Office Blood Pressures, Arterial Compliance Characteristics, and Estimated Cardiac Load
Because of rising interest in new methods to detect arterial diseases, we compared data from 3 different compliance-related techniques to measure arterial stiffness: systolic pulse contour analysis, diastolic pulse contour analysis (modified Windkessel model), and muscular (brachial) artery compliance by cuff plethysmography. Variables measured in the sitting position were compared with each other, with clinic blood pressures (BPs), and with the cardiac time-tension integral (CTTI) in 63 established hypertensive and 28 age-matched normotensive subjects. Hypertensives demonstrated marginal reductions in C1 (thought to represent reduced large vessel compliance) and increased central systolic BP augmentation. In contrast, muscular artery compliance tended to be greater in the hypertensives despite normal brachial arterial diameters. C2, suggested to be an indicator of small artery properties, was similar in both groups. CTTI was strongly related to systolic pressure (r=0.81), integrated mean arterial pressure (r=0.83), and systolic pressure-heart rate product (r=0.85) and was less strongly related to diastolic (r=0.71) or pulse pressure (r=0.57). Weak correlations were observed between CTTI and measured compliance-related variables, which also showed absent or weak correlations among themselves. We conclude that the weak relationships among BP and compliance-related variables could be due to intrinsic differences in the properties of large and small arteries, theoretical methodological weaknesses, measurement artifacts, or intrinsic hemodynamic differences of the sitting position. At present, compliance-related variables provide little additional advantage over cuff BP in the office estimation of cardiac work.
Systolic blood pressure (BP) is strongly related to age,1 and systolic hypertension in older individuals occurs primarily as a result of increased stiffness of large arteries.2,3 There is now recognition that systolic BP is a more robust predictor of overall cardiovascular morbidity and mortality than is diastolic BP4,5 and that systolic pressure is a critical indicator of health in older people.6,7 The assumption that systolic hypertension and widened pulse pressure are late manifestations of arterial stiffening2 has prompted the development of new techniques to allow more precise estimation of vascular biomechanical properties earlier in the course of arterial disease.
To date, no large studies have been performed to directly compare simultaneous measurements of BPs and compliance-related variables derived from current techniques. This present work reports early results investigating the relationships among office BPs and compliance-related variables derived from systolic pulse contour analysis, diastolic pulse contour analysis by modified Windkessel model, and cuff plethysmographic estimation of brachial artery compliance (BAC) and wall tension (BWT). Issues of specific interest in this survey analysis included potential differences between normotensive and hypertensive groups, potential differences among the various techniques, relationships between compliance-related variables and office BPs, and the utility of any of the variables as estimates of cardiac load.
All subjects were studied after they gave informed consent. Normotensives reported no occurrence of BP elevations at any time. Hypertensives were well characterized, having had systolic BP values >140 mm Hg or diastolic BP values >90 mm Hg on ≥3 occasions previously; and none received antihypertensive medications for the month prior to study. All analyses occurred during the morning hours. All measurements were obtained after subjects had been sitting comfortably in a chair for ≥5 minutes, with the back supported and feet on the floor.
Office BP Measurements
Observers certified for the measurement of cuff pressures obtained standard cuff BPs by auscultation. Phase I and V Korotkoff sounds were recorded either 2 or 3 times after subjects had been seated quietly for ≥5 minutes. BP was deemed to be stable if values were within 10 mm Hg. The mean of the final 2 values was recorded. Simultaneous heart rates were also obtained.
Systolic Pulse Contour Analysis
Subjects were then taken to the clinical laboratory and again allowed to sit undisturbed for ≥5 minutes. Systolic pulse contour recordings were obtained using the Sphygmocor BP Analysis System, which has been described elsewhere.6,8 In brief, a tonometer is used to record high-fidelity radial arterial pulse contour tracings. These values are then transformed using a proprietary transfer function to estimate central arterial waveforms. Central augmentation pressure (CAP) was defined as the late systolic increment caused by pulse wave reflection; central augmentation index (CAI), as CAP normalized to central pulse pressure. Cardiac time-tension index (CTTI) is the integral of the central pulse waveform beginning at the pulse upstroke and ending at the dicrotic notch. Mean arterial pressure was also derived from the central waveform using the mean of the integral of pressure over the entire cardiac cycle. Readings were obtained from ensemble averages of waveforms collected for 11 seconds, in which the coefficients of variation of the diastolic pressure was <10%.
Diastolic Pulse Contour Analysis
Pulse contours derived from the radial arterial waveform were also used to calculate 2 separate compliance variables according to a published model.9–11 This modified Windkessel system involves 4 model parameters: C1, C2, L, and R. The distal compliance C2 is in parallel with resistance R. The systemic circulation is represented by this pair in series with an inertance L. The circulation is in parallel to a central compliance C1. Curve-fit coefficients are used to calculate the model parameters.10,11
The data used to obtain C1 and C2 as reported here were ensemble-averaged signals from a radial tonometer (Sphygmocor) digitized at 128 Hz. The model was fitted to the waveforms using Marquardt’s algorithm (SigmaStat, SPSS Science). This method was tested against a proprietary commercial system (HDI Inc), with good concordance: in 15 subjects with systolic BPs ranging from 105 to 156 mm Hg and diastolic pressures ranging from 61 to 100 mm Hg, mean±SD values for C1 and C2 were 160±50 and 6±3 mL/mmHg, respectively, for our method versus 170±60 and 6±2 mL/mmHg for the HDI system. Coefficients of variation of C1 and C2 were <35% for our calculations. They are not reported for the proprietary HDI system.
BAC, Brachial Arterial Diameter, and BWT
BAC and BWT estimates were derived using a cuff plethysmographic technique (CuffSoft Inc), which has been described and validated on a mechanical model elsewhere.12,13 In brief, this system depends on a volume-calibrated oscillometric cuff to explore the range of pressure-area relationships of the brachial artery from occlusion to operating pressure. Incremental pressure-volume relationships are determined for each heartbeat as the cuff is deflated and both brachial arterial diameters (BDs) and distending pressures are measured. Curve fitting allows the calculation of volume-pressure (compliance) relationships, which can be reported at systolic, diastolic, or mean pressures. BWT can also be calculated using Laplace’s law, with units expressed in mm Hg · cm. In the present study, BAC and BWT values were reported at mean arterial pressure.
Values are expressed as mean±SD. Intermethod comparisons were effected using standard Pearson correlation coefficients. Statistical significance was accepted at the 0.05 level.
Results from 91 subjects (43 men and 48 women) were included. In all, 28 subjects were normotensive (<140/90 mm Hg), and 63 were hypertensive, as noted in Table 1. Systolic BPs ranged from 88 to 200 mm Hg; diastolic pressures, from 60 to 120 mm Hg. The age range was 24 to 81 years. The means of BPs and each of the compliance-related variables for the normotensive and hypertensive populations are reported in Table 1. Hypertensive patients showed significantly higher systolic, diastolic, mean, and pulse pressures, as well as higher CAI, CTTI and BWT, compared with those of normotensive subjects (P<0.05). In this population, there was no significant difference between hypertensive and normotensive subjects in BD at mean pressure, BAC at mean pressure, C1, or C2.
Interrelationships among clinic BPs and the derived compliance variables are reported in Table 2. Systolic pressure was strongly correlated with CAP, CAI, C1, and BWT but was not correlated with C2, whereas diastolic pressure was strongly correlated with CAP, CAI, and BWT, only weakly correlated with C2, and not correlated with C1. Pulse pressure was moderately predictive of CAP, C1, and CAI and was only weakly related to BWT but not to C2.
Table 3 is a matrix of Pearson correlation coefficients representing the interrelationships among the derived compliance variables. Correlations were absent or weak among variables derived from the different methods. C2 was only marginally related to C1.
Table 4 includes the Pearson correlation coefficients among CTTI and BPs and compliance variables. CTTI, which was derived from the integral of the estimated central pressure waveform, is representative of cardiac workload and was closely related to systolic pressure. Because of the relationship between age and systolic pressure, the whole group was dichotomized by age: the younger group (age: 24 to 50 years, n=39) demonstrated a correlation coefficient between CTTI and systolic BP of 0.85 (r2=0.71), whereas the older group (age 52 to 81 years, n=52) demonstrated a corresponding CTTI to SP correlation coefficient of 0.79 (r2=0.60). Thus, the correlation of CTTI and systolic pressure was age-independent. Diastolic pressures correlated substantially less well with CTTI in the older group (r=0.65, r2=0.42) than in the younger group (r=0.79, r2=0.60). Using all available variables in a stepwise forward regression model for the whole group, a marginal increase in the predictability of CTTI was obtained; a maximal multiple-r value of 0.92 occurred with the equation CTTI=[(0.57×SPHR)+(0.25×SP)+(0.22×DP)], where SPHR=systolic BP×heart rate.
Compared with normotensive controls, established hypertensive subjects, who were studied while they were seated comfortably, demonstrated higher brachial BP, higher CAP and CAI, increased BWT, and increased cardiac load as measured by CTTI. Sitting heart rate did not differ between groups, and there was no difference in C1, C2, or internal BDs. There was a trend toward increased sitting BAC (at a higher mean pressure) in the hypertensive group, but the observed data were much more scattered in the hypertensive population than in the normotensives.
Present results differ somewhat from previous reports that have described decreased C2 in supine hypertensive subjects.14 A major difference between studies could be our use of sitting compliance measurements, which we deemed to be more appropriate than supine values because we used sitting BP values. In most studies, supine hemodynamics have been reported, but it should not be assumed that supine and sitting values are interchangeable. In the supine position, venous return is passive and not rate-limiting, whereas in the sitting position, there is reduced venous return, lower cardiac output, and increased systemic vascular resistance. Because C2 depends so closely on systemic resistance,11 it can be inferred that systemic resistance values in the sitting position were similar between normotensives and hypertensives. It is also conceivable that present results were influenced by prior antihypertensive therapy and that only those who had never received treatment should have been studied. However, the available number of never-treated hypertensives was very small, and it would probably have been unethical to withhold medications for a sufficient period of time to allow all chronic hypertensive vascular changes to become manifested. Most hypertensives in the present study had been off medications only ≈4 weeks before study. ACE inhibitors have been found to cause a greater increase in carotid distensibility than do β-blockers.15 Yet the most important influence on any compliance-related variable is known to be the arterial pressure at the time of measurement.
Present results demonstrate only low-grade relationships among compliance-related parameters derived from systolic and diastolic pulse contour analysis or cuff plethysmography. The reasons for these poor intermethod correlations are not yet fully clear, but it is easily possible that each method describes different properties of different parts of the peripheral circulation. It would be expected, for example, that the compliance-related properties of large arteries are different from those of muscular arteries or smaller arterioles because of their intrinsic geometric differences. Thus, the lack of a strong correlation between C1 and C2, or between BWT and C2 may reflect expected biological and functional differences. It is also clear that variables derived from pulse wave contour analysis represent aggregate properties of heterogeneous vascular beds. These whole-body estimates have unclear biological significance, and the poor intermethod correlations may exist because real regional inhomogeneities are lost in a regression-to-the-mean effect. The potential for significant measurement artifacts and the absence of a gold standard for arterial compliance are further impediments to confident widespread clinical application of current techniques.
Newer studies defining relationships between compliance-related variables and cardiovascular disease outcomes offer hope that a clinical rationale for the measurement of arterial compliance is justified. Increased pulse wave velocity is a predictor of cardiovascular morbidity and mortality,16,17 but augmentation pressure demonstrates no significant relationship with atherosclerotic outcomes,17 and no large follow-up studies exist for variables derived from Windkessel or cuff plethysmographic methods. It would be expected that augmentation indices would be related more closely to cardiac hypertrophy and heart failure, but such follow-up studies remain to be completed.
With respect to cardiac load in the present study, CTTI could be predicted acceptably well by knowing only office systolic BP, with or without simultaneous heart rate determination. Pulse pressure, which has been touted as being somewhat more predictive of coronary disease risk than is systolic pressure,18,19 was no more closely related than systolic pressure to any of the arterial stiffness-related parameters except C1. Pulse pressure was not as closely related to CTTI as was mean or systolic pressure. Thus, although mean pressure appears to be minimally better than systolic alone at identifying aortic stiffness, there appears to be little clinical justification for its use. The correlation of CTTI with CAP also was weaker than that of mean, systolic, or diastolic pressures, and the correlation of CTTI and C1 was weaker still. The lack of correlation of office BPs with parameters derived from either systolic or diastolic pulse contour analysis is also disappointing. Taken together, present results indicate that it is premature to employ the existing compliance methodologies in everyday clinical medicine.
This study was supported in part by FDA grant No. FDT-000889.
- Received April 28, 2001.
- Revision received July 9, 2001.
- Accepted August 16, 2001.
Burt VL, Cutler JA, Higgins M, Horan MJ, Labarthe D, Whelton P, Brown C, Roccella EJ. Trends in the prevalence, awareness, treatment, and control of hypertension in the adult US population: data from the health examination surveys, 1960 to 1991. Hypertension. 1995; 26: 60–69.
Koch-Weser J. Correlation of pathophysiology and pharmacology in primary hypertension. Am J Cardiol. 1973; 32: 499–499.
Franklin SS, Gustin WT, Wong ND. Hemodynamic patterns of age-related changes in blood pressure. Circulation. 1997; 96: 308–315.
Kelly R, Hayward C, Avolio A, O’Rourke M. Noninvasive determination of age-related changes in the human arterial pulse. Circulation. 1989; 80: 1652–1659.
Izzo JL Jr, Levy D, Black HR. Clinical advisory statement: importance of systolic blood pressure in older Americans. Hypertension. 2000; 35: 1021–1024.
Karamanoglu M, O’Rourke MF, Avolio AP, Kelly RP. An analysis of the relationship between central aortic and peripheral upper limb pressure waves in man. Eur Heart J. 1993; 14: 160–167.
Watt T, Burrus C. Arterial pressure contour analysis for estimating human vascular properties. J Appl Physiol. 1976; 40: 171–176.
Finkelstein SM, Collins VR, Cohn JN. Arterial vascular compliance response to vasodilators by Fourier and pulse contour analysis. Hypertension. 1988; 12: 380–387.
Finkelstein SM, Cohn JN. First- and third-order models for determining arterial compliance. J Hypertens. 1992; 10: S11–S14.
Barenbrock M, Spieker C, Hoeks AP, Zidek W, Rahn KH. Effect of lisinopril and metoprolol on arterial distensibility. Hypertension. 1994; 23 (Suppl I): I-161–I-163.
Benetos A, Safar M, Rudnichi A, Smulyan H, Richard JL, Ducimetiere P, Guize L. Pulse pressure: a predictor of long-term cardiovascular mortality in a French male population. Hypertension. 1997; 30: 1410–1415.
Asmar RA, Blacher J, London GM, Safar ME. Pulse pressure and aortic pulse wave are markers of cardiovascular risk in hypertensive populations. Am J Hypertens ion. 2001; 14: 91–107.
Franklin SS, Khan SA, Wong ND, Larson MG, Levy D. Is pulse pressure more important than systolic blood pressure in predicting coronary heart disease events? Circulation. 1999; 100: 354–360.