HEMODYNAMICS AND ARTERIAL PROPERTIES UNDERLYING PRESSURE
RESPONSES TO COGNITIVE STRESS IN BORDERLINE HYPERTENSIVES
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULLFILMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
This dissertation is dedicated to my husband and my son,
Chun-Jen and Michael Huang who have given me all these years of
love and support.
I gratefully acknowledge the assistance of Kristine Stouffer Calderon,
Ph.D., in data collection and of Sean Burket, Breah Elliott, and Elizabeth
Stevens, in subject recruitment; their dedication and hard work made this project
I extend my appreciation to my dissertation committee members, Carolyn
Yucha, Ph.D., Hossein Yarandi, Ph.D., Joyce Stechmiller, Wilmer Nichols, Ph.D.,
and Randy Braith, Ph.D, for their contributions and support of this project.
TABLE OF CONTENTS
ACKNOW LEDGEM ENTS ......... ...... ...... ..... ........................... iv
LIST OF TABLES .............. ................ ........... ... ... ....... vii
LIST OF FIGURES ................ ............ ........... ............. viii
ABSTRACT ................. ...... ............ ........ .. ......... x
1 INTRODUCTION ........................................... .. .......... 1
Background and Problem Statement ................. ........... ...... 1
Purposes of the Study ............ ............ ................... 3
Hypotheses ............. ........... ............ ........ ....... 4
Definitions of Terms ......... .......... ............ ....... 5
Assumptions .................................... ......... ......... 7
Lim stations .................. ..... ................................ ... .... 7
Significance of the Study ................................... 8
2 REVIEW OF LITERATURE ................................. 10
Mental Stress and Cardiovascular Reactivity in Hypertension ........... 10
Large-Artery Compliance .................................. ........ ............... 19
Noninvasive Measurements of Aortic Wave Reflection Amplitude by
Pulse W ave A analysis ............................................... ............... 23
Indirect Measurement of Total Arterial Compliance ......................... 26
Decreased Large-Artery Compliance and Arterial Structural
Adaptation in Hypertension ............. ... ......... ................ 27
Assessing Cardiovascular Reactivity ......... ...... .. ............ .......... 32
Im pedance Cardiography ............ ........................................... 34
Ambulatory Blood Pressure Monitoring ............... ...... ............ 36
Blood Pressure Hyperreactivity to Stress in Borderline
Hypertensives ......... .................... .......................... ......... 40
Summary ............. .......... .. .. ................. ..... ..... 48
3 PROCEDURES AND METHODS ...... ....... ........................... 50
D esign ............. ...................................... .. ...... 50
Population and Sam ple .............. ............. .. .......... ............... 50
S getting ........... .. ............. ...... .... ............... ... . . .... 52
Research Variables and Instruments .............. ............ ........... 53
Study Protocol .............................. .................... .......... 60
Methods of Statistical Analyses .............. .......... ................ 65
4 R ESU LTS ............. ......... ............................ ........ 66
Descriptive Results .......... ............. .............................. 66
Analytic Results for Hypotheses ............ .......... ... .......... 69
5 DISCUSSION AND CONCLUSIONS ............................ 97
Discussion of Results ........................... .... ......... . ........ 97
C conclusions .................................................................... 107
Implications for Clinical Practice ......................... ... ........... 109
Recommendations for Future Research ................................... 110
S u m m a ry .................................................................. ............ 1 12
A PATIENT CONSENT ...... ....... ........ ... ...... ............ 113
B DEMOGRAPHIC AND CLINICAL DATA COLLECTION SHEET... 119
REFERENCES ........ .... ......... ............................... ............ 120
BIOGRAPHICAL SKETCH ....... .................... ............... 131
LIST OF TABLES
3. Major variables measured or calculated during the laboratory
session ......... ... ......... ... .............. .. ................ ............. 64
4.1. Demographic data for the borderline hypertensive and
norm otensive subjects studied ...... ..... .................... .............. 68
4.2. Clinical characteristics of the borderline hypertensive and
normotensive subjects studied ............ ..... .............. ............ 69
4.3. Resting hemodynamics of the borderline hypertensive and
norm otensive subjects studied ...... ..... .................... .............. 70
4.4. Summary of beat-to-beat radial blood pressure during the stress
protoco l ............................................ .................... ... ... . ... 73
4.5. Group comparison of ambulatory blood pressure and variability
m measures ................................................... .... .......... ........... 81
4.6. Group comparison of nocturnal blood pressure and heart rate
reduction .. ........................................ .. 81
4.7. Group comparison of blood pressure reactivity and recovery ............. 93
4.8. Group comparison of ambulatory heart rate and variability
measures ......... .. ... ...... ................................... ........... 96
LIST OF FIGURES
2.1. Hypothesized link between stress and hypertension ........................... 14
2.2. The configuration of the ascending aortic pressure wave ................... 23
2.3. Hypothesized changes in hemodynamics underlying enhanced pressure
responses to cognitive stress of normokinetic borderline hypertensives
versus normotensives ............ ......... .................... ............ 49
3. Calculation of the augmentation index ....... .............. ... .. ............56
4.1. Systolic blood pressure changes across the three experimental periods
4.2. Diastolic blood pressure changes across the three experimental periods
.............. .................... ............ ............ .......... . ....... 75
4.3. Pulse pressure changes across the three experimental periods ............. 76
4.4. Heart rate changes across the three experimental periods .................. 77
4.5. A comparison of arterial compliance between the normotensive control
group and the normokinetic borderline hypertensive group ............... 80
4.6 Relationship between augmentation index (Al) and age (bivariate
correlation) ............................................. .......... 86
4.7. Relationship between augmentation index (Al) and height (bivariate
correlation) .............. .. ....................... ................... .... ......... 87
4.8 .Relationship between augmentation index (Al) and stroke volume
(bivariate correlation) ............ ............................. 88
4.9. Gender comparison of augmentation index (Al) ............. ............. 89
4.10. Augmentation index across the three experimental periods in the
normal control versus normokinetic borderline hypertensive groups ..... 90
4.11. Relationship between augmentation index (Al) and arterial compliance
(SV/PP) ........ ..................................................... .... ......... 91
4.12. Arterial compliance across the three experimental periods in the
normotensive control versus normokinetic borderline hypertensive
g ro u p s ........................................ ........................ ...... .... 9 2
4.13. A comparison of ambulatory, clinic and laboratory systolic blood
pressure ................................. .......... ... .......... 94
4.14. A comparison of ambulatory, clinic and laboratory diastolic blood
pressure ................................. .......... ... .......... 95
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
HEMODYNAMICS AND ARTERIAL PROPERTIES UNDERLYING PRESSURE
RESPONSES TO COGNITIVE STRESS IN BORDERLINE HYPERTENSION
Chairperson: Carolyn B. Yucha
Major Department: College of Nursing
It has been postulated that the blood pressure (BP) elevation observed in
borderline hypertension is caused by enhanced reactivity to stress. To further
elucidate the nature of pressure hyperreactivity in borderline hypertension,
hemodynamic and arterial responses to the Stroop Color Word Test (SCWT)
were studied in 23 borderline hypertensives and 19 normotensives, aged 20 to
63 and 24 to 64 respectively. Twenty borderline hypertensives had a cardiac
index less than the mean plus one standard deviation of the normotensive group
and were classified as the normokinetic borderline hypertensive subgroup. Blood
pressure reactivity to stress of the borderline hypertensives was compared to that
of the normotensives. Impedance-derived hemodynamics and arterial properties
underlying BP responses to stress were compared between the normokinetic
borderline hypertensive and the normotensive group.
Blood pressure was continuously recorded using a radial artery tonometer
during the stress protocol. Heart rate and stroke volume were measured using
the Minnesota impedance cardiograph. Augmentation index, an estimate of aortic
wave reflection, was measured using a radial applanation tonometer and a pulse
wave analysis system with a generalized transfer function. Arterial compliance
was estimated using the stroke volume to aortic pulse pressure ratio. Ambulatory
BP was also recorded at 30 to 60 minute intervals for a 24-hour period using an
autonomic noninvasive cuff-oscillometric recorder.
Total peripheral resistance was higher in borderline hypertensive than in
normotensive subjects. While arterial compliance was lower in normokinetic
borderline hypertensives than in normotensives, augmentation index was similar
between groups. The SCWT induced significant increases in systolic BP,
diastolic BP and mean arterial pressure, which were of similar magnitude for
borderline hypertensives and age-matched normotensives. Overall,
hemodynamic and arterial responses to the SCWT were similar in normokinetic
borderline hypertensives and normotensives. The BP reactivity to cognitive
stress likely resulted from tachycardia rather than a vasoconstriction response,
and was not associated with a change in wave reflection or compliance.
Evaluated with ambulatory BP monitoring, borderline hypertensive subjects had a
generally higher BP variability than normotensive subjects. Augmentation index
did not predict the degree of nocturnal BP reduction, nor did it differentiate
borderline hypertensives from normotensives, or dippers from nondippers.
This chapter will introduce the main variables under investigation including
cardiovascular reactivity, aortic wave reflection, arterial compliance, impedance-
derived hemodynamics, and ambulatory blood pressure. It will state the
background and main research problem to be investigated as well as the
hypotheses to be tested. The definitions of the major terms, assumptions,
limitations, and significance of the study will also be described.
Background and Problem Statement
A widely held hypothesis is that cardiovascular hyperreactivity is a risk
factor for the development of essential hypertension. The reactivity hypothesis
draws on the response of the sympathetic nervous system (SNS) to stressful
environmental stimuli. Studies of the SNS response to stress involve several
limitations, including limited reproducibility of stressful stimuli-induced responses
(Mancia, 1997), limited external validity of laboratory stressors (Olga, Lucio,
Gueseppe, Stefano, & Paolo, 1995; Seibt, Boucsein, & Scheuch, 1998; Steptoe,
Roy, Evans, & Snashall, 1995), and the limited role the SNS is believed to play in
the long-term regulation of arterial pressure (Mark, 1996). Decreased arterial
compliance, which is associated with cardiovascular risk (Blacher, Asmar, Djane,
London, & Safar, 1999; Bortolotto, Blacher, Kondo, Takazawa, & Safar,
2000; Lehmann et al., 1998) and resetting of baroreceptors (Shepherd, 1990),
may provide the link between cardiovascular hyperreactivity and the
development of hypertension. Ambulatory monitoring of blood pressure (BP) and
heart rate (HR) offers valuable information to describe cardiovascular reactivity to
naturally occurring stress. Therefore it is important to incorporate measurements
of arterial compliance and ambulatory BP (ABP) monitoring in studies that
investigate the physiological mechanisms of cardiovascular reactivity in order to
further our knowledge of the link between cardiovascular reactivity and
Borderline hypertensives are individuals whose BP fluctuates around the
normal cut-off point defining hypertension. They are perfect human subjects for
studies of reactivity hypothesis because 1) cardiovascular changes during
psychological stressors resemble the pattern of cardiovascular activity observed
in borderline hypertension (Turner, 1994), and 2) the risk of developing
hypertension is increased in borderline hypertensives (Julius, 1986). However,
circulatory patterns are not uniformly identical among all subjects with borderline
hypertension. Borderline hypertensives with different circulatory states may
exhibit differences in the degree of nocturnal BP reductions and may respond to
stress via different physiological pathways (Jern, Bergbrant, Hedner, & Hansson,
1995). Arterial wall properties may explain the degree of nocturnal BP reduction
(Amar et al., 1997; Asar et al., 1996).
Hence, this study was proposed to undertake measurements of BP,
impedance-derived hemodynamics, and arterial properties noninvasively in
normotensives and age-matched borderline hypertensives (the latter subdivided
into normokinetic and hyperkinetic) before, during, and after a Stroop Color Word
Test. Ambulatory BP was also recorded at 30 to 60 minute intervals for a 24-hour
period. Measurements of the physiological basis and hemodynamic patterns
underlying blood pressure responses to laboratory stressors further our
understanding of possible mechanisms of hypertension development. Assessing
impedance-derived hemodynamic parameters and arterial properties during
stress helps to elucidate the mechanism of enhanced blood pressure responses
to cognitive stress in borderline hypertensives. Ambulatory BP recordings
document the circadian variations and clarify BP reactivity during everyday life.
Measuring arterial compliance as indexed by the SV/PP ratio contributes to a
better understanding of individual differences in circadian blood pressure
Purposes of the Study
1. To examine impedance-derived hemodyamic parameters underlying BP
responses to laboratory cognitive stress in borderline hypertensives and age-
matched normotensive subjects. Borderline hypertensives were further
divided into hyperkinetic and normokinetic subgroups on the basis of cardiac
index (CI). Blood pressure responses were measured using continuous
noninvasive arterial tonometry. Heart rate, stroke volume (SV) and total
peripheral vascular resistance (TPR) were measured using impedance
2. To characterize the difference in aortic wave reflection between
normotensives and normokinetic borderline hypertensives. Al measured by
applanation tonometry and a generalized transfer function indexed aortic
wave reflection amplitude.
3. To characterize the difference in arterial compliance between normotensives
and normokinetic borderline hypertensives. Arterial compliance was
estimated by the ratio of stroke volume and aortic pulse pressure: SV/PP.
4. To explore the pattern of changes in arterial compliance underlying BP
responses to cognitive stress in normotensives and borderline hypertensives.
5. To investigate the difference in 24-hour ABP profiles between borderline
hypertensives and normotensives.
6. To determine the association between augmentation index and the day-night
7. To determine the association between arterial compliance and the day-night
8. To determine the predictive power of pressure reactivity to cognitive stress on
1. Borderline hypertensives have a greater increase in BP in response to
cognitive stress as compared to age-matched normotensive subjects.
2. Normotensives and normokinetic borderline hypertensives differ significantly
in their responses in HR, SV, and TPR to cognitive stress.
3. Normotensives and normokinetic borderline hypertensive subjects differ
significantly in aortic wave reflection as indexed by Al.
4. Normotensives and normokinetic borderline hypertensive subjects differ
significantly in arterial compliance as indexed by SV/PP.
5. Borderline hypertensives exhibit a greater ABP variability as determined by a
greater standard deviation of the average 24-hour BP, as compared to the
6. Normotensives exhibit a significantly greater day-night BP difference as
compared to normokinetic borderline hypertensives.
7. Aortic wave reflection as indexed by Al is predictive of the day-night BP
8. Arterial compliance as indexed by SV/PP is predictive of the day-night BP
9. Pressure reactivity is predictive of the average 24-hour and daytime BP.
Definitions of Terms
Aortic Wave Reflection Amplitude
In this study, radial pulse waves were obtained via applanation tonometry
and aortic pressure waveforms were then derived using a transfer function. Al
was calculated as an index of aortic wave reflection amplitude (Kelly, Hayward,
Avolio, & O'Rourke, 1989). As arterial compliance decreases, Al increases.
Compliance (inverse of stiffness) is defined as the change in volume per
unit change in pressure. In this study, arterial compliance is estimated as the
ratio of SV to aortic PP (Chemla et al., 1998; Ferguson, Julius, & Randall, 1984).
Impedance cardiography refers to the measurement of thoracic blood
volume changes in relating to myocardial performance derived from changes in
the electrical impedance (Sherwood et al., 1990).
Impedance-derived hemodynamics refer to cardiac output (CO), SV, HR,
and TPR derived by measuring impedance changes occurring with each
heartbeat. In this study, these hemodynamic measures were obtained via the
Minnesota Impedance Cardiogram and SV was derived by the formula
developed by Kubicek et al. (1974):
SV = p x (L/Zo)2 x Tx (dZ/dt)min
where Rho (p) is the resistivity of blood, L is the mean distance between the two
inner electrodes in cm, Zo is the mean thoracic impedance between electrodes 2
and 3, T is the ventricular ejection time in seconds, and (dZ/dt)min is the minimum
value of dZ/dt occurring during the cardiac cycle in ohms per second. CO was
calculated from the formula: SV x HR. TPR was derived from the formula: MAP -
CO, where MAP is the mean arterial blood pressure derived from simultaneous
measurement of BP.
Pressure reactivity was defined as the absolute change of BP from the
baseline to the stress level. The absolute reactivity score was calculated by
subtracting the baseline BP from the stress level.
Pressure recovery was defined as the absolute change of BP from the
stress to the recovery level. The absolute recovery score was calculated by
subtracting the recovery BP from the stress level.
There is no universally accepted operational definition for borderline
hypertension, which is usually referred to pressures fluctuating around the border
of the cutoffs defining hypertension (Sherwood, Hinderliter, & Light, 1995). In this
study, borderline hypertension is referred to as a transition from normotension to
hypertension. The definition of the "high-normal" and "hypertension-stage 1"
defined by the Sixth Report of the Joint National Committee (JNC-VI) on
Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC-
VI, 1997) was selected as the definition of borderline hypertension.
Subjects were categorized as borderline hypertensive if both of the two
screening systolic BP (SBP) measurements obtained on two separate occasions
fell in the range of 130 to 159 mm Hg and/or diastolic BP (DBP) measurements
fell in the range of 85 to 99 mm Hg.
Subjects were categorized as normotensive if both of the two screening
SBP measurements obtained on two separate occasions fell below 130 mm Hg
and DBP fell below 85 mm Hg.
1. Blood pressure response is obligatory and stereotypical.
2. A decrease in cardiac responsiveness must be accompanied by an
increase in vascular responsiveness.
Categorization of hyperkinetic versus normokinetic borderline
hypertension was based on impedance-derived CI measurements. Borderline
hypertensive subjects who had a CI greater than the mean plus one standard
deviation of the normotensive control group were categorized as the hyperkinetic
borderline hypertensive subgroup. Those who had a CI below that level were
referred to as the normokinetic borderline hypertensive subgroup. As will be
discussed in Chapter 2, impedance cardiography is highly reliable in measuring
changes in CO but may not reflect individual absolute values. One might argue
that the accuracy of hemodynamic subgrouping within borderline hypertensives
was questionable. It has been suggested that the sources of measurement error
with impedance-derived measures are assumed to be systematic (Sherwood et
al., 1990). Based on this assumption, hemodynamic subgrouping remained
unchanged whether there was a systematic underestimation or overestimation of
Only 3 out of 23 borderline hypertensive subjects studied were
hyperkinetic by the subgrouping strategy. This uneven distribution made the
original intent to compare the two subgroups of borderline hypertensives
impossible. Nevertheless, a comparison between the normokinetic borderline
hypertensives and normotensives still offered valuable information.
Significance of the Study
Hypertension is one of the major causes of death in the United States
(Murphy, 2000). Borderline hypertensives have an increased risk of developing
essential hypertension (Julius, 1986). Enhanced BP responses to laboratory
stress have been found in borderline hypertensive subjects. The underlying
physiologic changes of enhanced pressure response to stress in borderline
hypertensives are still poorly understood. Whether increased BP responses in
borderline hypertensives are caused by an autonomic hyperreactivity or caused
by structural changes in peripheral vasculature remains controversial. This study
characterized differences in impedance-derived hemodyamic parameters and
arterial properties underlying BP responses to laboratory cognitive stress
between normokinetic borderline hypertensives and normotensives. A better
understanding of the physiologic mechanisms of BP variability in borderline
hypertensives may help in the selection of treatments for hypertension.
It is important to examine blood pressure variability during everyday life
when evaluating the reactivity hypothesis, given the limited external validity of
laboratory stressors. Therefore, this study also investigated the differences in BP
variability between borderline hypertensives and normotensives using ABP
This study also utilized ABP monitoring to gather information on the day-
night BP difference. Determining the association between arterial compliance
and the day-night BP difference may help to increase our understanding of the
nature of individual differences in circadian BP variation.
REVIEW OF LITERATURE
This chapter will present a literature review of the following areas of
research: mental stress and cardiovascular reactivity in hypertension, large-
artery compliance, noninvasive measurements of aortic wave reflection
amplitude by pulse wave analysis, indirect measurement of total arterial
compliance, decreased large-artery compliance and arterial structural adaptation
in hypertension, assessing cardiovascular reactivity, impedance cardiography,
ambulatory blood pressure monitoring, and pressure reactivity to mental stress in
borderline hypertensives. A summary linking these areas together to provide the
research rationale for this study will conclude this chapter.
Mental Stress and Cardiovascular Reactivity in Hypertension
Overview of Stress and Reactivity in Hypertension
The reactivity hypothesis holds that exaggerated cardiovascular
responses to stress may contribute to the long-term development of essential
hypertension, or more precisely to the initiation of its development (Turner,
1994). No definitive evidence supports the link between cardiovascular
hyperreactivity and hypertension in human subjects. However, several lines of
evidence suggest that high reactors respond to behavioral stressors with
inappropriate cardiovascular responses (Light, 1987). The inappropriate
cardiovascular responses to stressful stimuli set pathophysiologic alterations in
motion that may eventually lead to the development of hypertension.
The reactivity hypothesis is based on the observations that offspring of
hypertensive parents (Falkner, 1991; Fredrikson, Tuomisto, & Bergman-Losman,
1991; & Perini et al., 1990) and borderline hypertensive subjects (Falkner, 1991;
Perini et al., 1990; Spence et al., 1990) exhibit greater circulatory and/or
neurohormonal reactivity to mental stress than normotensives. The evidence that
cardiovascular changes during psychological stressors resemble the pattern of
cardiovascular activity observed in borderline hypertension further supports the
notion of a hypothesized link between reactivity and hypertension (Turner, 1994).
Fredrikson (1991) suggested that a neurogenically mediated hyperreactivity to
stress is a precursor not a consequence of hypertension. Yet, the mechanism of
cardiovascular reactivity to stress in the etiology of hypertension has not been
A critical question underlying the role of the reactivity hypothesis in
hypertension is whether reactivity to laboratory stress can predict future
hypertension. Higher SBP and DBP responses to a combination of mental and
physical stressors were associated with subsequent higher resting DBP among
206 middle-aged adults 6.5 years later (Matthews, Woodall, & Allen, 1993). In an
attempt to further clarify the predictive value of stress reactivity on later BP, 103
men originally tested at age 18 to 22 were reevaluated 10 years later (Light et al.,
1999). High reactors (top 25% on the basis of BP and cardiac responses during
both reaction time and cold pressor test) with positive family history of
hypertension demonstrated significantly higher SBP and DBP at follow-up. In a
subgroup of subjects who provided ratings of daily stress, the interactions among
positive family history, high daily stress and high stress reactivity were significant
in predicting follow-up SBP and DBP, suggesting that stress reactivity as a long-
term predictor is modulated by both genetic and environmental factors.
The reactivity hypothesis emphasizes individual differences in response to
stress with little attention paid to the nature of stress. Evidence suggests that
enhanced reactivity among hypertensive individuals may be related to the nature
of stressors. More specifically, mild hypertensives display exaggerated BP
reactivity to stress requiring active coping but not passive coping (Fredrikson,
1992; Steptoe, Melville, & Ross, 1984). Tasks involving active coping such as the
Stroop Color Word Test (SCWT) elicit predominantly beta-adrenergically
mediated cardiac activation (Hjemdahl, Freyschuss, Juhlin-Dannfelt, & Linde,
1984). Moreover, some researchers suggested that certain conditions such as
the presence of an incentive condition (Waldstein, Bachen, & Manuck, 1997) or
the level of task difficulty (Callister Suwarno, & Seals, 1992) also modulate the
hemodynamic basis of reactivity. However, in a study assessing hemodynamics
of BP responses during active and passive coping in 90 young healthy adult male
subjects, hemodyamic patterns of reactivity were an individual characteristic only
partially modified by coping demands (Sherwood, Dolan, & Light, 1990).
Hypothesized Link between Stress Hyperreactivity and Hypertension
Manuck and Kranz (1984) proposed two models to explain the relationship
between stress and hypertension: the recurrent activation model and the
prevailing state model. In the recurrent activation model, repeated transient
episodes of sympathetic arousal caused by stress increase the susceptibility to
hypertension. In the prevailing state model, a relatively persistent sympathetic
arousal alters autonomic regulation of BP. In the recurrent activation model,
laboratory-induced BP or HR reactivity corresponds to the magnitude of BP peak
during the everyday life, whereas in the prevailing state model laboratory-induced
BP or HR reactivity is indicative of the magnitude of the constant elevation of BP
during the waking hours (Manuck & Krantz, 1984). These two models differed
from each other not only in the presumptions but also in the underlying
pathophysiologic mechanisms. The differences with reference to underlying
mechanisms will be discussed later. Hypothesized links between hyperreactivity
to stress and the development of hypertension have been proposed, including
autoregulation theory, structural adaptation, and baroreceptor resetting (see
Autorequlation. Obrist (1976) suggested that transient episodes of SNS
arousal result in excessive CO that may lead to the development of hypertension.
Obrist's notion appears to be in line with the recurrent activation model. Obrist
(1981) further indicated that the presence of exaggerated CO increases during
psychological stress lead to overperfusion of the body that eventually increases
peripheral vascular resistance through compensatory autoregulatory
vasoconstriction. Obrist (1981) also maintained that autoregulation is a transition
mechanism between the early stage and the more advanced stage of
hypertension. As described by Guyton and Coleman (1969), the whole body
autoregulates vascular tone and vascular resistance increases under conditions
of volume expansion. As vascular resistance increases a new equilibrium of
normal CO will be reached accompanied by a high BP.
3. Resetting of
T HR T CO
Increased Arterial Pressure
Figure 2.1. Hypothesized links between stress and hypertension.
Fredrikson (1991) described two interrelated theories on SNS overactivity
in the development of essential hypertension. The "hyperreactivity theory"
suggests that individual differences in the autonomic nervous system activity are
the pathophysiological mechanism in the development of essential hypertension
and the "symptom-specificity theory" suggests that stereotypical SNS responses
enhance the risk of hypertension. The hyperreactivity theory is consistent with
the recurrent activation model and the above mentioned hypothesis proposed by
Obrist that transient episodes of increased CO induce hypertension by increasing
peripheral vascular resistance (Fredrikson, 1991).
Structural adaptation. As cited by Manuck and Krantz (1984), Julius and
colleagues proposed a model resembling the prevailing state model, which
stated that certain behavioral characteristics induce a permanent state of
enhanced arousal accompanied by alterations in centrally integrated
cardiovascular autonomic tone that accounts for an increased SNS activity that
increases CO and consequently BP. Julius and Nesbitt (1996) further showed
evidence against autoregulation as the mechanism explaining the transition of
borderline hypertension to established hypertension. They found a lack of volume
expansion and no evidence of increased oxygen consumption in borderline
hypertension, suggesting that the increased CO is appropriate for the increased
metabolic needs of the body.
As previously discussed, once hypertension is established CO is usually
not increased, but rather peripheral vascular resistance is increased. The whole-
body autoregulation theory does not apply to the chain of hemodynamic
transitions including a decrease in CI and an increase in TPR over time (Lund-
Johansen, 1986; Julius & Nesbitt, 1996) as the whole-body oxygen consumption
is increased (Lund-Johansen, 1986) and plasma volume is relatively low (Julius
& Nesbitt, 1996) in borderline hypertensives.
Folkow (1993) suggested that in the early phase essential hypertension is
mainly manifested as enhanced central neurohormonal activity and CO is
increased more than TPR. As the interaction among genetic predisposition,
environmental effects, and aging effects progresses, structural upward resetting
of the heart and vessels becomes more evident and manifested as increased
Julius and Nesbitt (1996) indicated that increased sympathetic stimulation
and decreased parasympathetic tone mediate the early phase (hyperkinetic
state) of hypertension. The involvement of both enhanced alpha- and beta-
adrenergic tone combined with a blunted parasympathetic tone suggests a
centrally integrated abnormality. Continued sympathetic stimulation in
hypertension eventually leads to myocardial (i.e., decreased beta-receptor
responsiveness and cardiac compliance) and vascular (i.e., vascular
hypertrophy) structural changes, which then cause decreased cardiac
responsiveness and increased vascular responsiveness. Decreased beta-
adrenergic responsiveness and decreased end-diastolic distention of the heart in
combination with increased alpha-adrenergic responsiveness of the resistance
vessels lead to the hemodynamic transition (from high CO to high resistance)
over the course of hypertension (Julius, 1991). In parallel to the hemodynamic
transition, the SNS tone is down regulated since less sympathetic drive is
needed with exaggerated vascular responsiveness (Julius, 1991). Although
focusing on different aspects of pathogenesis, the pathophysiological
mechanisms underlying hemodynamic transition over the course of hypertension
proposed by Julius seem to be consistent with Folkow's structural upward
Resetting of arterial baroreceptors. Arterial baroreceptors are sensitive to
arterial pressure. In the short term, they detect a change in pressure and regulate
blood pressure through a sino-aortic baroreflex feedback loop (Guyton, 1991).
Arterial baroreceptors respond to an increase in blood pressure by increasing the
afferent traffic to the cardiac acceleration center in the medulla. As a result, the
efferent sympathetic traffic is decreased. In chronic hypertension, the
baroreceptors can become reset over time, so that the pressure is maintained at
a higher level.
Several mechanisms for resetting the arterial baroreceptors have been
suggested, including viscoelastic relaxation of wall elements, genetic
abnormalities in membrane permeability, decreased arterial compliance,
alterations in endothelium-derived factors (Shepherd, 1990), and a defect in
central mediation of the baroreflex (Chapleau, Cunningham, Sullivan, Wachtel, &
Abboud, 1995). Of more relevance for present concerns, decreased arterial
compliance seems to provide the link among stress, sympathetic overactivity and
hypertension. Markedly increased carotid artery compliance has been shown
following sympathectomy in rats, suggesting a sympathetic modulation of arterial
compliance (Mangoni, Mircoli, Giannattasio, Mancia, & Ferrari, 1997). In chronic
hypertension, an increase in collagen as a result of structural changes in the
vessel wall secondary to increased sympathetic outflow might stiffen the large
arterial walls. Stiffening of vessel walls in the aortic arch and carotid sinus alters
the responsiveness of baroreceptors to stretch. Hence, decreased
mechanoreceptor activity and chronic resetting could occur.
On the other hand, however, baroreflex impairment has not been
demonstrated for the blood pressure component of the reflex in hypertension
(Mancia, 1997). Thus, the baroreflex does not appear to be the origin of
Limitations of the Reactivity Hypothesis
Despite years of investigation, the role played by cardiovascular
hyperreactivity in the development of hypertension remains largely unknown.
Reactivity hypothesis draws on the response of the SNS to stressful
environmental stimuli. Studies of the response of the SNS to stress involve
several limitations. First, the effect of any given stressful stimuli has limited
reproducibility (Mancia, 1997). An individual can respond to one stressor with
profound cardiovascular responses and respond to other stressors without
cardiovascular arousal. Second, laboratory stressors have limited external
validity. Some suggest that the BP response (Olga et al., 1995) or reactivity
(Seibt et al., 1998; Steptoe et al., 1995) to laboratory stressors does not predict
ambulatory blood pressure during everyday life. Finally, it has been argued that
the SNS is not important in the long-term regulation of arterial pressure (Mark,
1996). This argument is based on the observation that several pathological
conditions with marked sympathetic activation, such as heart failure and
cirrhosis, are not accompanied by increased arterial pressure (Mark, 1996).
To date, no definitive evidence supports the link between stress
hyperreactivity and the development of hypertension in human subjects, although
several possible links between stress hyperreactivity and hypertension have
been proposed. The following section will review arterial compliance, an
important concept in the natural history of hypertension. Investigating the
contribution of arterial compliance to cardiovascular hyperreactivity may provide
new insights into the role of stress hyperreactivity in the development of
Arterial Compliance and Pulse Wave Amplitude, Velocity, and Reflection
Compliance, distensibility and stiffness are often used interchangeably by
researchers. Mathematically and conceptually, they are different. Compliance of
an arterial segment represents the increase in its cross-sectional volume for a
given increase in pressure, which does not account for the diastolic dimension
before distension begins. Distensibility is compliance normalized by arterial
diameter. Stiffness is the reciprocal of dispensability.
Recently, the concept of hypertension as a condition characterized by
elevated peripheral vascular resistance has been challenged because it ignores
the pulsatile component, PP, of arterial pressure. Ventricular ejection interacting
with viscoelastic properties of the large arteries and wave reflection are two
major determinants of the pulsatile component of BP. An increased PP reflects
decreased arterial compliance (inverse of stiffness). Thus, elevated SBP and PP
are now considered to be more reliable indicators of cardiovascular risk than is
DBP (Alderman, 1999; Benetos, 1999; Safar, 1999).
In order to better understand this reevaluation of the indicators of
cardiovascular risk, it is important to reexamine the concept of large arteries as
simple passive conduits. Large arteries cushion pulsatile flow generated by
ventricular contraction during each cardiac cycle, transforming intermittent flow
into a steady flow of blood in the periphery and reducing the pressure oscillations
caused by the intermittent ventricular ejection (London & Guerin, 1999 & Safar,
1989). The viscoelastic properties of the arterial wall, termed arterial compliance
or distensibility, determine the artery's ability to perform this cushion function.
During systole, large arteries (e.g., the proximal aorta) expand to accommodate
SV and rebound during diastole to facilitate forward blood flow. When arterial
compliance is decreased, less cushioning of SV occurs in the arterial bed during
systole, so a greater proportion of SV is forwarded to the periphery.
Consequently, the amplitude of the arterial pulse wave during systole increases
and diastolic pressure falls (London & Guerin, 1999).
Arterial compliance also determines pressure wave travel and reflection in
the arterial system. The incident wave generated by the ventricular ejection
moves away from the heart at a finite speed. The speed of pulse wave
propagation (pulse wave velocity) increases as arterial compliance decreases.
The proximal aorta is relatively compliant to accepting the stroke volume. As the
aorta extends distally, it becomes less compliant.
The compliance of the arteries also decreases with distance from the
heart. Because the velocity of wave travel is faster in stiffer and smaller vessels,
this structure causes a progressive increase in pulse wave velocity along the
arterial tree (Nichols & O'Rourke, 1998). There is also a progressive increase in
pulse pressure from the central aorta to peripheral sites. This pulse pressure
amplification is less evident with aging (Nichols & O'Rourke, 1998), probably due
to the increased aortic pulse wave velocity, which is secondary to aortic stiffness.
In healthy adults, the pulse wave velocity is approximately 7.3 m/s from the heart
to the radial artery, and 8.0 m/s from the heart to the femoral artery (Nichols &
The incident (forward) wave is reflected backward along the arterial
system from peripheral reflection sites to the ascending aorta. A high level of
arterial stiffness (decreased compliance) causes an increase in pulse wave
velocity and earlier return of the reflected wave to the ascending aorta (Nichols &
O'Rourke, 1998; O'Rourke, 1990). Early return of reflected pressure waves adds
to the amplitude of the incident wave during the systolic phase. Thus, the
augmentation in the systolic part of the ascending aortic pressure further
increases systolic pressure and ventricular afterload.
In summary, as arterial compliance decreases, both the amplitude of the
pressure wave generated by ventricular ejection and pulse wave velocity
increase, causing an early return of the reflected pressure waves from the
periphery to the aorta. An increase in large-artery stiffness and a consequent
increase in the amplitude of wave reflection results in a disproportionate increase
in SBP and arterial pulsatility. This disproportionate increase in SBP increases
pulse pressure, producing the phenomenon known as isolated systolic
hypertension (Nichols, Nicolini, & Pepine, 1992).
Aortic Pressure Wave Contours
The aortic pulse pressure waveform generally contains an inflection point,
caused by the reflected wave, that divides the pressure wave into an early and a
late systolic peak (Type A and Type B contours, according to Murgo's
classification) (Nichols & O'Rourke, 1998). Middle-aged adults or individuals with
hypertension exhibit Murgo Type A and B aortic pressure wave contours with
peak systolic pressure (Ps) occurring in late systole after an inflection point (Pi).
(Ps Pi) / (Ps Pd) has been termed the augmentation index (Al), where Pi is
the first systolic shoulder or inflection point, Ps is the second systolic shoulder or
peak systolic pressure (SBP), and Pd is the minimum DBP (Kelly et al., 1989). A
Type A pressure wave contour has an Al of greater than 12%. Al is between 0
and 12% in a Type B pressure wave contour. Younger adults exhibit a Murgo
Type C contour with a smaller reflected wave beginning in late systole or
diastole. In such cases, Ps precedes Pi. Al is less than zero in a Type C pressure
wave contour. The Type A and Type C pressure wave contours are graphically
presented in Figure 2.2. A Type D beat classified by Nichols and colleagues, is
only seen in adults over 65 years of age (Nichols & O'Rourke, 1998). In Type D
beats, there is no inflection point because the reflected pressure wave begins in
early systole and merges with the incident wave. In summary, the aortic pressure
wave contour and Al change with age. Augmentation index is an index of aortic
wave reflection amplitude and will be discussed in the following section.
Figure 2.2. The configuration of the ascending aortic pressure wave. Pi: inflection
point, Ps: peak systolic pressure, Pd: minimum diastolic pressure. The diagram
on the left is a Type A contour. Notice that in a Type A contour the peak systolic
pressure (Ps) occurs in late systole after an inflection point (Pi). The right
diagram is a type C contour. Notice that in a Type C contour the peak systolic
pressure (Ps) precedes the inflection point (Pi).
Noninvasive Measurements of Aortic Wave Reflection Amplitude by Pulse
In studying arterial modifications in hypertension, the determination of
arterial compliance and wave reflection amplitude can be estimated by pulse
wave analysis (PWA) (O'Rourke & Gallagher, 1996; Nichols & O'Rourke, 1998).
PWA or sphygmocardiography is the analysis of aortic pulse waveforms to
identify systolic and diastolic periods and to generate indices of ventricular and
vascular interaction, including the aortic augmentation index (O'Rourke &
Gallagher, 1996). Modern PWA allows synthesis of the ascending aortic pressure
wave, central arterial BP and related indices from carotid or radial artery pressure
waves by use of noninvasive arterial applanation tonometry with a validated
generalized transfer function (Karamanoglu, O'Rourke, Avolio, & Kelly, 1993;
Nichols & O'Rourke, 1998; O'Rourke & Gallagher, 1996; Siebenhofer, Kemp,
Sutton, & Williams, 1999).
The application of applanation tonometry is based on the principle that
when the curved surface of a rounded pressure-containing chamber (artery) is
partially flattened, pressures are normalized and a sensor on the flattened
surface can record the pressure within the artery (O'Rourke & Gallagher, 1996).
Augmentation Index (Al)
The aortic Al is an indicator of the magnitude of wave reflection, which is
closely linked to aortic pulse wave velocity (Kelly et al., 1989). Aortic pulse wave
velocity is an well-established measurement of arterial compliance. Pulse wave
augmentation of the synthesized wave is determined by identification of the tidal
wavefoot and expressed as a percentage of pulse pressure (O'Rourke &
Gallagher, 1996). Al measures the relative contribution of the reflected wave to
the SBP in central arteries. Kelly et al. defined Al as the percentage of aortic
pulse pressure due to the late systolic peak, which is attributed to the reflected
wave. Al is calculated as the difference between early and late aortic systolic
peaks divided by aortic PP.
Results of a study by Karamanoglu et al. (1993) using a generalized
transfer function in adult humans indicate acceptable accuracy (> 90 %) in
generating an aortic waveform from the pressure wave in the radial or brachial
artery. In healthy adult subjects with a wide age range, Liang et al. (1998) have
demonstrated that measurement of Al with applanation tonometry is highly
reproducible and precise, with a correlation coefficient of .98 between visits and a
coefficient of variation of 1.3%. Likewise, Siebenhofer et al. (1999) also reported
that Al by applanation tonometry and PWA has excellent interobserver
reproducibility, with interobserver measurement difference being only 2.7%.
The reproducibility of Al measurements via applanation tonometry has
also been tested on subjects with cardiovascular risk factors. Wilkinson et al.
(1998) demonstrated in a group of 33 subjects (12 diabetes, 16 hypertensives,
and 5 controls) that PWA using a radial applanation tonometry is a highly
reproducible method for determining Al. They observed a low standard deviation
for within-observer and between-observer measurement differences (5.37% and
3.80 %, respectively).
Possible correlates of Al include age (Cameron, McGrath, & Dart, 1998;
Kelly, Millasseau, Ritter, & Chowienczyk, 2001; Liang et al., 1998, Yasmin &
Brown, 1999), body height (Cameron et al., 1998; Rietzschel, Boeykens, De
Buyzere, Duprez, & Clement, 2001; Yasmin & Brown, 1999), and HR (Cameron
et al., 1998; Wilkinson et al., 2000). A positive correlation between age and Al
was found by previous reports with a correlation coefficient ranging from .40 to
.70. A negative correlation was found between height and Al with a correlation
coefficient ranging from -.43 to -.49. The correlation coefficient for HR and Al
varied among studies and ranged from -.28 to -.76.
In summary, Al is an index of the magnitude of aortic wave reflection. Al
measurements via applanation tonometry and PWA demonstrated acceptable
accuracy and reproducibility in previous studies. Yet, this measure is slightly to
moderately correlated with HR, height, and age.
Indirect Measurement of Total Arterial Compliance
Arterial compliance provides information about the viscoelastic properties
of the arteries. The SV/PP ratio has been proposed as an estimate of total
arterial compliance (Chemla et al., 1998; Ferguson et al., 1984; Randall,
Westerhof, van den Bos, & Alexander, 1986). Randall et al. demonstrated a
close relation between compliance estimated by the SV/PP ratio and by the ratio
of the diastolic-decay time constant using a Windkessel model (r = .58). Chemla
et al. demonstrated that the difference between the SV/PP ratio and compliance
estimated by the area method was not significant and there was a linear
correlation between the two measurements (r = .98). Resnick et al. (2000)
demonstrated that the ratio of SV to PP was correlated with both capacitive (C1)
and oscillatory (C2) compliance of the arterial tree using a modified Windkessel
model (r = .92 and .68, respectively). Such studies have validated the accuracy
of the SV/PP ratio as an estimate of arterial compliance.
SV/PP may also be a possible predictor of cardiovascular morbid events.
de Simone et al. (1999) demonstrated that the SV/PP ratio was an independent
predictor of total cardiovascular events in Cox proportional hazards analysis
independent of age and left ventricular mass.
Possible correlates of the SV/PP ratio include age, body size, and HR
(Gudbrandsson et al., 1992). The distribution of the SV/PP ratio was skewed
toward larger values in normal young adults with larger body size. Those
subjects with the lowest SV/PP ratio had higher SBP, lower DBP, higher HR,
and higher left ventricular wall thickness ratio compared to subjects with higher
Decreased Large-Artery Compliance and Arterial Structural Adaptation in
Although decreased large-artery compliance has been identified as a
major risk factor for cardiovascular events (Blacher et al., 1999; Bortolotto et al.,
2000; Lehmann et al., 1998), its role in the pathogenesis of arterial hypertension
remains largely unknown. Decreased arterial compliance is most markedly seen
in central elastic arteries such as the aorta (O'Rourke, 1990; O'Rourke, 1999;
Smulyan & Safar, 1997). The compliance or distensibility of the aorta, which is
determined by its wall structure, its smooth muscle tone, and the BP level,
decreases with aging and arterial hypertension (O'Rourke, 1990; O'Rourke,
Since Folkow (1987) demonstrated that structural adaptation occurs in
response to altered functional demands in hypertension, increased arterial wall
thickness, termed intima-media thickness, is widely regarded as a hallmark of
hypertension. Increasing strong evidence suggests that the increased intima-
media thickness is caused by rearrangement of existing materials around a
vascular lumen and a consequent reduction in lumen diameter, a process called
vascular remodeling (Kaplan, 1998).
Decreased arterial compliance may play a role in the pathogenesis of
hypertension. It may occur irreversibly as a result of arterial changes (mainly an
increased intima-media thickness). Alternatively, decreased arterial compliance
may be a consequence of elevated arterial pressure, which occurs without any
structural change in the artery.
Gribbin, Pickering and Sleight (1997) found that an acute rise in arterial
distending pressure induced a rise in pulse wave velocity. Pulse wave velocity is
a well-accepted index of arterial compliance (Nichols & O'Rourke, 1998; Smulyan
& Safar, 1997). However, no difference in pulse wave velocity was found
between hypertensives (n = 17) and normotensives (n = 21) after mechanically
inducing identical arterial distending pressures of 120 mm Hg. The findings from
this study suggest that decreased compliance of large arteries in hypertension is
a consequence of high arterial pressure and is not due to irreversible arterial
changes. Laurent et al. (1993) also found no difference in isobaric compliance
and distensibility of the radial artery between hypertensive and normotensive
Reduced compliance of large arteries in patients with mild essential
hypertension might occur in the absence of structural vessel wall alterations
(Barentbrock et al., 1999). In this study 10 mild untreated hypertensive subjects
with normal intima-media thickness, 10 with increased intima-media thickness
(above 95% CI of normal controls) and 13 age-matched normotensives were
studied. Both groups of hypertensives did not differ in systolic and diastolic blood
pressure or heart rates. Brachial and carotid distensibility coefficients were
significantly lower in both groups of hypertensives when compared with
normotensives. Analysis of covariance further revealed that brachial artery
distensibility coefficients were significantly lower in both groups of hypertensives
when compared with normotensives, independent of age and brachial artery end-
diastolic diameter. There was no correlation between carotid artery intima-media
thickness and carotid or brachial artery distensibility coefficients.
In the Atherosclerosis Risk in Communities study, the relationship
between hypertension and arterial compliance assessed in 10,712 men and
women of 45 to 64 years of age (Arnett et al., 2000). Carotid artery compliance
was assessed using ultrasound technique with concurrent measurement of
brachial artery blood pressure. Hypertension was classified using the blood
pressure classification from the Fifth Joint National Committee. Using statistical
models to control for DBP and PP, arterial diameter change was calculated in
normotensive, medicated and nonmedicated hypertensives. Hypertension was
associated with a smaller adjusted carotid diameter change (greater stiffness)
when compared to normotensives, or medicated hypertensives with optimal
blood pressure. The pressure-adjusted diameter change was inversely related to
DBP in all gender-medication groups. Moreover, the relationship between PP
and carotid arterial diameter change (the slope of the PP and diameter change)
did not differ between hypertensives and normotensives, suggesting that
hypertension-related carotid arterial stiffness is the effect of distending pressure
rather than structural changes in the carotid artery.
Animal studies, however, have shown that hypertension decreases arterial
distensibility or compliance by inducing arterial structural changes, including
hypertrophy and increasing collagen in the extracellular matrix (Albaladejo at al.,
1994; Bunkenburg, van Amelsvoort, Rogg, Wood, 1992). In a newly published
report, van Gorp et al. (2000) assessed thoracic aortic wall properties using an
ultrasound technique combined with invasive BP measurements in 1.5, 3, and 6
month-old spontaneously hypertensive rats (SHR) and normotensive rats. In 1.5
month-old SHR, both compliance and the distensibility coefficient were
significantly lower and media thickness and cross-sectional areas were
significantly larger than normotensive rats, but blood pressure was not
significantly different. Compliance was significantly lower and blood pressure was
significantly higher in 3- and 6-month-old SHR than normotensive rats at the
corresponding age. Yet, the total collagen density and content of the thoracic
aortic wall were not significantly different between strains, van Gorp et al.
suggested that 1) in SHR in which hypertension develops over weeks, altered
thoracic aortic properties precede the development of hypertension; and 2) the
alterations in aortic properties are most likely caused by media hypertrophy
rather than a change in the composition of the wall.
The association of large arterial wall hypertrophy and decreased
compliance has been confirmed in human subjects. In 1996 Tice et al. examined
carotid artery intima-media thickness using B-mode ultrasound imaging and
carotid arterial waveforms using applanation tonometry in a group of 20 patients
with newly diagnosed essential hypertension, 18 patients with chronic
hypertension, and 32 subjects with normal BP. These investigators found that
carotid intima-media thickness was significantly greater in patients with early
hypertension and chronic hypertension as compared with that in control subjects.
In patients with chronic hypertension, an increased incidence of early waveform
refection on carotid arterial waveform was evident. Using multiple regression
analysis, age, SBP and Murgo class of arterial waveform (type A, B and C aortic
pressure wave contours) are independent predictors of increased intima-media
thickness. Tice et al. concluded that conduit artery thickness is significantly
related to the degree of pressure elevation and the arterial waveform contour.
Again, in a study investigating arterial structure and function in 50 healthy
volunteers (aged 20 70), all indices of arterial compliance measured by
ultrasonography technique (DC), Doppler technique (systemic arterial
compliance) or applanation tonometry (pulse wave velocity and Al) were
significantly correlated with carotid intima-media thickness (Liang et al., 1998).
Similarly, Boutouyrie et al. (1999) demonstrated a prominent influence of
pulsatile mechanical load, as determined by carotid PP, on arterial remodeling. In
this study, carotid PP, a surrogate measure of arterial compliance was
associated with carotid internal diameter (r = .33) and intima media thickness (r =
.42) in healthy (n = 43) and untreated hypertensive subjects (n = 124). Kohara,
Jiang, Igase and Hiwada (1999) also demonstrated a significant positive
correlation between carotid Al assessed with a strain gauge transducer and
carotid intima-media thickness (r = .28, p = .018) in patients with essential
Despite the evidence of pressure-independent changes on arterial wall
properties, the exact nature of the alterations in the aortic structural properties
that lead to a decrease in aortic compliance with hypertension is still poorly
understood. Nevertheless, animal and human studies have confirmed the
association between decreased compliance and hypertrophic structural changes
in large arteries, suggesting an important role played by decreased large-artery
compliance in the pathogenesis of hypertension.
The role of arterial compliance in cardiovascular reactivity has not yet
been elucidated. As discussed earlier, decreased arterial compliance is one of
the mechanisms leading to resetting of baroreceptors. Resetting of baroreceptors
is a hypothesized link between stress hyperreactivity and hypertension.
Therefore, decreased arterial compliance may provide the link between
hyperreactivity and the development of hypertension.
Assessing Cardiovascular Reactivity
Measuring Blood Pressure Reactivity
Reactivity is defined as a change in physiological activity from a resting
baseline period to a task period (Turner, 1994). Blood pressure reactivity is
calculated by subtracting the baseline blood pressure value from the blood
pressure value in the task period. Both mean raw changes and percent change
from baseline have been reported in the literature (Turner, 1994). Since
anticipatory psychological arousal may affect pretask baseline values and
consequently the estimation of reactivity, a period of stabilization and relaxation
of 15 30 minutes is commonly adopted.
Blood pressure responses to stress were often measured by noninvasive
methods, which were employed either manually or automatically to sample blood
pressure at intervals (de Champlan, Petrovich, Gonzalez, Lebeau, & Nadeau,
1991; Sherwood et al., 1995; Nawarycz, Ostrowska-Nawarycz, & Kaczmarek,
1999). Sampling blood pressures at intervals may not be able to detect the
reactivity. It is suggested that cardiovascular responses to stressors have such a
dynamic time course that they can only be assessed by continuous intraarterial
blood pressure monitoring (Parati et al., 1991). Such an invasive measure
excludes a large number of participants and renders the issue of self-selection.
Also the subject's discomfort and distraction may confound the results of testing.
In this regard, Parati et al. suggested that a noninvasive apparatus that provides
continuous beat-to-beat blood pressure values is preferable.
Hemodynamic Measurements of Reactivity
Noninvasive hemodynamic measures, such as SV, CO, and TPR are
facilitated by impedance cardiography in studies on hemodynamic effects of
physiological and psychological interventions. Measuring thoracic impedance
changes during each cardiac cycle derives SV. CO is calculated from the
formula: SV x HR. TPR is derived from the formula: MAP CO, where MAP is
the mean arterial blood pressure derived from simultaneous measurement of BP.
Recording impedance-derived hemodynamic parameters during stress
can help to assess cardiac and vascular components of responses underlying BP
changes. The application of impedance cardiography is useful in differentiating
underlying cardiovascular changes in blood pressure responses between two
stress tasks and between two study groups performing an identical task (Parati et
Impedance cardiography involves the creation of an electrical field in the
chest by passing a small current between four electrodes placed around the neck
and lower thorax (Kubicek et al., 1974). Resistance associated with alternating
current is known as impedance. The change in impedance seen during each
cardiac cycle is indicative of the amount of blood flowing in the aorta and
therefore being ejected by the heart into the arterial tree. Greater decreases in
impedance are associated with greater amounts of blood.
The impedance cardiograph generates the following through the electrical
signals: 1) the mean thoracic impedance between electrodes 2 and 3 (Zo); 2) the
impedance change during the cardiac cycle, delta Z (AZ); 3) the first derivative of
AZ with respect to time; and 4) the electrocardiogram (EKG). When blood is
ejected from the left ventricle, changes in electrical impedance can be monitored
electronically to obtain a measurement of SV using the following formula derived
by Kubicek et al. in 1974:
SV = p x (L/Zo)2 x T x (dZ/dt)min
where Rho (p) is the resistivity of blood, L is the mean distance between the two
inner electrodes in cm, Zo is the mean thoracic impedance between electrodes 2
and 3, T is the ventricular ejection time in seconds, and (dZ/dt)min is the
minimum value of dZ/dt occurring during the cardiac cycle in ohms per second.
The validity of the impedance-derived measurements has been
established by comparison with reference measures. White et al. (1990)
demonstrated a linear correlation between SV derived from impedance
cardiography using the Kubicek equation and SV derived from an
electromagnetic flowmeter in dogs at any level of hematocrit over a wide range of
SV. These findings were also supported by studies in humans and rabbits where
impedance-derived SV were compared simultaneously with Fick and
thermodilution-derived SV (White et al., 1990). A close agreement between
impedance derived-CO and CO measured by the thermodilution method in
humans has been reported in the absolute values (r = .85) and in the percent
change (r = .87) (Goldstein, Cannon, Zimilichman, & Keiser, 1986). A high
correlation (r = .82) has also been shown between impedance-derived SV and
SV measured by nuclear ventriculography in humans (Wilson, Sung, Pincomb, &
Lovallo, 1989). However, researchers have shown that impedance cardiography
tends to overestimate the absolute values of CO (Mehlsen et al., 1991).
Investigators have provided evidence that supports outstanding
reproducibility for impedance-derived measurements. SV measured by
impedance cardiography showed good reliability across and within days in the
same subjects (r = .92 ~ .96) (Wilson et al., 1989). In 1991 Mehlsen et al. studied
37 healthy subjects and 25 unmedicated patients with ischemic heart disease
and obtained highly reproducible measures within the same day (standard
deviation of the differences in CO = 0.12 L/min) and on different days (standard
deviation of the differences in CO = 0.45 L/min) based on a total of 270
measurements. Mehlsen et al. also reported low intraobserver variability
(standard deviation of differences in CO = 0.12 L/min).
Although impedance-derived measures do not reflect individual absolute
values, the sources of measurement error with impedance-derived measures are
assumed to be systematic (Sherwood et al., 1990). Hence, the applicability of
relative changes in these measures within subjects is generally accepted.
In conclusion, impedance cardiography is reliable in measuring changes in
CO and suitable for repeated measurements in studies on the hemodynamic
effects of physiological and psychological interventions.
Ambulatory Blood Pressure Monitoring
Overview of Ambulatory Blood Pressure Monitoring
Ambulatory BP monitoring has now become an established research tool
in clinical trials. This methodology limits threats to external validity and the effects
of observers on psychological and physiological response. In addition, ABP
provides various useful parameters, including the average 24-hour BP (SBP,
DBP, MAP, and PP) and HR, BP variability (the standard deviation of the
average 24-hour, daytime, and nighttime measures), diurnal BP and HR changes
(day-night BP and HR difference), and BP load (percentage of systolic and
diastolic readings greater than 140 and 90 mm Hg during the day or greater than
120 and 90 mm Hg during the night).
Several of these variables correlate with the extent of target organ
damage or risk of cardiovascular events. In the HARVEST study, the rate of
excretion of albumin was highly related to 24-hour BP (Palatini, 1999). In the
ELSA trial, the intima-media thickness of the common carotid artery correlated
significantly with average 24-hour SBP (r = .22, p < .0001), average 24-hour PP
(r = .31, p < .0001), 24-hour SBP variability (r = .11, p < .0001), and 24-hour PP
variability (r = .23, p < .0001) (Mancia, Giannattasio, Failla, Sega, & Parati,
1999). Verdecchia (2000) also indicated that ambulatory SBP, DBP, and PP
were independently and directly associated with cardiovascular risk.
Circadian Blood Pressure Variations
One of the most important parameters of ABP is the day-night BP
difference, or nocturnal BP reduction. Those subjects in whom a nocturnal BP
reduction (BP reduction from day to night) is 10% or greater are classified as
dippers, whereas nondippers are defined as those who show a reduction in BP of
less than 10% between the day and night (Mallion, Baguet, Siche, Tremel, & de
Gaudemaris, 1999). May, Arildsen, & Damsgaard (1998) indicated that the
nocturnal BP reduction calculated from individually defined day and night times
was larger than the fall calculated from every possible fixed day/night definitions
and concluded that the assessment of the nocturnal BP dipping should be based
on individually defined periods of day and night.
It has been postulated that the absence of nocturnal BP reduction is
related to more severe target organ damage, either left ventricular hypertrophy or
disease of major arteries, though the data remain inconclusive (Mallion et al.,
1999). In the study by Mallion et al. nondippers had a significantly higher
frequency of stroke and higher urinary excretion of albumin. The degree of
nocturnal BP was inversely related to cardiovascular morbidity (Verdecchia et al.,
The mechanisms responsible for differences between individuals who
show a nocturnal reduction in BP and those who demonstrate no such reduction
in BP are still unclear. Schillaci et al. (1996) studied a large population of
subjects with essential hypertension to determine the independent predictors of
diurnal BP changes. Schillaci et al. demonstrated that predictors of day-night BP
difference included age, clinical BP, diabetes, the reported duration of sleep,
smoking habit and working activity during ABP monitoring. The nocturnal BP
reduction decreased with age and the prevalence of nondippers was greater in
elderly than in younger subjects (Di Lorio, Marini, Lupinetti, Zito, & Abate, 1999).
Both the absolute value and the ratio of nocturnal BP reduction showed a
curvilinear correlation with age with the smallest reduction and the largest ratio
observed in the elderly (> 70 years) (Staessen et al., 1997).
Arterial physical properties have also been studied to determine their
association with nocturnal BP reduction. Arterial compliance as indexed by aortic
pulse wave velocity predicts nocturnal SBP reduction in normotensives (Asar et
al., 1996). Arterial compliance as estimated by pulse wave velocity
measurement and its relationship to nocturnal BP reduction have been studied in
a group of treated hypertensive patients on hemodialysis (Amar et al., 1997).
Results from this study indicated that pulse wave velocity was significantly higher
in nondippers. A stepwise regression analysis further revealed that pulse wave
velocity was one of the independent variables related to the lack of or inverse
nocturnal BP reduction.
Ambulatory Blood Pressure and Blood Pressure Reactivity
An important question with regard to reactivity observed in laboratory
settings is: Do those people who show exaggerated responses to laboratory
stressors also show large responses to stress in the natural environment during
normal daily activities? Combining laboratory stress research with 24-hour
ambulatory monitoring of blood pressure and heart rate has become a common
approach in reactivity research (Krantz & Wing, 1984). Ambulatory monitoring of
blood pressure and heart rate offers valuable information to describe
cardiovascular reactivity to naturally occurring stress. Ambulatory blood pressure
monitors can be programmed to measure blood pressure and heart rate every 30
to 60 minutes during a 24-hour period. Blood pressure variability (e.g., the
standard deviation of the 24-hour BP) can then be calculated. Greater variability
corresponds to heightened cardiovascular reactivity during naturalistic conditions
(Parati et al., 1991).
Two models of laboratory-field generalization have been adopted: (1) to
obtain ambulatory measures when a selected real-life stress occurs, or (2) to
obtain ambulatory measures when subjects go about their normal daily activity
(Turner, 1994). Given the complexity of the context in which stressors occur, and
the factors operating on subjects to produce the BP responses, subjects should
keep a brief diary, which describes circumstances each time a recording occurs
(Krantz & Wing, 1984). Behavioral information such as posture, physical activity,
consumption of caffeine, alcohol, and nicotine, and the subject's affective state
are of particular interest (Krantz & Wing, 1984; Turner, 1994). This behavioral
information is also useful in disentangling the influences resulting from factors
other than psychological stressors on cardiovascular responses (Turner, 1994).
Contradictory findings have been reported in studies using correlational
analyses to examine the relationship between psychologically induced laboratory
reactivity and reactivity obtained by ambulatory measurements (Fredrikson,
Blumenthal, Evans, Sherwood, & Light, 1989; Georgiades, Lemne, de Faire,
Lindvall, Fredrikson, 1996; Jern et al., 1995; & Seibt et al., 1998). Confounding
influences on ambulatory measures of cardiovascular activity may obscure this
Blood Pressure Hyperactivity to Stress in Borderline Hypertensives
Characteristics of Borderline Hypertensives
Borderline hypertension represents a transitional phase from
normotension to hypertension. The risk of developing hypertension is increased
in borderline hypertensives (Julius, 1986). Lindvall, Kahan, de Faire, Ostergren,
and Hjemdahl (1991) indicated that mild hypertension is associated with
enhanced arousal and cardiac activation during rest. Increased CO and HR have
been found in young borderline hypertensives. Young (age 40 and less)
borderline hypertensives have a higher resting Cl and HR than normotensive
age-matched controls (Lund-Johansen, 1986). Circulatory patterns in the history
of borderline hypertension exhibit a hyperkinetic state characterized by high Cl
and HR preceding a normokinetic state characterized by normal/low Cl and high
TPR index (Lund-Johansen, 1989). In a sample of 458 subjects who developed
hypertension four years after the initial screening in the Framingham heart study,
subsequent onset of hypertension was associated with initially increased HR
(men) and CI (both sexes) (Post, Larson, & Levy, 1994). However, none of the
hemodynamic variables of a hyperkinetic state was a significant predictor of the
development of hypertension after adjusting for age and baseline BP.
Petrin, Egan and Julius (1989) indicated that baseline arterial compliance
as indexed by the SV:PP ratio was significantly higher in a hyperkinetic
borderline hypertensive group in comparison to a normokinetic group. After the
administration of beta-adrenergic blockade, the intergroup difference in arterial
compliance was no longer significant, suggesting that the group difference in
arterial compliance is attributed to observed disparities in vascular beta-
adrenergic tone between hyperkinetic and normokinetic borderline hypertensives.
In 1993 Bergbrant, Hansson, and Jern further demonstrated that minimal
forearm vascular resistance was significantly higher in the normokinetic
borderline hypertensive group than that observed in the hyperkinetic borderline
hypertensive group, suggesting the presence of structure vascular changes (i.e.,
vascular hypertrophy) in the former.
Hemodynamic and Physiological Basis of Enhanced Pressure Reactivity
Pressure hyperreactivity to stress is characteristic of borderline
hypertensives. However, the hemodynamic determinants and physiological
mechanisms of reactivity to stress have not been fully elucidated. Elliason (1985)
demonstrated that borderline hypertensives showed a tendency towards
enhanced alpha-adrenergic vasoconstriction during isometric handgrip test when
compared with both established hypertensives and normotensives. An
enhancement of the hypothalamic defense reaction (increased BP and plasma
adrenaline levels) during mental stress was evident in borderline hypertensives
as compared to established hypertensives and normotensives, whereas physical
stress produced more similar responses among groups.
de Champlain et al. (1991) examined cardiovascular reactivity to isometric
exercise with M-mode echocardiography, phonocardiography, carotidography
and circulating catecholamines in 15 borderline and 42 mild hypertensives and
25 normotensives of either sex. These investigators found that sympathoadrenal
reactivity, pressor and chronotropic responses were similarly increased during
exercise in the two groups of hypertensives and normotensives. In
normotensives, an increase in cardiac contractility and performance accounts for
the increase in BP, whereas in both groups of hypertensives the increase in BP
is mainly associated with an increase in TPR.
In contrast, a recent study using the impedance cardiography to assess
cardiovascular reactions to a passive orthostatic test, a cold pressor test, and a
hyperventilation test of 30 borderline hypertensive young men and 29 age-
matched normotensive men found that CI, SV, MAP and left ventricle work index
were increased in borderline hypertensives accompanying normal TPR
(Nawarycz et al., 1999).
Schulte and Neus (1983) demonstrated increased BP, HR and CO and
slightly decreased TPR during a mental arithmetic test in 20 hypertensive men
(10 borderline and 10 mild) as well as 10 age-matched normotensive men. The
rise in BP was greater in these two hypertensive groups than in the normotensive
group. The increase in BP also correlated with the CO elevation. Schulte and
Neus (1983) suggested that cardiovascular hyperreactivity in early hypertension
is attributed to increased beta-adrenergic stimulation of the heart.
Schulte, Ruddel, Jacobs, and von Eiff (1986) assessed cardiovascular
reactivity to mental stress in 41 borderline hypertensive men and 37 age-
matched normotensive men and found that the former group had higher absolute
or percentage of increased SBP and DBP during stress. In this study, the
increases in HR and CO and the decreases in TPR were higher in the borderline
hypertensive group as compared with the normotensive group, suggesting a
dominant cardiac reaction underlying the pressor response to mental stress in
Sherwood et al. (1995) examined cardiovascular reactivity to mental
stress (adverse reaction time test) that requires active coping and physical stress
(forehead cold pressor test) that elicits passive tolerance in borderline
hypertensive and age-matched normotensive men. These authors demonstrated
that the SBP increases were greater in borderline hypertensives when compared
with age-matched normotensives during the active task but not during the
passive task. During the active coping stress (mental stress), BP increases were
caused by marked augmentation of CO in borderline hypertensives despite
evidence of reduced cardiac beta-adrenergic receptor responsiveness.
Sherwood et al. indicated that the pressor hyperreactivity to a mental test in
borderline hypertensives was caused by an excessive rise in plasma epinephrine
and a consequently greater rise in CO. Similarly, Lindvall et al. (1991)
demonstrated a mental stress-induced CO-dependent increase in BP in both
normotensives and mild hypertensives. However, in the study by Lindvall et al.
smaller relative increases in DBP and MAP to mental stress were found in mild
hypertensives than in normotensives, suggesting that reduced SV
responsiveness caused by structural changes (i.e., left ventricular hypertrophy)
might account for this attenuated BP reactivity.
Sherwood et al. (1995) demonstrated that borderline hypertensives
reacted to passive coping (cold pressor test) with greater increases in TPR
possibly caused by vascular hypertrophy (as evidenced by significantly higher
minimal forearm vascular resistance) and an attenuation of CO in the presence
of normal baroreceptor reflex sensitivity. These changes lead to no greater net
increase of BP than that in normotensives. Other groups have also shown no
difference in the BP response to the cold pressor test between borderline
hypertensives and normotensives (Lafleche, Pannier, Laloux, & Safar, 1998;
Lindvall et al., 1991; & Ostergren et al., 1992).
Jern et al. (1995) studied hemodynamic responses to both mental and
physical stressors as well as daily-life ambulatory blood pressure in borderline
hypertensive men and normotensive control subjects from an unbiased
population of healthy young males. The group of borderline hypertensives were
further subdivided into hyperkinetic (CI greater than the mean + 1 standard
deviation of the normotensive control group) and normokinetic (CI below the
mean + 1 standard deviation of the normotensive control group) subgroups.
Resting MAP and HR were significantly higher in two borderline hypertensive
subgroups than those observed in the normotensive control group. Resting TPR,
forearm vascular resistance and calf vascular resistance were significantly lower
in the hyperkinetic borderline hypertensives than in the normotensive control
group. In contrast, TPR was higher in the normokinetic borderline hypertensive
group than in the normotensive control. These resting hemodynamics suggest
that in the hyperkinetic borderline hypertensive subgroup BP was elevated
through an increase in stroke volume, whereas in the normokinetic borderline
hypertensives BP was increased through an increase in TPR. SBP, DBP, and
MAP increased significantly during mental stress in all three groups. In
hyperkinetic borderline hypertensives, SBP, DBP, and MAP level during mental
stress were significantly greater than those observed in the normotensive control
group. Absolute and relative DBP reactivity were significantly greater in the
hyperkinetic borderline hypertensives than both the normokinetic borderline
hypertensives and normotensives. The stress-induced increase in CI was
significantly greater in the hyperkinetic borderline subgroup than in the
normotensive control group and tended to be higher than that in the normokinetic
borderline subgroup. Interestingly, TPR decreased in all groups during the
mental stress. While both hyperkinetic and normokinetic hypertensives exhibited
enhanced pressure reactivity to mental stress, the hyperkinetic hypertensive
subgroup demonstrated a greater nocturnal BP reduction than the latter
subgroup. Jern et al. concluded that enhanced stress reactivity in hyperkinetic
borderline reactivity is also reflected by greater day-night BP gradients and is not
explained by structural changes in the peripheral vasculature as determined by
forearm and calf minimal vascular resistance.
The choice of different stress tests by previous studies may explain the
disparity in results among studies that investigated hemodynamic determinants
of reactivity to stress. In addition, circulatory patterns are not uniformly identical
among all subjects with borderline hypertension. Therefore, different
hemodynamic patterns should have been taken into consideration when studying
stress reactivity of borderline hypertensives.
Physical Mechanisms of Enhanced Pressure Reactivity
As previously discussed, borderline hypertensives reacted to the cold
pressor test with greater increases in TPR (Sherwood et al., 1995) but no greater
increase in BP than that observed in normotensives (Lafleche et al., 1998;
Ostergren et al., 1992; & Sherwood et al., 1995). Sherwood et al. suggested that
the observed enhanced vascular reactivity to cold pressor stimulation might be
attributed to vascular hypertrophy evidenced in borderline hypertension.
In the study by Lafleche et al. (1998), arterial distensibility, compliance,
pulse wave velocity, and Al were assessed in 10 borderline hypertensives and 10
age- and sex- matched normotensive controls during the cold pressor test.
Arterial distensibility was determined by the relative change in diastolic diameter
during the systolic and diastolic period using the formula: 2 (Ds Dd) /PP, where
Ds is the systolic diameter, Dd is the diastolic diameter, and PP is pulse
pressure. Compliance was the absolute change in diastolic diameter during the
systolic-diastolic period and was calculated using the formula: (Ds-Dd) /2PP.
Pulse wave velocity was the delay from the foot of the pressure wave to the
inflection point representing the travel time for the incident wave to reach the
peripheral reflection sites and return.
In both groups, the cold pressor test induced a significant increase in BP
without a significant difference between groups. Both groups experienced an
increase in PP, a decrease in arterial distensibility and compliance and an
increase in pulse wave velocity and Al. The increase in Al was explained by the
increase in pulse wave velocity and the decrease in the distance from the
peripheral reflection site to the heart secondary to the cold pressor test-induced
vasoconstriction. The Al increased more in normal subjects than borderline
hypertensives, suggesting a different hemodynamic pattern between groups at
the central artery.
The decreases in carotid and brachial compliance were more pronounced
in normotensives than in borderline hypertensives. In normotensives, the cold
pressor test induced both an increase in PP and a decrease in pulsatile diameter
leading to a decrease in compliance in the brachial artery. In subjects with
borderline hypertension, only the increase in PP contributed to the decrease in
distensibility, thereby a smaller change in arterial compliance occurred. Lafleche
et al. (1998) maintained that in normotensives both a mechanical (increase in
PP) and a nonmechanical (decrease in pulsatile diameter) mechanism contribute
to the decrease in arterial distensibility during the cold presssor test, whereas in
borderline hypertensives only the mechanical mechanism (increase in PP)
account for the decrease in distensibility, presumably a consequence of a higher
basal sympathetic tone in the latter group.
Circulatory patterns in the history of borderline hypertension exhibit a high
cardiac output (hyperkinetic) state preceding a normal cardiac output
(normokinetic) state. Hyperkinetic borderline hypertensives have a higher resting
cardiac index than normotensive age-matched controls. Normokinetic borderline
hypertensives have significantly higher peripheral vascular resistance, whereas
hyperkinetic borderline hypertensives have significantly higher arterial
compliance, possibly due to enhanced beta-adrenergic tone in the latter
subgroup or the presence of vascular changes in the former. While both
hyperkinetic and normokinetic hypertensives exhibit enhanced pressure reactivity
to mental stress, the hyperkinetic hypertensive subgroup demonstrate a greater
nocturnal BP reduction than the latter subgroup.
The investigator speculated that increased pressure reactivity to stress in
hyperkinetic borderline hypertensives is caused by autonomic hyperreactivity
presumably due to enhanced beta-adrenergic tone. In contrast, increased
pressure reactivity to stress in normokinetic borderline hypertensives is caused
by increased peripheral vascular resistance, presumably due to structural
changes in the vasculature and/or partially explained by underlying decreased
arterial compliance. Hypothesized changes in hemodynamics underlying BP
reactivity to cognitive stress in normokinetic borderline hypertensives as
compared with normotensives are graphically displayed in Figure 2.3.
Figure 2.3. Hypothesized changes in hemodynamics underlying enhanced
pressure responses to cognitive stress of normokinetic borderline hypertensives
versus normotensives. A: change; T: increase; 1: decrease; HR: heart rate; SV:
stroke volume; TPR: total peripheral resistance; BP: blood pressure.
PROCEDURES AND METHODS
This study utilized a comparative design. Both the borderline hypertensive
group and the age-matched normotensive control group performed the Stroop
Color Word Test (SCWT). The dependent variables, blood pressure (BP),
augmentation index (Al), and the impedance-derived hemodynamic parameters
were measured before, during and after the SCWT. The design for the laboratory
experiment is graphically displayed as follows:
Group Baseline Stress Recovery
Borderline Hypertensives 01 02 03
Normotensive Controls 01 02 03
where O1 represents measurements of the dependent variables during the 6-
minute baseline period; 02 represents measurements of the dependent variables
during the SCWT; and 03 represents measurements of the dependent variables
during the 6-minute recovery period.
Population and Sample
The population under investigation was adult borderline hypertensives,
aged 18 to 65, in one North Central Florida County. Subject recruitment was
done through both flier advertising and BP screening. The investigator offered BP
screening several times a month for over 7 months at various locations.
Before BP was measured, potential subjects were told that they would be offered
a chance to take part in a research study if they qualified.
Blood pressure inclusion criteria was based on the classification of blood
pressure for adults, age 18 and older from the sixth report of the Joint National
Committee on Prevention, Detection, Evaluation, and Treatment of High Blood
Pressure (JNC- VI, 1997). Subjects were categorized as normotensive if SBP fell
below 130 mm Hg and DBP fell below 85 mm Hg measured on at least two
separate occasions. The "high-normal" and "hypertension-stage 1" groups
defined by the JNC-VI were selected as the BP criteria for inclusion to the
borderline hypertensive group. Borderline hypertension was therefore defined as
a resting DBP of 85 mm Hg to 99 mm Hg and/or a SBP of 130 mm Hg to 159 mm
Hg measured on at least two separate occasions.
Blood pressure was measured twice 2 minutes apart after the subject sat
quietly for 3 to 5 minutes. If the BP readings differed from each other by more
than five mm Hg, more measurements were taken. The initial screening BP was
obtained by averaging the two readings that agreed within 5 mm Hg. Subjects
whose initial screening BP met the inclusion criteria were invited to visit the study
site for a second screening to confirm their eligibility for inclusion in the study.
Specific Inclusion Criteria
Subjects selected were healthy individuals, age 18 to 65, free from overt
cardiovascular diseases, or other chronic or acute medical or psychiatric
Specific Exclusion Criteria
Subjects were excluded from the study if they were taking
antihypertensive or antidiuretic medicines, or if their history demonstrated
evidence of significant cardiovascular or renal diseases. There was no exclusion
of subjects from the study based on gender or race.
Hemodynamic Subgrouping of Borderline Hypertensive Subjects
The borderline hypertensive group was divided into two subgroups, the
hyperkinetic and normokinetic subgroups, on the basis of CI. Borderline
hypertensive subjects who had a Cl greater than the mean plus one standard
deviation of the normotensive control group were categorized as the hyperkinetic
borderline hypertensive subgroup (Jern et al., 1995). Those who had a Cl below
that level were referred to as the normokinetic borderline hypertensive subgroup.
Cardiac index was measured using the formula: CO/BSA, where BSA was the
individual's body surface area and CO was obtained from impedance
This study was conducted at a human research laboratory in the
University of Florida College of Nursing.
Research Variables and Instruments
Weight was measured in bare feet using a calibrated balance-beam scale.
Height was measured in bare feet using a wall-mounted height
Body Mass Index (BMI)
Body mass index (BMI) was calculated as weight (kg) divided by height
Body Surface Area (BSA)
The Impedance Cardiogram Microcomputer automated the computation of
BSA from the formula: BSA = (H 0.725) x (W 0.425) x (0.007184), where H is height
in cm and W is weight in Kg.
The Stroop Color Word Test (SCWT)
The independent variable in this study was the SCWT. The SCWT is
composed of a color-name reading task, a color-naming task, and an
interference task. The SCWT is printed on three pages of white paper and each
page has 100 items, presented in five columns of 20 items each. Page one is the
color-naming reading task and consists of the words "RED", "BLUE", and
"GREEN" printed randomly in black ink. Subjects were presented with this page
and instructed to read the words, going from top to bottom. Page two is the color-
naming task and consists of "XXXX" printed in red, blue, or green ink and
arranged randomly. Subjects were presented with this page and instructed to
name the colors of "XXXX" from top to bottom. The interference task consists of
the words on Page 1, printed in the colors on Page 2 and with no word or color
match. Subjects were handed Page 3 and were asked to ignore the printed word
and name the color of the ink each word was printed in.
The SCWT is associated with significant increases in respiratory rate,
plasma and urinary adrenaline, electrodermal activity (Tulen, Moleman, van
Steenis, & Boomsma, 1989), cardiac output (Hjemdahl et al., 1984), heart rate
(Hjemdahl et al., 1984;Tulen at al., 1989), and blood pressure (Hjemdahl et al.,
1984; Seibt at al., 1998). Such studies have validated the efficacy of the SCWT
in inducing sympathetic responses.
Beat-to-Beat Radial Blood Pressure
Continuous noninvasive BP was measured using radial artery tonometry
(Model 7000, Colin, San Antonio, TX) coupled with a computerized data
acquisition system (WINDAQ, Dataq Instruments, Akron, Ohio). Continuously
noninvasive BP provided beat-to-beat blood pressure values and a high-fidelity
arterial pressure waveform. The tonometric sensor contained pressure
transducers in an array. When a pneumatic pump and bellows press the
transducer array against the skin and tissue above the artery, the sensor records
the pressure wave. The radial tonometry is also equipped with an oscillometric
cuff, a pressure transducer with electronic processing to determine brachial SBP,
DBP and mean MAP. These pressure values are used to calibrate radial sensor
Continuously noninvasive BP measurements by radial artery tonometry
have been shown to offer a reliable trend indicator of pressure changes (Weiss,
Spahn, Rahmig, Rohling, & Pash, 1996).
Augmentation Index (Al)
Kelly et al. (1989) initially defined Al as the ratio of augmentation pressure
and PP expressed as a percentage. Augmentation pressure is defined as the
difference in pressure between the early and late systolic shoulders of central
aortic pressure waveforms. In this study the central aortic wave was synthesized
from a recorded peripheral wave recorded using a radial applanation tonometer
(Millar Pressure Tonometer, Millar Instruments) and a pulse wave analysis
system with a generalized transfer function (SCOR-Px/P, SphygmoCorTM pulse
wave analysis system, PWV Medical, Sydney, Australia). This device automates
the assessment of Al, expressed as a percentage, using the formula: Al = 100 x
(Ps Pi) / (Ps Pd), where Pi is the first systolic shoulder (inflection point), Ps is
the peak systolic pressure (SBP), and Pd is the minimum DBP (see Figure 3).
Using the transfer function to synthesize the central aortic wave from the
peripheral wave, agreement between the central aortic and peripheral pressure
wave is good (Karamanoglu et al., 1993). Linear relationships have also been
demonstrated between brachial blood pressures and corresponding central
pressures derived by the transfer function method (Karamanolglu et al., 1993).
Several studies (Liang et al., 1998; Seibenhofer, 1999; Wilkinson et al., 1998)
have previously reported excellent reproducibility for Al measurements with a
between-visit correlation coefficient of .98, standard deviation of intraobserver
measurement difference of 2.70% to 5.37%, and standard deviation of
interobserver measurement differences of 3.80%.
Pi / AP
Figure 3. Calculation of the augmentation index. The augmentation index
is calculated as the difference between Ps and Pi (AP), expressed as a
percentage of the difference between Ps and Pd (pulse pressure, PP). Tis
the time between the foot of the wave and the infection point, which
provides a measure of the travel time of the pressure wave to and from
the major reflection site.
In terms of validity, slight but significant correlations were found between
Al and aortic pulse wave velocity obtained via applanation tonometry in healthy
adults (r = .386, p < .01) (Liang et al., 1998) and adults with a parental history of
hypertension (r = .29, p, < .005) (Yasmin, & Brown, 1999).
Hematocrit was measured using the Micro-Capillary Reader (Damon / IEC
Division, Needham Hts, MA).
Impedance-derived Hemodynamic Parameters
Cardiac output was measured by the Minnesota impedance cardiograph
(MIC Model 304B, Surcom'nc, Minneapolis, MN). Three waveforms were
generated by the impedance cardiograph: (a) electrocardiogram (EKG); (b) basal
impedance (Zo); and (c) the first derivative of thoracic impedance (dZ / dt). The
impedance stroke volume in milliliters was calculated using the equation
developed by Kubicek et al. in 1974: SV = p x (L / Zo)2 x (VET) x (dZ / dt). Based
on the subject's hematocrit, the computer calculated the electrical resistivity of
blood Rho (p). VET is the left ventricular ejection time in seconds, as obtained
from the point where the impedance wave leaves the baseline to the sharp notch
or downward deflection in the wave. L is the front distance between the inner two
electrodes in cm. Zo is the mean thoracic impedance between the two inner
electrodes in ohms. dZ / dt is the maximum rate of change in impedance during
the ejection of blood from the left ventricle. dZ / dt was measured from baseline
to the maximum height of the waveform and recorded in ohms per second.
Stroke volume was calculated for each wave, and three to four waves were
averaged to obtain one value for each data collection point.
Heart rate in beats per minute was calculated from the R-R interval on the
EKG waveform. CO in liters per minute was calculated using the formula, CO =
(SV x HR) / 1000. TPR was estimated using the formula: TPR = MAP / CO x 80
(dynes-.cm-5), where MAP is the mean arterial blood pressure calculated using
the formula: (SBP + 2 DBP)/ 3.
Support for the validity of the impedance-derived measurements has been
well documented by comparison with reference measures. A close agreement
between impedance derived-CO and CO measured by the thermodilution method
in humans has been reported in the absolute values (r = .85) and in the percent
change (r = .87) (Goldstein et al., 1986). A high correlation (r = .82) has also
been shown between impedance-derived SV and SV measured by nuclear
ventriculography in humans (Wilson et al., 1989).
Mehlsen et al. (1991) have provided evidence that supports outstanding
reproducibility for impedance-derived measurements with low measurement
differences on the same day (standard deviation of the differences in CO = 0.12
L/min) and on different days (standard deviation of the differences in CO = 0.45
L/min), and low intraobserver variability (standard deviation of differences in CO
= 0.12 L/min).
Arterial compliance was estimated by the SV/PP ratio. PP was calculated
by subtracting aortic DBP from aortic SBP. The aortic pressure wave was derived
from a radial pressure wave recorded using a radial applanation tonometer with a
generalized transfer function. SV was measured using the impedance
cardiography. The SV/PP ratio has been demonstrated to be a valid estimate of
total arterial compliance (Chemla et al., 1998; Randall et al., 1986; Resnick et al.,
Ambulatory Blood Pressure (ABP) Parameters
An autonomic noninvasive cuff-oscillometric recorder (Model 90207,
SpaceLabsTM Inc., Redmond, WA) measured ambulatory blood pressure. This
monitor measures BP by detection of oscillations transmitted from the brachial
artery to the cuff. The SpaceLabsTM monitor was equipped with four different size
adult cuffs. A SpaceLabsTM Model 9029 Data Interface Unit was used for report
Average pressures and standard deviations were calculated. Ambulatory
BP variability was defined as the standard deviation of the 24-hour APB average.
The day-night BP difference, or nocturnal BP reduction was defined as the
difference between the nighttime pressure average (NP) and the daytime
pressure average (DP), using individually defined periods of sleep and wake time
parameters. Percentages of nocturnal BP reduction were computed by the
formula: 100 % x (DP-NP).
The validity of the SpaceLabs 90207 ABP monitor has been established.
According to the validation protocols of British Hypertension Society and the
Advancement of Medical Instrumentation, the SpaceLabs 90207 satisfies the
criteria for accuracy (O'Brien, Atkins, & Staessen, 1995).
Clinic Blood Pressure and Heart Rate
Clinic brachial BP and HR were measured manually with the ABP monitor
(SpaceLabs 90207, SpaceLabsTM Inc., Redmond, WA) that was used in the
recording of 24-hour ABP. The ABP monitor was programmed to show readings
of SBP, DBP, and HR. Multiple resting measurements 2 minutes apart were
taken 5 minutes after subjects were seated. The average of the two readings that
agreed within 5 mm Hg was defined as clinic BP. The average of the HR
recorded from the same set of BP readings selected was defined as clinic HR.
Laboratory Session Protocol
The subject was instructed to keep a regular sleep and wake pattern and
to avoid abnormal physical exertion and psychological stress starting 3 days
before the laboratory session. Alcohol or caffeine containing beverages and
cigarette smoking was prohibited before or during the laboratory session. All
subjects signed a consent document before any measurements were made.
All subjects filled out a health history inventory consisting of information
about major medical history, life style, and family history of hypertension. Clinic
BP and HR were measured manually with the ABP monitor 5 minutes after
subjects were seated in a quiet room.
The subject was assessed individually in the laboratory session.
Consistency between subjects was assured by using a standardized test
administration protocol. The test administration was scripted. Persons
administering the test were trained by the investigator. All instruments were
calibrated prior to starting data collection.
Height and weight were measured in bare feet. One drop of capillary blood
was collected by the finger-stick method in a capillary tube, and centrifuged for 5
minutes for hematocrit determination. The investigator placed four pieces of
aluminum-coated Mylar tape around the subject's neck and chest, which were
attached to the electrodes of the MIC. Four pairs of impedance electrodes were
attached to the aluminum surface of the Mylar tape. The inner two electrodes
(Electrodes 2 and 3), or voltage electrodes, were used to record the EKG, basic
impedance of the thorax (Zo), and the small impedance change that occurred
during the cardiac cycle, and its first derivative (dZ/dt). The outer two electrodes
(Electrodes 1 and 4), or the current electrodes, were connected to a constant
current source (4mA, 100 kHz). Electrode 2 was placed at the base of the neck
as close as possible to the clavicles, and electrode 3 was placed at the
xiphisternal joint. Electrodes 1 and 4 were placed at least 3 to 5 cm above and
below the voltage electrodes
The sensor of the CNIBP radial tonometry was placed over the preferred
sensor site of the subject's left radial artery with proper hold down pressure and
optimal signal strength indicated on the tonometry screen. A pressure cuff was
applied to the ipsilateral arm over the brachial artery for initial and intermittent
calibration of the radial tonometry during the recording of CNIBP. Al was
obtained from measuring the pressure waveform with a radial applanation
tonometer applied to the right arm.
Before testing a 30-minute adaptation and resting period was employed to
accustom the subject to the laboratory environment. Subjects were instructed to
relax in the seated position with their eyes open for the 30-minute adaptation and
resting period. A 6-minute baseline period was followed by the 6-minute SCWT.
After a brief introduction to the test the subject performed the SCWT. A 6-minute
recovery period followed the SCWT.
Blood pressure waves were continuously recorded with the CNIBP radial
tonometry coupled with a computerized data acquisition system during the 6-
minute baseline period, the SCWT test period, and the 6-minute recovery period.
The peak and valley of each BP wave were captured and transformed into
numerical data using a computerized data browsing system (WINDAQ, Dataq
Instruments, Akron, Ohio). The average BP and standard deviation for each
period were then calculated.
Multiple measurements of impedance and Al were obtained before,
during, and after the SCWT. Three to four sitting impedance measurements
during the baseline period, one measurement during each page of the SCWT
(three during the stress period), and three to four measurements during the
recovery period were obtained by a trained research assistant.
Likewise, three to four sitting Al measurements during the baseline period,
one measurement during each page of the SCWT (three during the stress
period), and three to four measurements during the recovery period were
obtained by the investigator. The investigator made sure that each impedance
measurement was obtained simultaneously with each Al measurement. A list of
the variables measured or calculated during the laboratory session is provided in
Ambulatory Blood Pressure Monitoring Protocol
The subject was instructed to keep a regular sleep and wake pattern and
to avoid abnormal physical exertion and psychological stress during the ABP
recording day. Ambulatory BP was recorded on an ordinary workday or school
Subjects were fitted with an ABP monitor and familiarized with its
operation. The monitor was programmed to measure blood pressure over a 24-
hour period at the frequency of every 30 minutes during the day and evening (6
AM to 10 PM) and every 60 minutes during the sleep time (10 PM to 6 AM). A
cuff of the proper size, determined by upper arm circumference was placed on
the subject's nondominant arm and attached by flexible tubing to the monitor.
The center of the inflatable bladder of the cuff was placed directly over the
brachial artery. The investigator inserted a finger between the cuff and the arm to
ensure the cuff was not too tight. The monitor was then strapped to the patient on
Rigorous calibration of the equipment was made prior to ABP monitoring.
A calibration procedure comprised of 3 calibration readings in the seated
position, each taken with the ABP monitor and a mercury column simultaneously
by means of a "T" connector between two instruments. Readings for both SBP
and DBP agreed within 5 mm Hg on all 3 attempts.
The monitor emits a series of alarm sounds 5 seconds prior to cuff
inflation. The investigator instructed the subjects to keep the limb quiet and allow
their arm to hang freely at their side during cuff inflation and deflation. To avoid
reading errors due to hydrostatic pressure differences, the patient was instructed
to keep the level of the cuff near the heart level. Subjects were given a diary and
asked to record their activities each time the monitor recorded a pressure.
Subjects were asked to return the monitor and the diary to the investigator after a
minimum elapsed time of 24 hours. The 24-hour data for each subject were then
downloaded, printed out, and compared with the subject's diary entries. The
individually defined periods of sleep and wake time indicated on the diary were
used to compute the subject's nocturnal blood pressure reduction.
Maior Variables Measured or Calculated During the Laboratory Session
Systolic BP (SBP)
Diastolic BP (DBP)
Pulse Pressure (PP)
Mean Arterial Blood
Aortic augmentation index
First systolic shoulder (Ps)
Second systolic shoulder
Resistivity of blood (p)
Basal impedance (Zo)
Distance between the two
inner electrodes (L)
Left ventricle ejection time
First derivative of thoracic
Stroke volume (SV)
Cardiac output (CO)
Body Surface Area (BSA)
Cardiac Index (CI)
Total peripheral resistance
Measured or Calculated
PP = SBP DBP
MAP = (SBP +2DBP) + 3
Al (%) = (Ps Pi) + (SDB -
Calculated from blood
Calculated from R-R interval
Measured over chest
SV (ml) = p x (L / Zo)2 x
(VET) x (dZ / dt)
CO (L) = (SV x HR)+ 1,000
BSA (m2) = (H 0.725) x (W
0.425) x (0.007184)
CI (L / m2) = CO + BSA
TPR= (MAP + CO) x 80
Methods of Statistical Analyses
Data were analyzed using SAS (SAS Institute Inc, Cary, North Carolina).
Descriptive statistics were used to obtain the summary measures for the data.
Addressing Hypotheses 1 and 2, Repeated Measures Analysis of Variance was
performed, using a mixed linear model, to determine the changes in response
variables within subjects, between subjects, and within-subjects-by-between-
subjects interactions. Student t-test was performed to address Hypotheses 3, 4
5, and 6. Simple and multiple regression analyses were used to address
Hypotheses 7, 8, and 9. The significance level of .05 for rejecting the null
hypothesis was chosen.
The primary purpose of this study was to determine the differences in
aortic reflection wave amplitude as estimated by the augmentation index (Al),
arterial compliance as estimated by the SV/PP ratio, and hemodynamic patterns
underlying pressure responses to cognitive stress in normokinetic borderline
hypertensives and age-matched normotensives. The secondary purpose was to
compare the differences in ambulatory blood pressure parameters between
borderline hypertensives and normotensives. The relation between arterial
compliance and nocturnal BP reduction was also investigated in this study.
Finally, the predictive power of pressure reactivity to stress on ambulatory BP
(ABP) parameters was examined.
This chapter will first present descriptive results, including means,
standard deviations, and frequency data for each variable. The nine hypotheses
posed in Chapter 1 will be addressed using repeated measures analyses of
variance, Student's t-test, and regressional analyses.
Over 60 potential subjects met the specific inclusion criteria in the initial
screening and were invited for a second screening to be included in the study.
However only 46 subjects actually showed up for the second screening. One
subject was excluded because she was taking antihypertensive medications.
Three subjects did not meet the BP inclusion criteria in the second screening.
A total of 42 subjects participated in this study. Nineteen subjects
comprised the normotensive control group and 23 comprised the borderline
hypertensive group. Of 23 borderline hypertensives, 20 were normokinetic and 3
Subject demographics expressed in numbers and percentage were
gender, race, and age, family history of hypertension, and habits of cigarette
smoking and exercise. Table 4.1 shows the subject demographics, providing a
comparison among groups. For completeness, both the borderline hypertensive
and normokinetic borderline hypertensive groups are shown. The borderline
hypertensive group includes the normokinetic borderline hypertensive group.
The normotensive group ranged in age from 20 to 63 years with a mean of
40.5 years. The age of the borderline hypertensive group ranged from 24 to 64
years with a mean of 44.4 years. The age of the normokinetic borderline
hypertensives ranged from 24 to 64 years with a mean of 46.4 years. As shown
in Table 4.1, the borderline hypertensives and normotensives were comparable
in gender, race, age, and family history of hypertension, and habits of smoking
The clinical characteristics, including weight, height, BMI, clinic BP and
heart rate of the borderline hypertensive and normotensive subjects studied are
presented in Table 4.2.
Demographic Data for the Borderline Hypertensive and Normotensive Subjects
BH (N = 23) NBH (N = 20) Normotensive
N % N % N %
Male 9 39 7 35 8 42
Female 14 61 13 65 11 58
Asian 1 4 0 0 2 11
African-American 4 18 4 20 1 5
Hispanic 1 4 1 5 0 0
Caucasian 17 74 15 75 16 84
30 and under 3 13 1 5 5 26
31-40 5 22 4 20 4 21
41-50 9 39 9 45 6 32
51 60 4 17 4 20 2 10.5
61-65 2 9 2 10 2 10.5
Family History of
Yes 17 74 15 75 13 68
No 6 26 5 25 6 34
Yes 6 26 5 25 3 16
No 17 74 15 75 16 84
Yes 12 52 9 45 14 74
No 11 48 11 55 5 26
BH: borderline hypertensive, NBH: normokinetic borderline hypertensive.
Table 4.2. Clinical Characteristics of the Borderline Hypertensive and
Normotensive Subjects Studied
BH Group NBH Group Control Group
(N = 23) (N = 20) (N = 19)
Weight (kg) 89.9 + 18.4** 93.0 + 17.46** 72.0 + 16.05
Height (cm) 171.7 10.5 171.60+ 11.07 171.78.34
BMI (kg / m2) 30.7 + 6.9** 31.7 + 6.80** 24.1 +3.60
SBP (mm Hg) 142.6 + 7.4** 142.4 + 7.4** 117.3 +9.2
DBP (mm Hg) 88.2 + 9.3** 88.4 + 9.4** 69.6 + 6.7
HR (beats / min) 79.1 + 11.1** 80.3 + 10.2** 67.7 +9.5
Values are expressed as means SD. *p < .05, **p < .01, versus normotensive
controls, by t-test. BH: borderline hypertensive, NBH: normokinetic borderline
hypertensive, BMI: body mass index, SBP: systolic blood pressure, DBP:
diastolic blood pressure, HR: heart rate.
Clinic BP and HR were significantly higher in the borderline hypertensive
subjects than in the control subjects (refer to Table 4.2). The borderline
hypertensive subjects and the normotensive subjects also differed significantly in
weight and BMI. Therefore BMI was treated as a covariate in the repeated
measures analysis of variance.
Resting hemodynamics during the baseline period are summarized in
Analytic Results for Hypotheses
For Hypotheses 1 and 5, the pattern of BP reactivity to cognitive stress
and BP reactivity during everyday life in the borderline hypertensive group was
compared to the normotensive group. For Hypotheses 2, 3, 4, hemodynamic and
arterial responses underlying BP reactivity to stress was examined in the
normokinetic borderline hypertensive subgroup and compared to the
normotensive group. For Hypothesis 6, the degree of nocturnal BP reduction in
the normokinetic borderline hypertensive subgroup was compared to the
normotensive group. Hypotheses 7 and 8 tested the predictive power of
augmentation index and arterial compliance, respectively for the day-night BP
difference in the whole study sample. Lastly, Hypothesis 9 tested the predictive
power of BP reactivity for ambulatory BP parameters in the whole study sample.
Table 4.3. Resting Hemodynamics of the Borderline Hypertensive and
Normotensive Subiects Studied
BH Group NBH Group Control Group
(N = 23) (N = 20) (N = 19)
SV (ml) 62.8 + 34.3 52.7 + 22.0** 74.3 22.2
CO (L / min) 4.5 +2.0 4.0 +1.5 4.8 +1.2
CI (L / min / m2) 2.2+ 1.1 1.9 + 0.8** 2.6 0.60
TPR(dynes-s-cm-5) 2102.2 +637.2** 2196.0 +876.1** 1298.5 + 388.0
Values are expressed as means SD. *p < .05, **p < .01, versus normotensive
controls, by t-test. BH: borderline hypertensive, NBH: normokinetic borderline
hypertensive, CO: cardiac output. CI: cardiac index, TPR: total peripheral
Hypothesis 1. Borderline hypertensives have a greater increase in BP in
response to cognitive stress (SCWT) as compared to age-matched normotensive
subjects. Repeated measures analysis of variance was performed using a mixed
linear model (PROC MIXED). A mixed linear model is a generalization of the
standard linear model. In a mixed model, the data are permitted to exhibit
nonconstant variability and correlation, thereby providing the flexibility of
modeling means, variances, as well as covariances. The mixed model is
therefore preferable to other standard linear models, because it utilizes not only
the response means but also their variances and covariances (Little, Miliken,
Stroup, & Wolfinger, 1996).
There was a significant change in SBP within subjects across
experimental periods with the SCWT (F = 9.86 p = .0001) and a significant
difference in SBP between the two groups (F = 97.60, p < .0001). However, the
pattern of change in SBP was similar between the borderline hypertensive and
normotensive control group (i.e., no time and group interaction, F = 0.42, p =
.6580). The effect of BMI on SBP was statistically significant (F = 11.61, p =
Likewise, there was a significant difference in DBP between the two
groups (F = 41.07, p < .0001) and a significant change in DBP within subjects (F
= 4.21, p = .0165) with the SCWT. However, there was no time and group
interaction effect on DBP (F = 0.41, p = .6667).
There was a significant difference in PP between the two groups (F =
20.95, p < .0001). However, there was no time effect (F = 2.02, p = .01281) and
no time and group interaction effect on PP (F = 0.09, p = .9121). The effect of
BMI on PP was significant (F = 3.2, p = 0.0765).
Furthermore, there was a significant difference in MAP between the two
groups (F = 67.64, p < .0001) and a significant change in MAP within subjects
with the SCWT (F = 7.28, p = .001). In addition, the effect of BMI on MAP was
statistically significant (F = 6.23, p = .0139). Again, there was no time and group
interaction effect on MAP (F = 0.32, p = .7284).
The means and standard deviations of BP during each experimental
period are summarized in Table 4.4. Analysis revealed that all baseline BP
measurements, including SBP, DBP, MAP, and PP were significantly higher in
the borderline hypertensive group than the normotensive group. Both groups had
significant increases in SBP, DBP, MAP, and PP from baseline to stress. SBP
and MAP decreased significantly from stress to recovery in both groups. DBP
and PP decreased significantly from stress to recovery in the normotensive group
but not in the borderline hypertensive group. The changes in SBP, DBP, and PP
across the three experimental periods in the two study groups are graphically
presented in Figure 4.1, 4.2 and 4.3.
Hypothesis 2. Normotensives and normokinetic borderline hypertensives
differ significantly in their responses in HR, SV, and TPR to cognitive stress. For
Hypothesis 2, repeated measures analysis of variance was performed using
PROC MIXED. There was a significant difference in SV between the two groups
(F = 9.84, p < .0001). However, there was no significant change in SV within
subjects (F = 0.12, p = .8851) and no time and group interaction effect on SV (F
= 0.01, p = .9920).
There was a significant difference in HR between the two groups (F =
5.06, p < .0265) and a significant change in HR within subjects with the SCWT (F
= 12.62, p < .0001). Additionally, the effect of BMI on HR was significant (F =
3.95, p = .0497). Again, there was no time and group interaction effect on HR (F
= 1.18, p = .3112). The change in HR across the three experimental periods in
the two study groups is presented in Figure 4.4.
Lastly, there was a significant difference in TPR between the two groups
(F = 6.18, p = .0145). In addition, the effect of BMI on TPR was significant (F =
14.71, p = .0002). However, there was no significant change in TPR within
subjects with the SCWT (F = 1.95, p = .1475) and no time and group interaction
effect on TPR (F = 0.48, p = .6185).
Hypothesis 3. Normotensives and normokinetic borderline hypertensive
subjects differ significantly in wave reflection amplitude as indexed by Al.
Student t-test was used to address Hypothesis 3. The resting Al was
16.9% + 16.8% in the normotensive subjects and 22.1% + 11.9% in the
normokinetic borderline hypertensive subjects. The difference in Al between
these two study groups was not statistically significant (t = -1.10, p = .28).
Table 4.4. Summary of Beat-to Beat Radial Blood Pressure During the Stress
Group / BP mm Hg Baseline Stress Recovery
SBP 133.7 +12.3 144.2 +13.0 138.1 + 11.5
DBP 72.6+ 10.4 78.5+ 11.0* 75.5 +10.5
MAP 93.4 +10.7 101.2 +11.0 96.5 + 10.1*
PP 61.1 + 10.0 65.7 + 12.7** 62.5+ 10.0
SBP 105.0+11.9 117.1 +13.6 106.2+ 13.6
DBP 56.6 +10.0 63.9 + 11.5** 57.0 + 12.0*
MAP 72.8 + 9.5 82.0 + 11.4** 73.6 + 12.2*
PP 48.4 +9.4 53.3 + 9.5 49.5 + 9.1
Values are expressed as means + SD, p < .01 versus control group, p < .01
versus baseline, p < .01 versus stress, by t-test. BH: borderline hypertensive,
SBP: systolic blood pressure, DBP: diastolic blood pressure, MAP: mean arterial
blood pressure, PP: pulse pressure.
CO 110, *** Baseline
100 N M Recovery
Figure 4.1. Systolic blood pressure changes across the three experimental
periods. NC: normal controls, BH: borderline hypertensives. p < .01
versus baseline, p < .01 versus stress, by t-test.
Figure 4.2. Diastolic blood pressure changes across the three
experimental periods. NC: normal controls, BH: borderline hypertensives.
p < .01 versus baseline, p < .01 versus stress, by t-test.
Figure 4.3. Pulse pressure changes across the three experimental
periods. NC: normal controls, BH: borderline hypertensives. p < .01
versus baseline, p < .01 versus stress, by t-test.
Figure 4.4. Heart rate changes across the three experimental
periods. NC: normal controls, NBH: normokinetic borderline
hypertensives. p < .01 versus baseline, p < .01 versus stress,
Hypothesis 4. Normotensive and normokinetic borderline hypertensive
subjects differ significantly in arterial compliance as indexed by SV/PP.
Student t-test was used to test Hypothesis 4. The resting arterial
compliance was 2.25 + 0.85 (ml/mm Hg) in the normotensive group and 1.18 +
0.48 (ml/mm Hg) in the normokinetic borderline hypertensive subjects (See
Figure 4.5). The normotensive group had significantly higher arterial compliance
than the normokinetic borderline hypertensive group (t = 4.9, p < .0001).
Hypothesis 5. Borderline hypertensives exhibit a greater ABP variability as
determined by a greater standard deviation of the average 24-hour BP, as
compared to the age-matched normotensives.
Student t-test was used to address Hypothesis 5. The results of the
comparison of 24-hour, daytime, nighttime BP parameters and BP variability
measures between the two groups using a t-test are shown in Table 4.5.
The standard deviations for 24-hour and daytime SBP were significantly
higher in the borderline hypertensives than in the control group. The standard
deviations for 24-hour, daytime, and nighttime DBP were significantly higher in
the borderline hypertensives than in the control group. Likewise, the standard
deviations for 24-hour, daytime, and nighttime MAP were significantly higher in
the borderline hypertensives than in the control group. In addition to the
variability measures, the two groups differed significantly in all the BP measures
as shown in Table 4.5. Of course if the Bonferroni method of controlling the
overall error was used, the new level of significance will be .003. Based on this
new level of significance, none of the ABP variabilities were different between
Hypothesis 6. Normotensives exhibit a significantly greater day-night BP
difference as compared to normokinetic borderline hypertensives. Student t-test
was used to test Hypothesis 6.
The two groups did not differ significantly in the day-night SBP difference
(t = -0.90, p = .3765), day-night DBP difference (t = -0.90, p = .3765), and day-
night MAP difference (t = -1.46, p = .1539). Additionally, the day-night HR
difference was not significantly different between the two groups (t = -0.44, p =
.6599). The results of the day-night BP and HR difference are shown in Table
Hypothesis 7. Aortic wave reflection amplitude as indexed by Al is
predictive of the day-night BP difference. Regression analysis was used to
address Hypothesis 7.
There was no significant relationship between Al and the day-night SBP
difference (R2 = .0010, p = .8417). Additionally, Al did not explain the variations
in the day-night DBP difference (R2 = .0006, p = .8793) and day-night MAP
difference (R2 = .0018, p = .7930).
Hypothesis 8. Arterial compliance as indexed by SV/PP is predictive of the
day-night BP difference. Regression analysis was used to address Hypothesis 8.
The SV/PP ratio did not explain the variations in the day-night SBP
difference (R2 = .205, p = .197). Likewise, the variations in the day-night DBP
difference was not explained by the level of SV/PP (R2 = .121, p = .452).
N 19 20
Figure 4.5. A comparison of arterial compliance between the normotensive
control group and the normokinetic borderline hypertensive group. NC:
normal controls, NBH: normokinetic borderline hypertensives. Values
expressed as means + SD, p < .01 versus normal controls, by t-test.
Table 4.5. Group Comparison of Ambulatory
Blood Pressure and Variability
ABP and Variability Control Group BH Group t p-value
Parameters (mm Hg) (N = 19) (N = 23)
Mean + SD Mean + SD
24-hour SBP 118.0+9.4 136.6 +10.4 -5.92 < .0001
24-hour SBPV 10.5 +3.1 13.5 +3.2 -3.00 .0046
Daytime SBP 121.2 + 9.4 140.4 +10.4 -6.02 < .0001
Daytime SBPV 7.6 + 1.9 11.0 +2.6 -2.09 .0433
Nighttime SBP 108.2 +9.9 122.6 +13.4 -3.80 .0005
Nighttime SBPV 7.4 +2.3 12.3+ 14.6 -1.56 .1330
24-hour DBP 71.3 + 5.9 83.2 + 9.9 -4.82 < .0001
24-hour DBPV 9.4 +2.3 11.8 +2.9 -2.83 .0073
Daytime DBP 74.4 + 6.3 86.9 + 9.7 -4.95 < .0001
Daytime DBPV 7.6+ 1.9 9.5 +3.2 -2.32 .0260
Nighttime DBP 61.7 +4.5 70.9 11.4 -3.53 .0014
Nighttime DBPV 6.7 +2.8 9.1 +3.2 -2.55 .0149
24-hour MAP 87.0 +5.8 101.5 +9.6 -5.64 < .0001
24-hour MAPV 9.2 +2.6 12.0 +3.0 -3.15 .0031
Daytime MAP 89.9 +6.5 105.2 + 9.5 -6.30 < .0001
Daytime MAPV 9.2 +2.6 9.7 + 3.1 -2.26 .0295
Nighttime MAP 78.0 + 5.5 88.5 +11.5 -3.85 .0005
Nighttime MAPV 6.9 + 2.4 9.0 + 3.4 -2.27 .0288
ABP: ambulatory blood pressure, BH: borderline hypertensive, SBP: systolic
blood pressure, DBP: diastolic blood pressure. MAP: mean arterial pressure,
SBPV: systolic blood pressure variability, DBPV: diastolic blood pressure
variability. MAPV: mean arterial pressure variability. p < .05, p < .003.
Table 4.6. Group Comparison of Nocturnal Blood Pressure and Heart Rate
Day-Night Difference Control Group NBH Group t p-value
(mm Hg) (N= 19) (N = 20)
Mean + SD Mean + SD
SBP 13.0 +6.5 15.6 +11.2 -0.90 .3765
DBP 12.8 +5.3 14.7 +7.5 -0.93 .3605
MAP 11.8 6.0 15.2 8.0 -1.46 .1539
HR 10.5 6.5 11.3 4.2 -0.44 .6599
NBH: normokinetic borderline hypertensive, SBP: systolic blood pressure, DBP:
diastolic blood pressure. MAP: mean arterial pressure
Hypothesis 9. Pressure reactivity is predictive of the average 24-hour and
daytime BP. The reactivity score was calculated by subtracting the baseline from
the stress score. The recovery score was calculated by subtracting the recovery
from the stress score. Multiple regressions were used to test Hypothesis 9. Each
ambulatory BP measure was regressed on all reactivity and recovery measures,
clinic BP and HR, and BMI. The stepwise regression procedure in multiple
regression analysis was performed to determine the optimal number of variables
necessary to account for explaining the maximum variance in the response
Daytime SBP was significantly explained by DBP reactivity (p = .0073),
DBP recovery (p < .0001), MAP reactivity (p = .0045), MAP recovery (p < .0001),
HR reactivity (p = .0494), SV reactivity (p = .0201), CI recovery (p = .0112), and
clinic SBP (p < .0001) with an R2 of .87.
Daytime DBP was significantly explained by DBP recovery (p = .0002),
MAP reactivity (p = .0337) and recovery (p = .0001), clinic DBP (p < .0001) and
BMI (p = .0126) with an R2 f .86.
In addition, the variation in daytime MAP was explained by DBP recovery
(p < .0001), MAP reactivity (p = .0286) and recovery (p < .0001), clinic SBP (p =
.0002) and DBP (p = .0003), and BMI (p = .0262) with a R2of .86.
Furthermore, the variations in 24-hour SBP were significantly explained
(R2 = .86) by reactivity (p = .0458) and recovery DBP (p < .0001), reactivity (p =
.0284) and recovery MAP (p < .0001), and clinic SBP (p < .0001). The variations
in 24-hour DBP were significantly explained (R2 = .86) by recovery DBP (p =
.0458), reactivity (p < .0001) and recovery MAP (p = .0284), clinic DBP (p <
.0001), and BMI (p < .0001). The variations in 24-hour MAP were significantly
explained (R2= .84) by recovery DBP (p < .0001), reactivity (p = .0473) and
recovery MAP (p < .0001), clinic SBP (p = .0027) and DBP (p = .0001).
Augmentation index. A gender comparison of resting Al was performed
using the t-test. Female subjects had significantly higher (t= -2.46, p = 0.019) Al
values than male subjects (24.6% + 14.5% and 13.7 % + 13.8 %, respectively).
Further analysis revealed that female subjects had significant lower body height
than male subjects (t = 8.11, p < 0.0001). There was no difference in resting Al
between subjects with and without family history of hypertension (t = 1.023, p =
.313). Again, the difference in resting Al between regular exercisers and
nonexercisers was not statistically significant (t = 0.320, p = .750).
Correlational analyses were performed to examine possible correlates of
Al. Age significantly correlated with resting Al (r = .58, p < .0001). Height and Al
had a weak but significant inverse relationship (r = -.33, p = .036). SV and Al was
inversely correlated (r = -.40, p = .008). The correlation between the resting HR
and Al was not statistically significant (r= -.03, p = .835). However, the correlation
between resting HR and Al was barely significant in the normokinetic borderline
hypertensive subjects (r = -.43, p = .056). Figure 4.6 shows the correlation
between Al and age. Figure 4.7 shows the correlation between Al and height.
The correlation between SV and Al is presented in Figure 4.8. The gender
comparison of Al is presented in Figure 4.9.
The study subjects were further divided into dippers (the day-night MAP
difference equal to or greater than 10 mm Hg) versus nondippers (the day-night
MAP difference less than 10 mm Hg). Thirty-one subjects were categorized as
dippers (74%) and 11 subjects were categorized as nondippers (26%). A
comparison of resting Al between the dippers and nondippers was performed
using a t-test. There was no significant difference in Al between the dippers and
nondippers (t = 0.43, p = .6664).
A comparison of the change in Al from the baseline to the stress level was
made between the normokinetic borderline hypertensive group and the
normotensive group using the t-test. Both groups showed a nonsignificant trend
toward a decrease in Al from the resting baseline in response to the SCWT.
However, there was no significant difference in the change in Al from the
baseline to the stress level between the two groups (t = -0.98, p = .3326). The
change in Al from the stress to the recovery level was also compared between
the normokinetic borderline hypertensive group and the normotensive group.
Both groups showed a nonsignificant trend toward an increase in Al from the
stress to the recovery level. Again, no significant difference in the change in Al
from the stress to the recovery level between the two groups was found (t = -
0.52, p = .6064). Al across the three experimental periods in both groups is
displayed in Figure 4.10.
Arterial compliance. Correlational analyses were performed to identify
correlates of the SV/PP ratio. The SV/PP ratio was inversely related to age (r = -
.60, p < .0001), BMI (r = -.47, p < .002), HR (r = -.45, p = .003), and Al (-.65, p <
.0001). The SV/PP positively correlated with height (r = .34, p = .03). The
relationship between the SV/PP ratio and Al is shown in Figure 4.11. The SV/PP
ratio strongly correlated with SV (r = .83, p < .0001) and moderately inversely
correlated with aortic PP (r = -.56, p < .0001).
A comparison of the change in arterial compliance from the baseline to the
stress level was made between the normokinetic borderline hypertensive group
and the normotensive group using the t-test. There is no difference in the change
in Al between groups (t = -1.67, p = .103). Arterial compliance across the three
experimental periods in both groups is displayed in Figure 4.12.
Repeated measures analysis of variance was performed using a mixed
model to determine the difference in the pattern of change in arterial compliance
during the stress protocol. Age and BMI were adjusted in the model. The change
in Al was not significantly different within subjects (F = 2.07, p = .1311), though
there was a between-group difference (F = 26.66, p < .0001). The pattern of
change in arterial compliance was not statistically different between groups (F =
0.79, p < .4548). The effects of BMI and age on arterial compliance were
significant (F = 18.02, p < .0001 and F = 27.83, p < .0001, respectively).
Blood pressure reactivity and recovery. A comparison of the BP reactivity
and recovery scores between the borderline hypertensive group and the
normotensive group was performed using a t-test. Table 4.7 shows the BP
reactivity and recovery for the two study groups. No significant differences in the
reactivity and recovery BP scores between the two groups were found.
.4- r= .58, p < .01
.1 I O O
10 20 30 40 5"0 60 70
Figure 4.6. Relationship between augmentation index (Al) and age
(bivariate correlation). BH: borderline hypertensive subjects, NC:
normal control subjects.
.4 o r=-.33, p<.05
150 160 170 180 190 200
0 0 NC
150 160 170 180 190 200
Figure 4.7. Relationship between augmentation index (Al) and height
(bivariate correlation). BH: borderline hypertensive subjects, NC:
normal control subjects.
.4. 0 r= -.40, p< .01
.1 I OB O
o D BH
D D NC
20 40 6"0 80 1 20 140 160
Stroke Volume (ml)
Figure 4.8. Relationship between augmentation index (Al) and stroke
volume (bivariate correlation). BH: borderline hypertensive subjects, NC:
normal control subjects.
N= 17 25
Figure 4.9. Gender comparison of augmentation index (Al). Values
expressed as means + SD, p < .05 versus males, by f-test.