by the use of external transducers accurately placed over the artery (Landowne, 1957; Gribbin, 1974). A modification of this method is to record the time delay between the R wave of the EKG and the upstroke of a single peripherally recorded pulse wave (Weltman, Sullivan, & Brendon, 1964). PWV values derived from such experiments have been found to increase during mental arithmetic and are considered likely to parallel changes in arterial pressure (Brod, Fend, Hejl, & Jirka, 1959; Williams & Williams, 1965). It is the purpose of this paper to examine the relationship of PWV to changes in arterial pressure within individuals, as opposed to the betweensubject comparisons more commonly made. The application of PWV monitoring to psychophysio^ logical experiments will also be considered. Methods
where c = PWV metres/sec, Ap = change in pressure, Av == change in volume, V = initial Volume, and p = density of fluid. One can observe from this equation that PWV is inversely related to true compliance or distensibility (Av/Ap) although also influenced by the dimensions of the vessel at the time of study. Because of the latter, PWV may be used only as an index of distensibility. Animal tissues obviously differ greatly in structure from elastic tubing but the relevance of PWV to arterial distensibility studies was confirmed by experimental work in animals which ''evealed a very close agreement between observed PWVs and values computed from vessel elastic "moduli and dimensions (Gow & Taylor, 1968; Patel, Janicki, & Carew, 1969). The fact that arteries and arterial segments become progressively more resistant to stretch, that is less distensible, at higher distending pressures, has been recognized for almost a century and subsequently confirmed in many studies (Roy, 1880; Bergel, 1961). Arterial PWV, as would be expected, also rises with distending pressure and a near linear relationship over a wide pressure range for excised artery was reported by Bramwell & Hill (1922) and more recently by Yoshimura, Sugai, Hashimoto, Okamura, Yamagishi, Hasegawa, Hayashi, Otsuka, Ishikawa, Yamashita, & Kozakai (1968). Measurements of PWV have also been "^ade in intact humans, the dependence of velocity on pressure being similar to that found with in vitro preparations, although PWV is substantially influenced by pathological cardiovascular conditions (Hemingway, McSwiney, & Allison, 1928; King,
Twenty-six patients were studied. Twelve were patients referred to the hospital because of high blood pressure, 3 of whotn were taking antihypertensive drugs at the time of study; and 14 were volunteers, members of the medical staffer students. Their ages ranged from 8 to 80 yfs and their resting mean arterial pressure (MAP) from 75 to 150 mm Hg. MAP was taken as diastolie pressure +'A pulse pressure. PWV was measured between two 'Pixie' strain gauge transducers (Endevco Ltd.) whieh were applied to the skin over the most prominent parts of the braehial and radial arteries of one arm; in the antecubital fossa and at the wrist. The eharaeteristics of these small transducers, which accurately reflect the intra-arterial pressure pulse contour, have been described by Van der Hoeven, Monchv. & Beneken (1973). In order to allow measurement of PWV in each subject over a range of arterial distending pressure.s, the arm under study and with the transducers in position was placed in an airtight perspex box within which controlled steady pressures of between - 8 0 and +80 mm Hg were achieved by a vacuum cleaner attached at one end. Arterial pressure was monitored in the opposite arm either by using an intra-arterial eannula connected to a strain gauge transducer (Consolidated Electrodynamics Ltd) in 9 subjects, or by repeated sphygmomanomctric readings in the others. Cuff blood pressures when eompared in the two arms were found to be virtually identical in every subject. On the basis of previous work, both positive and negative box pressures were assutned to be transmitted through the arm to the depth of the arterial wali (Ludbrook & Collins, 1967; Caro. Foley, & Sudlow, 1968). The arterial transmural pre.ssure (TMP) in the limb being studied could therefore be altered at will over a wide range Thus if MAP was 80 mm Hg and the box pressure - 4 0 mm Hg then the mean TMP would be 120 mm Hg. Each .subject was studied resting and supine, and after suitable amplification the pujse waves and if available intra-arterial pressure were recorded on magnetic tape (Thermionic Recorder). Records were made at rest and during induced changes in TMP which were made in increments and decrements of 10 mm
88
Vol. 13, No, 1
GRIBBIN, STEPTOE, AND SLEIGHT Subjects
Hg. In most subjects the TMP range attained was greater than that which occurs physiologically. PWy was later calculated for each level of TMP by selecting artifact free pairs of pulses and measuring the time interval between corresponding points 10% up the wave fronts. This point was chosen because of its high frequency content which renders it free of distorting influences in its passage along the artery (MacDonaU, I960; Gribdin, 1974). This measurement of transit time was made in at least 8 pairs of pulses at each pressure level and measurement of the distance between transducers then allowed calculation of the PWV. The relationship of PWV to mean TMP was examined using linear regression analysis. Three subjects were retested after a number of months so that retest validity of the correlations could be assessed.
IW,
JE,
HA
and
MP
Results The coefficients of correlation relating PWV to TMP are all .9 or greater and are significant at the .001 level, indicating that PWV increases in a linear fashion with TMP. There is no evidence of any decline in the validity of the correlations with increasing age or resting pressure, but the slopes of the relationship, represented by the regression coefficients, vary considerably between individuals (range 0.04 to 0.25). The data from 4 subjects with very diiferent levels of blood pressure and of different ages have been plotted in Fig. 1. The test/retest data for the subjects examined twice are shown in Table 1. Although the range of pressures differs on the 2 occasions for each subject, a comparison of the regression coefficients indicates that these are not significantly different on retest (Snedecor & Cochran, 1967). Discussion It is important to realize that in the experiment reported here, the arterial distending pressure was changed; the changes in arterial distending pressure (TMP) being limited to the brachial-radial artery of the limb studied. The data presented strongly support the hypothesis that changes in PWV can be used to detect and to follow changes in blood
IW
•
•
JE
= 0.95 * 0, 90
0.1!
HA
» *—
MP
"
''
0.09
0.97
0. 96
T M P i n m m H g
pressure. The wide range of regression coefficients makes clear that in any one subject an absolute change in arterial pressure cannot be deduced from a given change in PWV alone, although the linearity is such that monitoring PWV at two pressure levels only may be sufficient to establish the slope. Firm conclusions about the test/retest reliability of the measure cannot be drawn, however, from the limited amount of data available here, for despite the lack of a significant statistical difference in test/ retest coefficients the disparity between the actual values of these coefficients is substantial. Further study of this problem in a larger number of subjects is in progress. Other possible reasons for changes in propaga-
Comparison of regression coefficients of PWV on TMP in 3 suhjects tested twice
Subject AW
8 months
DJ
3 months
TS
4 months
Resting MAP 83 82 150 126 76 82
0.22
Fig. 1. PWV plotted against TMP for 4 subjects; r is the linear correlation coefficient andm is the slope of each regression line. Subject IW: age 27, resting MAP 75 mm Hg. Subject JE: age 44, resting MAP 150 mm Hg. Subject HA: age 62, resting MAP 87 mm Hg. Subject MP: age 80, resting MAP 107 min Hg.
TABLE 1
Intervai Between Tests
0.09
TMP range
Correlation Coefficient
Regression Coefficient
32-120 40-152 95-180 53-147 28- 95 76-153
.984 .939 .922 .974 .994 .946
.074 .094 .087 .102 .082 .125
F Value of Difference Between Slopes
, Fiin6) = 1.90 FiUlS) = 1.31 f(l/14) = 3./I4
January, 1976
PULSE WAVE VELOCITY
tion velocities during these experiments should also be considered, and in particular the possibility that some of the changes in arterial distensibility and hence PWV are caused by factors other than an increase in distending pressure, for example autonomic influences on vessel smooth muscle activities, perhaps initiated by fright or pain, or perhaps changes in wall properties brought about by the high positive and negative pressures achieved in the pressure box. Consideration of the smooth muscle tension over the measurement pathway is particularly pertinent to studies involving stress or biofeedback where alterations in autonomic tone cannot be neglected. The marked influence of sympathetic tone on peripheral resistance vessels is well recognized, but the magnitude of any such effect on the major vessels or conduit arteries of man is unknown, although animal experiments suggest that any change in arterial distensibility would be minimal (Aars, 1971; Gow, 1972). Changes in smooth muscle tension which alter the stiffness of arterial walls can be produced pharmacologically but this requires topical application of agents in doses far exceeding any natural autonomic influence (Dobrin & Rovick, 1969). As far as the present work is concerned, the results are unlikely to have been influenced by changes in autonomic tone, because the heart rates and blood pressures of all subjects studied did not differ significantly during the procedure. The effect of autonomic tone together with any intrinsic myogenic response to the box-induced pressure change was also excluded by rapid 'cut-off' studies reported elsewhere (Gribbin 1974). A further point for consideration is the effect of external pressure changes on intra-arterial pressure, ^fone have been reported during negative pressure application (Caro et al., 1968) but Ludbrook & Collins (1967) reported that positive pressure applied to a limb could influence intra-arterial pressure if allowed to approximate the diastolie value. Because of this finding box pressures were never raised to more than 20 mm Hg below the diastolie blood pressure.
89
It might be expected that PWV is frequency dependent due to the viscoelastic properties of arterial walls. If this were the case, changes in heart rate could affect PWV independently of blood pressure. Leroyd & Taylor (1966) and Gow (1970) have shown that the elastic and viscous moduli of arterial wall tissues are markedly frequencydependent, but any such effect would be negligible in the present study since the heart rate changes were extremely small. Moreover Yoshimura et al. (1968) demonstrated that pulse rates of 40 to 90 bpm did not affect pulse wave velocity, and manipulation of heart rates with atropine was similarly shown to be of no consequence (Lyon & Sands, 1925). An alternative method of monitoring PWV has been used and may be more suitable for psychophysiological experiments than the system employed here. The braehial transducer is dispensed with and the R wave of the EKG complex is taken as the central trigger for measurement. The time interval from the electrical depolarization of the heart to the radial pulse is thus measured (Weltman et al., 1964; Williams & Williams, 1965). Although lack of knowledge about the distance travelled by the pulse from the heart to the peripheral recording site precludes calculation of PWV, this method does have the advantage of requiring only a single peripheral transducer, so is more stable and less prone to movement artifact than the two transducer system (Thomas, 1965). Conclusions Our experiment suggests that the dependence of changes in PWV on alterations in blood pressure is a reliable phenomenon. As a technique for the continuous monitoring of pressure it has some advantages over other methods, in that it can be used to detect both fast pressure changes (from one beat to the next), and slow alterations occurring over several hours. It must be emphasized that its application is limited to situations where changes in blood pressure are of interest. PWV may have considerable value in psychophysiology for the monitoring of short term pressure changes.
REFERENCES s, H. Effects of altered stnooth muscle tone on aortic diameter and aortic baroreeeptor activity in anaesthetised rabbits. Circulation Research, \91\,28, 254-262. Beamer, V., & Shapiro, A. P. Response in cardiovascular testing. Pjj-c/ioOTmaric A/^djc/fif, 1973, i 5 , 112-120. Bergel, D. H. The static elastic properties of the arterial wall. Journal of Physiology, 1961, 156, 44.5-457. Bramweli,J. C..&Hill,A. V. The velocity of the puise wave in man. Proceedings of the Royal Societv. London, 1922, 9? 298-306. Brener, i. A general model of voluntary control applied to the phenomena of learned cardiovascular change. InP. A. Obrist.
A. H. Blaek, J. Brener, & L, V. DiCara (Eds.), Cardiovascular psychophysiology: Current issues in response mechanisms, biofefdhack and methodology. Chicago: Aldine Atherton, 1974. Pp. 365-391. Brod. J., Fencl, V., Hejl, Z., & Jirka, J. Circulatory changesunderlying blood pressure elevation during acute emotional stress in normotensive and hypertensive subjects. Clinieal Science, 1959,/S, 269-279. Caro, C. Ci., Foley, T. H., & Sudlow, M. F. Early effects of abrupt reduction of local pressure on the forearm and its circulation. yoHr«a/o/P/,v.Ho,'ogy, 1968, 194, 645-658. Dobrin, P. B., & Rovick, A. A. Influence of va.scular smooth
90
GRIBBIN, STEPTOE, AND SLEIGHT
muscle on contractile mechanics and elasticity of arteries. American Journal of Physiology, 1969.2/7, 1644-1651. Geddes, L. A. The direet and indirect measurement of blood pressure. Chicago: Year Book Medical Publi.shers, Inc. 1970. Gow, B. S. Viscoelastic properties of conduit arteries. In R. Reader (Ed.), Hypertensive mechanisms, American Heart Association Monograph. 1970, 32, 113-123. New York: American Heart Association. Gow, B . S . Influence of vascular smooth muscle on the viscoelastic properties of blood vessels. In D. H. Bergel (Ed.), Cardiovascular fluid dynamics. Vol. 2. London: Academic Press, 1972. Pp. 66-110. Gow, B. S., & & Taylor, M. G. Measurement of viscoelastic properties of arteries in the living dog. Circulation Research, 1968,25,111-122. Greatorex. C. A. Indirect methods of blood pressure measurement. In B. W. Watson (Ed.), IEEE Medical Electronics Monograph. LonAotw Peter Peregrinus. 1971. Pp. 1-29. Gribbin, B. A study of baroreflex function and arterial distensibility in normal and hypertensive man. M.D. Thesis, University of Dundee, 1974. Hemingway, A., McSwiney, B. A., & Allison, F. R. The extensibility of human arteries. Quarterly Journal of Medicine, 1928, 2/. 489-498. Karvonen, M. J., Telivvo, L. J., & Jarvinen, E. J. K. Sphygmoinanometer cuff size and the accuracy of indirect measurement of blood pressure. American Journal of Cardiology, 1964, / i . 688-693. Kho.s!a, T., & Lowe, C. R. Arterial pressure and arm circumference. British Journal of Preventive & Social Medicine, 196.5. 79, 159-163. King, D.,Ceghlan, B., Gosling, R., Pickup, A., Newman, D., & Woodcock, J. Transcutaneous measurement of pulse wave velocity and mean blood pressure in man. inC. Roberts (Ed.), Blood flow measurement. London: Sector, 1972. Pp. 40-43. Landowne, M. A method using induced waves to study pressure propagation in human arteries. Circulation Research, 1957,5, 594-601. Lategola, M. T., Harrison, H., & Barnard, C. Continuous measurement of human blood pressure transients without puncture. Aerospace Medicine. 1966, J 7 . 228-233. Leroyd, B. M., & Taylor, M. G. Alterations with age in the viscoeleastic properties of human arterial walls. Circulation Research, 1966, 18, 278-292. Ludbrook, J., & Collins, G. M. Venous occlusion pressure plethysmography in the human upper limb. Circulation Research, 1967, 2 / . 139-147.
Vol. 13, No. I
Lyon, W., & Sands, J. Studies in pulse wave velocity. 4: Effects of adrenaline on pulse wave velocity. American Journal of Physiology, 1925. 71, 534-542. MacDonald, D. A. Blood flow in arteries. London: Arnold, 1960. Patel, D. J., Janicki, J. S., & Carew, T. E. Static anisotropic elastic properties of the aorta in living dogs. Circulation Research, 1969. 25, 765-779. Ragan, C., & Bardley, J. The accuracy of clinical measurements of arterial blood pressure. Bulletin ofJohns Hopkins Hospital. 1941,69, 504-528. Roberts, L. N., Smiley, R. A.. & Manning, G. W. A comparison of direct and indirect blood pressure determinations.
Circulation, 1953,5, 232. Roy, C. S. The elastic properties of the arterial wall. Journal of Physiology, 1880, i , 125. Shapiro, D., Schwartz, G. E., & Tursky, B. Control of diastolic blood pressure in man by feedback and reinforcement. Psychophysiotogy, 1972, 9. 296. Snedecor, G. W., &Cochran, W. G. Statistical methods. Ames: Iowa State University Press, 1967. Thomas, J. G. Continuous pulse wave velocity recording for indirectly monitoring blood pressure in man. Medical Electronics & Biological Engineering, 1965, 3, 331-332. Tursky, B. The indirect recording of human blood pressure. In P. A. Obrist, A. H. Black, J. Brener, & L. DiCara (Eds.), Cardiovascular psychophysiology: Current issues in response mechanisms, hiofeedback and methodology. Chicago: Aidine Atherton. 1974. Pp. 93-105. Tursky, B., Shapiro, D.. & Schwartz, G. E. Automated constant cuff-pressure system to measure average systolic and diastolic blood pressure in man. IEEE Transactions on Biomedical Engineering, 1972, 19, 271-276. Van der Hoeven. G. M. A., de Monchy, C , & Beneken, J. E. W. Systolic time intervals and pulse wave transit time in normal children. Bm(.vft//€arr/o«™a/, 1973. i 5 , 669-678. Weltman, G., Sullivan, G., & Brendon, D. The continuous measurement of arterial pulse wave velocity. Medical Electronics & Biological Engineering, 1964, 2, 145-154. Williams, J. G. L., & Williams, B. Arterial pulse wave velocity as a psychophysiological measure. Psychosomatic Medicine, 1965.27. 408-^14. Yoshimura. S., Sugai, J., Hashimoto, H., Okamura, T., Yamagishi, N., Hasegawa, M., Hayashi, T., Otsuka, E., Ishikawa, E., Yamashita, H., & Kozakai, M. The estimation of arteriosclerosis by the measurement of pulse wave velocity. CorVasa, 1968, 10, 173-182.
(Manuscript received April 10, 1974; revision received May 15. 1975; accepted for publication September 9, 1975)
