15315.».
4.}: .r,
.p 0!

1.“.
pr.

.2 3
c.

. y . .
. IA 0"» vnvmybw x
mum“ {an is». 2

“MARK:

t; " ‘

, at;

#9
“ .0 I
”:5

.‘x’ '
ﬁ

sh
‘z‘m

-‘ ; 5...
.21)....

1......u ‘ hm};
«(Kc-J" och-ft
anur... .! r
; .....3..-nn.»zl:1$....5.. .1... 1.1..
. but..to.1}htir .
lid” .1.

a n .

x) . Liv.

T yi‘nvtvtv 0.2,:
3. .t‘! .su: Hutu.

. .msmuuak.
s»... . >

1.9? .

V.’ n‘

 

 

 

 

i: ..l.. ..
.. . bun...“

..)7.n1.u.~é(nﬂu|.r}) ‘
.1; 3:: ...

Kstnb.2.,\..:.L
. W ‘

0.} u|.L
$31.. . Ll

- mfitgi-s; ,
4341;} 3' ‘ '
1.21,! :I’J"!
.

 

..I . . y n. u‘.. , A. . “ ‘ . . ‘ . (if...
. I. . J v. V K . _ IAN§W~$W€£TIMVY V. ‘ , . . 1..)1 1!.
ta .vrwnﬁm...d .;. ..re..41uu§:€ . . . . , . . . . . .. . . . . . .
. . .l ..... .

. v. .v..lr5;:.1 Wiltl‘v
.tfax.)): , ‘ . ‘ ‘ .
f r. L 3'. . . ‘

 

 

 

1111111171111W11"“VITISIH'imlﬁl'lﬁllsll i M 50 $1” __

:1 1
ll! 111 l

 

 

 

 

 

 

 

 

 

 

 

 

 

31293 006091171 "w "1
- " "DIARY
Michigan Stas a
University
Thisistocettifythatthe
(ﬁmmﬂnﬁonenﬁﬂui

THE EFFECTS OF EXPERIMENTAL HEART FAILURE ON
ARTERIAL BAROREFLEX CONTROL OF THE CIRCULATION
IN CONSCIOUS DOGS

[nawmuﬂtw

Nicholas Bari Olivier

luutxznaounmaiunuudsﬁdﬁﬂnmmt
ofﬂuuequhemwnudbr

 

 

Ph.D. degreein Physiology

ﬂak/5.3m:

Majorprofeuor ’

Date September 20, 1989

 

MSUBanAcﬁm/WQMIMM 042771

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or baton date duo.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11 1

MSU Is An Afﬁrmative Action/Equal Opportunity Institution

 

 

 

THE EFFECTS OF EXPERIMENTAL HEART FAILURE ON
ARTERIAL BAROREFLEX CONTROL OF THE CIRCULATION
IN CONSCIOUS DOGS

BY

Nicholas Bari Olivier

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Physiology

1 989

(505-3in

ABSTRACT

THE EFFECTS OF HEART FAILURE ON
ARTERIAL BAROREFLEX CONTROL OF THE CIRCULATION
IN CONSCIOUS DOGS

BY

Nicholas Bari Olivier

Aortic denervated dogs were prepared for reversible carotid sinus isolation and the
induction of heart failure by rapid ventricular pacing to determine if heart failure impairs
baroreflex control of blood pressure in conscious subjects. Complete open-loop
stimulus-response relationships were determined in quietly resting, conscious dogs
before (Phase I), during (Phase II) and after the resolution (Phase III) of heart failure.
Reﬂex range and gain were determined for baroreflex control of mean arterial pressure
(MAP), heart rate (HR), cardiac output (CO), and total peripheral resistance (T PR).
Ventricular pacing at rates of 255 bpm for 21 days resulted in bi-ventricular, low-output
congestive heart failure characterized by a 45% reduction in resting cardiac output
(3.51 z .6 to 1.91 z .6 L/min), an increase in right atrial pressure from 1:2 to 12:3
mmHg, and an increase in left atrial pressure from 3:2 to 21 :21 mml-lg. A dramatic
impairment in carotid sinus baroreﬂex control of both MAP and HR was associated
with the development of heart failure. The range of reﬂex control of MAP was reduced

by 52% while the average reﬂex gain decreased by 59 percent. These changes in

baroreﬂex control of MAP were due entirely to a similar impairment in the range (75%)
and average gain (66%) for reﬂex control of cardiac output since reﬂex control of TPR
was unaffected by heart failure. This selective impairment in CO responses while TPR
control remained intact is suggestive of an efferent neuro-eflector site for the overall
reﬂex impairment. Pacing was then discontinued and recovery from heart failure was
evident as early as day 7 of Phase III. During this recovery, reﬂex control of blood
pressure was largely restored. Near complete recovery in the range (76% recovery)
and average gain (75% recovery) were evident during this first week. These data
suggest a rapidly adapting mechanism for the impairments in the carotid baroreﬂex

caused by pacing-induced experimental heart failure.

The efforts and any beneﬁts associated with the information summarized within
are dedicated to my wife, and closest friend

Amanda Marie Jumonville

ACKNOWLEDGMENTS

I would like to acknowledge and thank the members of my guidance committee for
their contributions to my training. Each helped shape my graduate school experience
in a general sense, but also in individual and special ways.

My thanks to

Dr. George Eyster, for continued encouragement and enthusiasm and for his belief
in the value of thorough, fundamental research training.

Dr. Gregory Fink, for patience, and sparking my interest in statistical analysis. His
achievement of academic excellence with such apparent effortlessness sets a
standard I continue to strive for.

Dr. John Chimoskey, a true philosopher, for his continued support and wide
ranging perspectives. He taught me to avoid being satisfied with simply an acceptable
interpretation.

Dr. Harvey Sparks, for pure inspiration.

Of course, special thanks go to the chairman of my committee, Dr. Robert
Stephenson, a friend and trusted advisor. He was able to 'teach his graduate student
graduallness', as Larry Weed puts it. He somehow also managed to observe my
desire for independence while always being available for advice and gently redirecting
me when my efforts strayed. He taught me more than he will ever know.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES

INTRODUCTION

Homeostatic Cardiovascular Reﬂexes are impaired in Congestive
Heart Failure

Potential Mechanisms for impairment in the Arterial Baroreﬂex

Previous Studies Investigating Baroreﬂex Characteristics in Heart
Failure

Formulation of the Experimental Questions

Selection and Characterization of the Heart Failure Model

METHODS
Surgical Preparations
Confirmation of Aortic Denervation
Determination of Baroreﬂex Stimulus-Response Relationships
induction of Heart Failure
Experimental Protocols
Statistical Analysis

RESULTS

Effects of Pacing Alone on Resting Hemodynamics and Reﬂex
Function

Development of Congestive Heart Failure by Chronic Rapid
Ventricular Pacing

Heart Failure Persists During Brief Periods When Pacing is
Discontinued

Baroreﬂex Control of Blood Pressure and Heart Rate are Impaired in
Congestive Heart Failure

lmpainnent in Reﬂex Control of Blood Pressure is Due to a Selective
impairment in Reﬂex Control of Cardiac Output

Effects of Heart Failure on Reﬂex Function: Pacing Curves

viii

12
14

16
16
19
24

24
27

4O

40

44

46

Heart Failure and Reﬂex Impairment are Partially Reversible After

Pacing is Discontinued

DISCUSSION
Goals of the Study

Comparison With Results of Previous Studies

Signiﬁcance of the Results
Potential Sites of Reﬂex Impairment

Time Course and Potential Mechanisms for Reduced Cardiac

Responsiveness
SUMMARY
APPENDIX A

LIST OF REFERENCES

vii

88$8§8i$

72

74

LIST OF TABLES

TABLE

1. Control Experiment: Effects of brief periods of rapid ventricular
pacing and chronic low frequency pacing on resting hemodynamics
and baroreﬂex parameters. (* p <.05, NS = not
signiﬁcant for comparisons with control data).

2. Effect of chronic ventricular pacing on resting hemodynamics
illustrating the rapid development of congestive, low-output heart
failure. MAP = mean arterial pressure, HR = unpaced heart rate, CO
= cardiac output, TPR = total peripheral resistance, LAP = left atrial
pressure, RAP = right atrial pressure. (* p <.05 for comparisons with
control data)

3. Echocardiographic data illustrating cardiac enlargement and systolic
dysfunction resulting from chronic ventricular pacing. (n =5, * p < .05
compared to control values)

4. Effects of congestive heart failure on arterial baroreﬂex
characteristics. (* p <.05 compared to control, NS = not signiﬁcant)

5. Recovery Data: Effects of discontinuing rapid ventricular pacing on
resting hemodynamics and baroreﬂex characteristics for the 3 dogs
studied in Phase lll. Only two of these 3 dogs had a functional ﬂow
probe and left atrial catheter during Phase III. (* p < .05 for
comparisons of day III-7 with lI-21)

viii

31

37

37

42

LIST OF FIGURES

Eigure Base

1. Schematic diagram depicting the surgical preparation of the carotid 18
sinus regions for reversible vascular isolation. Heavy bars indicate
ligation of vessels. Stippled structures represent pneumatic
occluder cuffs.

2. Hypothetical baroreﬂex stimulus-response curve depicting the 23
physiologically relevant curve parameters.

3. Reﬂex stimulus response curves before and after short periods (30) 32-33
minutes of rapid ventricular pacing. The ranges and gains of
baroreﬂex control of mean arterial pressure, heart rate, cardiac
output, and total peripheral resistance were not signiﬁcantly
affected by this short duration of pacing.

4. Reﬂex stimulus response curves before and after 7 days of low 34-35
frequency ventricular pacing (LFP, 170 bpm) illustrating an absence
of effect of pacing alone on baroreﬂex control of blood pressure
and heart rate. The effects of LFP on reﬂex control of cardiac
output and total peripheral resistance are opposite in direction to
those seen with chronic RVP and heart failure.

5. Time course for the development of echocardiographic changes 39
associated with congestive heart failure. (n =5, * p < .05 compared
to control, day H)

6. Comparison of resting hemodynamics in the control (Phase i) 41
period, during pacing of Phase ii, and during a brief no-pace period
in Phase Ii. Congestive, low-output heart failure is maintained
during transient Phase iI periods when pacing is turned off. (* p
< .05 for comparisons with control)

ix

10.

11.

12.

13.

Heart failure reduces the range and average gain of baroreﬂex
control of both mean arterial pressure and heart rate.

Heart failure related impairments in the arterial baroreﬂex are due to
selective impairment in reﬂex control of cardiac output. Baroreﬂex
control of total peripheral resistance (T PR) is unaffected by
pacing-induced heart failure.

Pacing Curves: Heart failure reduces the range of reﬂex control of
blood pressure and cardiac output during pacing. Reﬂex control of
TPR was not affected by heart failure. Stimulus response curves
were generated during rapid ventricular pacing.

Phase III reversal of resting hemodynamic abnormalities.
Phase lll reversal of arterial baroreﬂex abnormalities.

Time course for the development of abnormalities in the arterial
baroreﬂex as a result of Phase II pacing (* p < .05 compared to
control values, day l-1)

Functional schematic of the arterial baroreﬂex control system.
Effects designated as 'External' represent inﬂuences that are
independent of the reﬂex loop. Their effects are simply added to
the reﬂex inﬂuences.

45

47-48

51

52

61

INTRODUCTION

WWW
iC I' II IE'I

Reﬂex adjustments in cardiac output and vascular resistance are necessary to
maintain a stable blood pressure during periods of hemodynamic stress. Postural
changes, such as assuming a standing position, elicit venous pooling of blood and a
reduction in venous return. The resulting drop in cardiac filling pressure, cardiac
output and arterial blood pressure are normally sensed by baroreceptors located in the
heart and lungs (cardiopulmonary receptors) and the aortic arch and carotid artery
bifurcation (arterial receptors). A change in the activity of these stretch receptors
during orthostasis initiates neuro-endocrine excitation including an increase in
circulating plasma concentrations of norepinephrine, renin, angiotensin, and
vasopressin (Campese, 1980). The result is a compensatory increase in heart rate and
vascular smooth muscle contraction which collectively prevent a postural drop in

systemic blood pressure.

Subjects with advanced heart failure, however, do not respond normally to
orthostatic stress. Several studies have shown that plasma levels of norepinephrine,
renin activity, and vasopressin do not rise during passive head-up tilt in symptomatic
heart failure patients. (Cody, 1921; Levine, 1983; Goldsmith, 1983) Wood and Zelis

and others have shown though, that venous compliance is reduced in heart failure

2

such that postural venous pooling may not be as profound. (Wood, 1956; Zelis, 1974)
The question then arises whether passive tilt represents an adequate baroreceptor
stimulus under these circumstances. More recent studies by Kassis (1987), and
Ferguson (1984), substantiated the adequacy of baroreceptor unloading by
documenting a signiﬁcant fall in both cardiac ﬁlling pressure and systemic arterial
pressure during either head-up tilt or lower body negative pressure. Mohanty et al.
(1989) measured ventricular dimensions during lower body negative pressure in
patients with heart failure and also conﬁrmed that such a stimulus invokes a signiﬁcant
reduction in end-diastolic ventricular diameter and left atrial diameter. Thus, even in
the face of decreased venous compliance, these provocative maneuvers are capable
of unloading baroreceptors. Nevertheless, the patients with heart failure in these
studies did not experience the normal increase in heart rate, forearm vascular
resistance or plasma catecholamines expected from the reduction in baroreceptor
load. Nonspeciﬁc impairments in vascular responsiveness to adrenergic stimuli were
ruled out by the demonstration of a normal pressor response to cold water hand

immersion.

These findings clearly demonstrate that heart failure is associated with signiﬁcant
alterations in cardiovascular reﬂex mechanisms. Despite general agreement for this
conclusion, uncertainty remained regarding which baroreceptor ﬁeld, cardiopulmonary
or arterial, is specifically impaired by heart failure. Since postural changes tend to
unload both baroreceptor ﬂelds simultaneously, it is difﬁcult to ascribe observed
impairments to one reﬂex or the other. Roweli (1986) pointed out that the fall in cardiac
ﬁlling pressure following passive tilt or lower body negative pressure usually occurs
before the drop in arterial blood pressure. Most of the reﬂex changes in forearm
vascular resistance also occur before the fall in arterial pressure while heart rate
changes tend to follow the hypotension. As a result, it has been suggested that the

forearm responses reﬂect the cardiopulmonary reﬂex while the heart rate response is

3

dependent on the arterial baroreceptors. This distinction however, is probably
incomplete, making it difﬁcult to differentiate the two reﬂex responses when both are
activated and potentially interacting.

 

Advanced congestive heart failure is typically accompanied by several
neuro-endocrine changes as well as varying morphologic changes in several
components of the baroreﬂex arc. For example, adrenergic tone is increased,
parasympathetic stimulation is reduced and resting plasma levels of renin,
angiotensin, atrial natriuretic peptides, and vasopressin are increased in naturally
occurring and experimental models of heart failure. (Francis, 1988) Many of these
changes have been studied in relative isolation, each being capable of modifying
overall baroreﬂex characteristics. As a result, each could be important as a possible

cause for alterations in baroreﬂex properties in heart failure.

For clarity, the effects of these neuro-endocrine and morphologic changes will be
considered separately and categorized according to the location of their effect as
afferent, central or efferent modiﬁers. it should be remembered however, that many of
these effectors can inﬂuence baroreﬂex properties at multiple sites and also that many

potentially interactive effectors can exist concurrently in patients with heart failure.

WM - The assertion that baroreﬂex properties are altered by the
existence of heart failure has led to an almost universal assumption that the major site
of impairment was the arterial baroreceptors themselves. Surprisingly, very little
experimental evidence exists to support this reasonable, but unproven assumption. In
addition, most of the available evidence is only indirect, supporting the possibility that
receptor function is altered secondary to the neural, endocrine, and morphologic

changes that accompany heart failure. For example, Zelis (1971) reported that arterial

4

wall sodium content is increased in dogs with pacing-induced heart failure. He
speculated that high sodium and water content reduces the compliance of arteries. if
so, then the pressure-volume relationship of the carotid sinuses might be altered in a
fashion compatible with a reduction in their sensitivity. Bagshaw (1972) reported
contrasting results when he perfused the isolated carotid sinuses of anesthetized dogs
with norepinephrine. He discovered that norepinephrine also reduces the compliance
of the vessel but can actually enhance the sensitivity of baroreceptors. A similar
ﬁnding was observed if postganglionic cervical sympathetic nerves were stimulated at
relatively high frequencies of 10-20 Hertz. (Peveler, 1980) This somewhat unexpected
ﬁnding was tentatively explained by a series relationship between the adventitial
receptors and the smooth muscle contractile elements. Despite the agreement of
these ﬁndings, the effects of adrenergic stimulation (endocrine or neural) on receptor
function, appears to be complex; with either enhancement (T omomatsu, 1981;
Sampson, 1970) or depression (Bolter, 1980; Keith, 1974)) of receptor activity and
sensitivity noted in different studies. Munch (1987) has attempted to explain these
differences by suggesting a bimodal, 'dose-dependent' effect. Mild adrenergic
stimulation appears to increase the resting activity of the receptors but reduces their
strain sensitivity. This 'low-dose' effect is thought to reﬂect signiﬁcant changes in the
vessel wall properties of the carotid sinuses since it is mimicked by other vasocon-
strictors and can be abolished by alpha blockade. Greater degrees of adrenergic
stimulation apparently enhance baroreceptor sensitivity through a direct action on the

receptor endings themselves.

Baroreceptor sensitization could also potentially occur as a result of increased
plasma levels of vasopressin. Schmid (1985) studied anesthetized, vagotomized and
aortic denervated rabbits in which the carotid sinus regions had been surgically
isolated. The isolated sinuses were then selectively pressurized with ﬂuids containing

either a vehicle or vasopressin. Baroreﬂex strength was determined by measuring the

5

degree to which baroreceptor loading reduced lumbar sympathetic nerve activity.
intrasinus vasopressin augmented the magnitude of baroreflex restraint on
sympathetic nerve activity. Since exposure to vasopressin was limited to the carotid
sinus region, he concluded that vasopressin has a sensitizing effect on the carotid
baroreceptors themselves. No such sensitization was observed when the carotid

sinuses were perfused with angiotensin ll.

Each of these studies presents evidence that baroreceptor function could be
altered in heart failure. Only one brief report by Zucker (1983) describes the changes
in arterial baroreceptor sensitivity actually associated with cardiac disease. He
reported that nigh-output heart failure caused by an aorto-caval ﬁstula reduced the
sensitivity of the pressoreceptors in the aortic arch by approximately 50 percent.
Fragmentation and degeneration of the receptor endings reported for other forms of

cardiovascular disease (Abraham, 1967) may be responsible for such ﬁndings.

Mechanisms - Alterations in baroreﬂex properties could also conceivably
arise secondary to changes in the central processing of baroreceptor afferent
information and/or the coordination of the afferent neural responses. Again, the
available evidence can only support a m39n_tiai role for neuro-endocrine factors in
mediating such a central effect during heart failure. In a companion experiment to the
one noted above, Schmid isolated the carotid sinuses in anesthetized rabbits but
administered vasopressin systemically. Enhanced baroreﬂex restraint of sympathetic
nerve activity was again observed and was abolished after complete sino-aortic
denervation. (Guo, 1982) Schmid implicated a central mechanism since the arterial
baroreceptors were isolated and therefore not exposed to the systemically
administered vasopressin. This conclusion is also supported by the work of
Undesser (1984), who demonstrated that destruction of the area postrema region of
the brainstem prevents the enhancement of baroreﬂex control of sympathetic nerve

activity by vaSOpressin.

6

Central enhancement of the baroreﬂex has also been reported following
angiotensin ll administration. Hull (1985) infused All in the lateral cerebral ventricles of
conscious dogs prepared for reversible carotid sinus isolation. The open-loop gain of
arterial baroreﬂex control of blood pressure was enhanced during these central
infusions. Whether CNS levels of angiotensin increase in parallel with plasma levels in
heart failure remains uncertain. When All was administered in the systemic veins, an
upward resetting of the reﬂex was noted without a change in gain. Characterization of
this response as either a peripheral action or a different central action due to different
sites of central exposure could not be made from his data. (Hull, 1985) it should be
noted, however, that the central effects of angiotensin II on baroreﬂex function may be
species dependent since Schmid has reported that peripheral infusions of All can
actually slightly depress reﬂex control of sympathetic nerve activity. (1985) He
attributed this to a central depressant effect-since All had no effect when conﬁned to
the carotid sinuses themselves as. no effect on the receptors) although he provided

no evidence to exclude some peripheral effect on efferent reﬂex components.

Cardiac ventricular receptor activity is likely to be increased, at least transiently,
during heart failure as a result of diastolic chamber enlargement and increased
sympathetic stimulation. (Thoren, 1977) The cardiac ventricular receptor reﬂex can
have a profound depressant effect on arterial baroreﬂex function (Denison, 1985).
Although the mechanism for this reﬂex interaction is not well understood, there is data
implicating a central site based on evidence that both reﬂex receptor ﬁelds send
afferent projections that converge in the nucleus tractus soiitarius of the
medulla. (Cottle, 1964; Lam, 1952) Sensitization of ventricular receptors and
enhancement of their inhibitory effect on the arterial baroreﬂex may also be the means
by which atrial natriuretic peptides depress the reﬂex tachycardic response to
baroreceptor unloading. (Ebert, 1988)

7

Although much less potent, atrial cardiac receptors can also influence arterial
baroreﬂex function. Experimental activation of atrial receptors augments carotid
baroreﬂex control of heart rate. (T so, 1985) Atrial receptor activity is naturally
increased by stretch of the atrial wall and is thought to be increased during the early
stages of heart failure. interestingly, Greenberg (1973) has shown that chronic heart
failure causes an adaptation or desensitization of these atrial receptors. This desensi-
tization involves a reduction in their tonic resting activity as well as a decrease in their
strain sensitivity. This chronic decrease in atrial receptor activity could potentially
inhibit baroreﬂex function.

EﬂerenLMechanisms - Abnormalities in efferent baroreﬂex mechanisms could
involve altered efferent autonomic nerve activity or transmission, or abnormal
end-organ responsiveness to this efferent activity. Most of the available studies
investigating efferent mechanisms in heart failure have focused on changes in
end-organ sensitivity. As might be expected, heart failure is associated with a
depression in cardiac responsiveness to autonomic stimuli. White created right heart
failure in dogs by inducing tricuspid insufﬁciency and pulmonic stenosis.
Approximately 4 weeks later, after clinical signs and hemodynamic evidence of heart
failure were evident, he cut the vagus nerve and stimulated the distal end at
frequencies ranging from .1 to 3 Hz. (White, 1981a) Dogs in heart failure had less of a
bradycardic response to vagal stimulation than control dogs. White then perfused the
sinoatrial node artery with different concentrations of acetylchoiine and recorded the
resulting change in heart rate. Again, the bradycardic response was depressed by
over 50% in affected dogs compared to normal controls. That this was a speciﬁc
insensitivity to choiinergic stimuli was conﬁrmed by demonstrating equivalent

bradycardic responses to hypertonic saline infusions in both groups of dogs.

8

Covell at al. (1966) reported a similar cardiac desensitization to adrenergic nerve
stimulation. Using the same model of heart failure in anesthetized dogs, they
stimulated the right and left post-ganglionic sympathetic cardio-accelerator nerves and
recorded the changes in heart rate and peak contractile force. Control animals
received right atrial stimulations only, in order to control for any positive inotropic
effects resulting from an increase in contraction frequency. Dogs with heart failure had
a signiﬁcantly blunted response to the sympathetic nerve stimulations (from 1 to 9 Hz)
in terms of both contractile force and heart rate. A two minute intravenous infusion of
norepinephrine elicited essentially equivalent increases in peak contractile force in the
two groups suggesting that the source of the desensitization may have resided in
nerve terminal release of norepinephrine. White’s study (see above) also reported a
normal tachycardic response to sinoatrial artery infusions of norepinephrine. This is a
somewhat surprising ﬁnding considering the well described reduction in
beta-adrenergic receptors in failing hearts. (Vatner, 1985; Bristow, 1982) Newman,
however, did document such a depression in inotropic responses to both
norepinephrine and isoproterenol infusions in dogs with experimental heart

failure. (Newman, 1977 & 1978)

The effects of cardiac end-organ insensitivity on reﬂex function are obvious.
Without a comparable increase in cardiac stimulation intensity, and /or compensation
by the vascular component of the baroreﬂex, blood pressure control by the arterial

baroreﬂex would be impaired.

The evidence regarding vascular responsiveness in heart failure is conﬂicting.
Forster (1989) presented a study in which a dorsal pedal artery and saphenous vein
were harvested before and after congestive heart failure was induced by rapid
ventricular pacing. Vascular rings were studied in-vitro in the presence of propranoioi,
indomethacin, and desipramine (to antagonize beta receptors, prostanoid production

and neuronal catecholamine re-uptake respectively). Dose response relationships for

9

force generation in response to norepinephrine and phenylephrlne were determined.
Vessels from heart failure dogs actually exhibited an increase in maximal tension and a
reduction in E050 suggesting an enhanced adrenergic sensitivity in the failure state.
Different results have been reported when vascular sensitivity is assessed in-vivo.
WIISOI‘I, for example, using the same model of heart failure, stimulated the lumbar
sympathetic chain and recorded the changes in regional vascular resistance. (Wilson,
1988) A high spinal block was performed to eliminate potentially varying background
sympathetic activity, and atropine was administered to abolish sympathetic choiinergic
vasodilation. The absolute increase in vascular resistance to nerve stimulation was not
different between heart failure and normal dogs. However, since resting resistance
was greater in heart failure, the fact that the change in resistance was not greater for
failure dogs is suggestive of a moderate insensitivity of the vasculature to nerve
stimulations. Since the responses to regional infra-arterial infusions of norepinephrine
were not different for the two groups, Vlﬁlson, like White, concluded that the apparent
deﬁcit was due to decreased neuronal release of norepinephrine. Admittedly, other
potential considerations would include changes in ganglionic transmission or the
concomitant neuronal release of a noncholinergic vasodilator substance. Kaiser
(1989) also used pacing-induced heart failure to examine the in-vivo vascular
responses to topically applied norepinephrine and nitroglycerin. No differences were
noted for these agonists between control and heart failure dogs although the
endothelial dependent vasodilator response to acetylchoiine was signiﬁcantly impaired

for the latter.

Although studies on individual neural or hormonal factors suggest that congestive
heart failure would be associated with altered arterial baroreﬂex characteristics, the net
effect on the baroreﬂex from these multiple neural and hormonal changes during heart

failure is difﬁcult to predict.

 

Only a few studies have attempted a speciﬁc examination of arterial baroreﬂex
characteristics in subjects with heart failure. Each of these studies however, was
potentially ﬂawed by the technical limitations described below.

Studiesﬁasenmhaarumemakﬂimdﬂessumﬂaiaﬁonships - Most of the studies

done in subjects with heart failure have employed modiﬁcations of a technique
originally described by Smyth (1969) where systolic arterial pressure (SAP) and the
R-R electrocardiographic interval (HI) are recorded during a transient change in blood
pressure induced by injections of vasoactive substances. A linear regression is
performed between the change in systolic pressure for each cardiac cycle and the
change in the next heart interval. The slope of this regression has been used as an
index of baroreﬂex sensitivity. However, interpretation of the resulting data is
potentially misleading for several reasons. For example, the relationship between
heart rate and heart interval is nonlinear. As a result, a decrease in heart interval slope
could occur with no change or even a slight increase in the slope for heart rate control.
This would be particularly evident when resting heart rates are different (higher); a
condition typical for heart failure. Moreover, the applied stimulus to the baroreceptors
is constantly changing with this technique so the analysis is sensitive mainly to fast
acting responses. While parasympathetic inputs to the heart can change quickly, the
sympathetic responses are considerably slower (Scher, 1970). Such a limitation might
create a bias toward detection of parasympathetic responses, which is suggested by
the consistent observation that slopes for hypertensive stimuli are greater than for
hypotensive stimuli (White, 1981). This could produce misleading results, particularly

since the relative importance of sympathetic and parasympathetic inﬂuences on heart

11

rate appear to change with heart failure. (Vatner, 1974) Despite these technical
limitations, this form of reﬂex analysis is widely used, and provides fairly consistent

results.

Higgins (1972) studied the Hl-SAP relationship in conscious dogs before and after
induction of low output heart failure using a tricuspid avulsion-pulmonic stenosis
model. Bolus phenylephrine injections were used to manipulate arterial pressure and
vary the stimulus to the arterial baroreceptors. The slopes for HI control were
depressed by 80% during heart failure. White reported similar ﬁndings for dogs, with
the additional ﬁnding that Hi slopes were also depressed when blood pressure was
deemesed by glyceryl trinitrate. Similar ﬁndings have been reported for humans with
naturally occurring heart failure. (Ellenbogen, 1989; Olivari, 1983)

Despite the consistency of these results, it remains difﬁcult to conclude how
baroreﬂex regulation of blmdmessum is affected by heart failure since several recent
studies have shown that reﬂex heart rate responses can be qualitatively different than
the responses for the other components of the reﬂex (total peripheral resistance for
example) and for the overall baroreﬂex regulation of blood pressure. (Hull, 1985;

Cowley, 1984; Angeli-James, 1980; and Guo, 1983)

Ware -- Since the regulated variable of the

arterial baroreﬂex is arterial pressure, an analysis of baroreﬂex characteristics should at
least include some measure of this capacity. Only two studies have attempted to
quantitate baroreﬂex control of blood pressure in experimental subjects with heart
failure. Higgins, in 1972, observed that bilateral carotid occlusion produced less of an
increase in blood pressure in dogs with right heart failure when compared to normal
dogs. He concluded that baroreﬂex control of blood pressure is impaired by heart
failure. However, data such as this does not provide enough information to evaluate
the 'strength' of the baroreﬂex, which is more rigorously quantiﬁed by the range and

gain of the complete reﬂex stimulus-response relationship. The response to bilateral

12

carotid occlusion alone cannot reliably distinguish between a reﬂex that is simply reset
and one with an altered range and/or gain. White, in 1981, provided additional
evidence to support Higgen’s conclusion in a study of anesthetized dogs with right
heart failure. He reported that the range of blood pressures resulting from changes in
isolated right carotid sinus pressure was reduced by heart failure. Unfortunately, this
potentially important ﬁnding was confounded by the use of general anesthesia.
Several studies have now shown that anesthesia can augment, impair, or reset
baroreﬂex function (Ebert, 1985; Cox, 1979; Stephenson, 1980) The dramatic effects
of anesthesia on reﬂex function are actually evident in White’s study as the effect of
vagotomy on baroreﬂex function. In White’s study, vagotomy in normal dogs caused a
negative ordinal shift and a deemse in the range of baroreﬂex control. This is
directionally opposite to the observations for the identical experiment in conscious
dogs. In the absence of anesthesia, vagotomy causes a shift up in the baroreﬂex
curve with an increase in range. (Stephenson, 1980) If anesthesia can depress the
response to vagotomy, might it similarly induce an artifactual depression in baroreﬂex

responses in heart failure?

WWW

Although abnormal baroreﬂex control of heart rate appears to be a consistent
ﬁnding in conscious subjects with heart failure, there is still no convincing evidence to
answer the question of whether reﬂex regulation of blood pressure is impaired. The
primary aim of this dissertation, therefore, is to answer this question. Does
experimentally induced congestive heart failure impair baroreﬂex regulation of blood
pressure and control of the circulation in conscious subjects? Baroreﬂex integrity will

be determined by analyzing the range and gain of the open-loop baroreﬂex stimulus

response

13

relations using the reversible carotid isolation model described by Stephenson. (1980)
The hypothesis to be tested is that W

 

As pointed out, many potential mechanisms could contribute to impairment of the
baroreﬂex in heart failure. Detecting a dominant mechanism would be difﬁcult in
conscious animals because of the invasive nerve stimulation and recording
experiments that would be required. An additional experimental aim is to provide a
preliminary search for the major site(s) of reﬂex impairment. This search is based on
the premise that abnormalities in the afferent and central components of the reﬂex
would be expected to produce a generalized effect on the efferent reﬂex components
(cardiac and vascular responses) whereas differential cardiac or vascular abnormalities
would suggest an efferent site of impairment, perhaps at the end-organ level. i
hypothesize that W
eﬂerentmechanism rather than the generalized pattern predicted from an afferent
defect. This hypothesis will be tested by quantifying and comparing the range and
gain of baroreﬂex control of both cardiac output and total peripheral resistance (T PR)

in conscious dogs, before and during experimentally induced congestive heart failure.

Finally, I propose a third hypothesis, that W
W (within 2 weeks) after resolution of the heart failure

state. in addition to the clinical importance of determining the reversibility of reﬂex
abnormalities, a study of the time course for the development and resolution of
observed changes may provide important clues about the mechanism of these
changes. Anatomic mechanisms, such as decreased vascular (baroreceptor)
compliance, would be expected to develop and resolve more slowly than changes
induced by neural or hormonal stimuli. Zelis, for example, has shown that the
increased vascular stiffness seen in humans with heart failure persists for at least 18

days after heart failure is resolved by cardiac transplantation. (Sinoway, 1988) The

14

evidence to date regarding the reversibility of CHF induced depressions in heart

interval control are conﬂicting. Dogs with high-output failure induced by an aorto-caval

ﬁstula continued to show a depression in the Hl-SAP relationship for as long as 8

months following surgical closure of the ﬁstula. (White, 1981) in contrast, the

depressed Hl-SAP slopes seen in humans with Iow.output heart failure are normalized

as early as two weeks after orthotopic cardiac transplantation and resolution of the

heart failure state. (Ellenbogen, 1989) Use of a reversible model of heart failure will

allow a test of this hypothesis through quantiﬁcation of baroreﬂex responses before,

during, and at selective time intervals after the reversal of failure.

SI |° ICI lil' III II lEil M II

The hypotheses to be tested in this dissertation impose certain demands on the

selection of an experimental model of heart failure.

1)

2)

3)

The model should produce low-output, congestive, biventricular failure to
accurately represent the conditions of most naturally occurring forms of heart
failure.

The severity of heart failure should be relatively stable throughout the three
week period of study. Preparations based on myocardial damage with
toxins or ischemia are therefore undesirable since they frequently produce
either too little damage with rapid recovery, or cause premature death from
fuiminating cardiac failure.

Previous experience has shown that open-loop baroreﬂex characteristics
remain relatively stable for an individual normal dog from clay to day. in
contrast, the coefficients of variation eetween dogs can be as high as
25-35%. A repeated measures design is therefore required to allow each
dog to provide their own normal control data. Since the functional lifespan of

15

the isolated carotid sinus preparation is typically 6 weeks or less, the
induction of heart failure cannot involve extensive interventions or surgical
manipulations requiring prolonged recovery times.

4) The model of heart failure must be reversible.

Only one established model of heart failure satisﬁes all these requirements, the rapid
cardiac pacing model originally described by Whipple (1962) and Coleman (1971).
Continuous ventricular pacing at rates between 240 and 280 bpm consistently
produces a signiﬁcant reduction in resting cardiac output and increases in atrial
pressures within 3-7 days of initiation. Plasma levels of norepinephrine, renin,
vasopressin, and atrial natriuretic peptide are all increased. (Riegger, 1981; 1984;
1988). The mechanism of cardiac impairment has not been elucidated although an
increase in afterload (wall stress) and a reduction in cardiac contractility have been
suggested from echocardiographic evaluations. (Wilson, 1987) Despite a 5-fold
increase in myocardial oxygen consumption, ischemia does not appear to be a major
pathophysiologic mechanism since neither the myocardial A-V oxygen difference or
coronary sinus lactate concentrations are signiﬁcantly increased. (Wilson, 1987)
Partial recovery of the reduced ejection fraction occurs within 48 hours of cessation of
pacing while diastolic cardiac dimensions and systolic wall stress remain elevated.
This improvement in systolic function in the absence of notable changes in cardiac
loads again implicates some fundamental deﬁcit in contractility. That structural cardiac
changes are also present is suggested by the persistence of cardiomegaly despite

restoration of cardiac output and atrial pressures during recovery.

METHODS

Thirty-ﬁve male and female adult mongrel dogs, weighing 20-30 kg, were used in
this study. Completion of the general experimental protocol was accomplished in 8 of
these dogs. Two additional dogs were used in a separate control experiment. The
results presented in this dissertation are from these dogs. Of the dogs not completing
the study, 1 was excluded because of refractory ventricular arrhythmias, 3 died
prematurely in congestive heart failure, 4 died as a result of thoracic surgical
complications, and 4 died from acute aortic rupture secondary to ﬂow-probe pressure
necrosis. Thirteen more dogs were not successfully studied because of premature
deterioration of the carotid sinus preparation and/or the implanted recording

instruments.

SurgicaLELenaLaiions

Two aseptic surgical procedures were required to prepare the dogs for study. The
ﬁrst surgery was a thoracotomy for implantation of recording devices. The second
surgery, performed approximately 10-14 days later, involved a cervical procedure to
prepare the carotid sinuses for reversible isolation. it was necessary to perform the
thoracotomy ﬁrst to maximize the use of the limited lifespan of the carotid sinus
preparation and to allow maturation time for the aorta-electromagnetic ﬂow probe

transducer system.

16

17

Iberacotemy - Dogs were premedicated with intramuscular acepromazine malate
(.1 mg/kg) and atropine (.05 mg/kg) and general anesthesia induced by a bolus
intravenous injection of thiamylal sodium (4-8 mg/kg). Anesthesia was maintained by
inhalation of a mixture of halothane (1 -2%) in oxygen. A left 4th intercostai
thoracotomy was then performed. A Tygon catheter (1.25 mm i.d.) was inserted into
the descending aorta through a stab wound for measurement of arterial pressure.
Similar catheters were also secured within the right and left atrial appendages. The
adipose and connective tisSues surrounding the ascending aorta were removed to
facilitate placement of an electromagnetic ﬂow probe around the vessel. Dacron
velour was placed between the aorta and ﬂow probe in 6 of the 31 dogs instrumented.
Two platinum-tipped pacing electrodes were embedded and sutured to the
myocardium of the left ventricular apex. Care was taken to avoid trauma to
surrounding coronary vessels. The pericardium was closed around the cardiac
implants. After the catheters and leads were exteriorized at the midscapular region,
the chest was closed, evacuated and the dogs were allowed to recover. Each dog
wore a protective canvas vest to protect the exteriorized ends of the implants.
Following surgery, the dogs were acciimatized to the experimental setting on a daily
basis while the catheters were irrigated with a heparinlzed penicillin-dextran (70,000

lU/ml) solution, and the surgical wounds were cleansed.

WWW - Each dog was allowed 10-14 days to

recover from the thoracotomy before the second surgical procedure. Anesthesia was
induced and maintained as described above. Reversible carotid sinus isolation was
accomplished using the method of Stephenson and Donald (1980). All vessels

surrounding the carotid sinuses were ligated except for the common and external

18

OCCipiiOI 0- \ Internal Carotid 0.

External

COTOIId 0- Carotid Sinus

Common Carotid a.

/ o‘e.e.e.o.e.e.e.e.e o
o e e o O
o o . e e n
e e o o e
e e e e
g ‘0’.
e
e

........
...................

 

 

     

 

Lingual a.

Anterior
Thyroid a.

indwelling /

Occlusion Cuffs

 

 

Figure 1. Schematic diagram depicting the surgical preparation of the carotid sinus regions
for reversible vascdar isolation. Heavy bars indicate ligation of vessels. Stippled
structures represent pneumatic occluder cuffs.

19

carotid arteries which remained as the only entrance to, and exit from, the sinus region
(Figure 1). Pneumaﬁc occluder cuffs were piamd around the common and external
carotid arteries. Inﬂation of the occluder cuffs completely isolated the carotid sinuses
from the rest of the systemic circulation. Tygon catheters (1.02 mm l.d.) were inserted
into the common carotid artery between the occluders through the stump of the
anterior thyroid artery. These catheters allowed independent control of carotid sinus
pressure when the occluders were inﬂated and the sinuses temporarily isolated. Aortic
denervation was accomplished by identifying (and transecting) the aortic depressor
nerves as a small nerve bundle that exited the cervical vagosympathetic trunk.
Veriﬁcation of their identity was accomplished by intraoperative whole nerve recordings
demonstrating the characteristic pulse-synchronous activity of aortic baroreceptors.
The dogs were allowed 5-7 days to recover before experiments were initiated.
Calibration of the aortic ﬂow signal was accomplished at the end of this recovery

period using a standard dye-dilution (indocyanin green) reference.

CI] I' [EI'D I'

Effective aortic denervation was veriﬁed in each dog by comparing the bradycardic
response to a bolus injection of phenylephrine (200-400 ug) or methoxamine (2-4 mg)
with the carotid sinus alternatively open and isolated. Pre-injection mean arterial blood
pressure was not signiﬁcantly different for the two respective states (129 vs. 124
mm Hg). Injection of the pressor agent in the closed-loop state caused an average
increase in mean blood pressure of 55 mm Hg with a corresponding drop in heart rate
of 45 bpm. in contrast, during carotid sinus isolation, pressor injections induced an 83

mm Hg rise in blood pressure yet only a 1 bpm average drop in heart rate. The slope

20

or 'gain' of this measure of baroreﬂex activity was signiﬁcantly reduced (-.855
bpm/mm Hg to .008 bpm/mm Hg) by sinus isolation for an average loss of 91%

(range 79-100%) of the aortic baroreceptor heart rate buffering capacity.

 

All experiments were conducted with conscious, quietly resting and unrestrained
dogs in lateral recumbency. The experiment environment was shielded from
extraneous visual or auditory stimuli. Heparin (5000 IU) was administered
intravenously and the following variables were continuously recorded on an 8 channel
Grass physiograph: phasic and mean aortic pressure, mean left and right atrial
pressure, phasic and mean (cardiac output) aortic ﬂow, carotid sinus pressure, and a
surface lead electrocardiogram. Heart rate was determined manually from the EKG
recording. Total peripheral resistance was calculated as the quotient of systemic
perfusion pressure and cardiac output. At the time of study the dogs were allowed
10-15 minutes to adjust to the environment at which time resting hemodynamic values
were recorded. The carotid sinuses were then temporarily isolated by inﬂation of the
occluders and subjected to a static pressure of 50 mm Hg through a connection
between the carotid catheters and an external reservoir of heparinized, lactated
Ringers equilibrated with a gas mixture of 95% oxygen and 5% carbon dioxide. The
steady state systemic reﬂex responses in heart rate, cardiac output, mean arterial
pressure, and total peripheral resistance were then derived. These steady state
responses were typically evident within 30-45 seconds of a change in static sinus
pressure. The sinus pressure was then increased in a step fashion to 70 mm Hg and
steady state responses again determined. This procedure was repeated throughout

the range of carotid sinus reﬂex activity, typically from 50-200 mm Hg, in 20 mm Hg

21

steps. Total sinus isolation time averaged 12 minutes for the dogs of this study.
Thereafter, the carotid occluders were deﬂated and communication between the

carotid sinuses and the systemic circulation was reestablished.

The resting closed-loop data, and the steady-state open-loop values for each
hemodynamic variable at each carotid sinus pressure (CSP) were entered into an
MS-DOS, 80286 based microcomputer. A best-ﬁt (least squares) asymmetrical
sigmoidal curve was approximated for the raw, steady-state data using a 5-parameter
logistic equation (lshikawa, 1984) and an iterative curve ﬁtting algorithm (Wilkinson,
1988). in each case, the hemodynamic variable of interest served as the dependent
variable with isolated carotid sinus pressure, the independent variable. The equation

for ﬁtting the reﬂex stimulus-response relationship for blood pressure, for example, is:

P,
MAP- +P5

1+ gunmen—Pan”4

 

Where: P1 = range of dependent response
P2 = slope coefﬁcient
P3 = domain centering coefﬁcient
P4 = asymmetry parameter

P5 =- minimum dependent response

Although most of the steady state reﬂex data assumed an obvious sigmoidal
relationship, some of the cardiac output data sets during heart failure (both with and
without pacing) were extremely ﬂat (negligible range and slope). Forcing a sigmoidal
ﬁt to such data occasionally resulted in an abrupt and artifactual inﬂection in the ﬁtted
curve. Reﬂex gains calculated under such circumstances become artiﬁcially high. To

eliminate this, linear ﬁts were made for these data sets. In every case, the alternate

22

use of the linear ﬁt was defended by comparing the goodness of ﬁt of the two curve
types. Since a higher number of parameters almost always leads to a lower sum of
squares (SS), the comparison was standardized for the degrees of freedom available

with each curve type according to the following F test. (Motulsky, 1987)

SSI-532
dil'dlz
532

d];

F-

Where: 881 = sum of squares, linear
$82 = sum of squares, sigmoidal
df1 = degrees of freedom, linear

dfz = degrees of freedom, sigmoidal

A linear ﬁt was not used if the probability of the obsewed F value was less than 0.1.

Several physiologically relevant parameters of the baroreﬂex were derived from the
ﬁtted curves (Figure 2). These parameters included the maximum value of the
dependent variable (MAX), the minimum dependent value (MIN), the range (RNG) of
the reﬂex (MAX - MIN), and two reﬂex gain measurements. The maximum slope of the
ﬁtted curve represented the maximum gain (MAX GN) while the average slope
between threshold and saturation values of CSP represented the average
gain (AVG GN). Threshold and saturation are deﬁned as the values of CSP bounding
the upper and lower 5% of the range of the dependent variable respectively (threshold

and saturation were undeﬁned for those data ﬁt with a linear model).

23

 

 

  
 
 
 

 

 

 

 

250 j r I I I
maximum
200 w. -
.9.
.0
1.1!
S 150 threshold
> _ _
4.3
1: range
3 1 -
sope = gain
‘3 100 r —
<1)
G:
0)
C: minimum
50 - 1
saturation
O 1 1 1 1 L

 

0 50 100 150 200 250 300

Carotid Sinus Pressure

Figure 2. Hypothetical baroreﬂex stimulus-response curve depicting the physiologically
relevant curve parameters.

24

Heart failure was induced by continuous left ventricular pacing at constant rates
between 240 and 280 bpm. The group average for the pacing rate employed in this
study was 255 bpm. Pacing stimuli were supplied by a portable pulse generator
(Medtronics, #5320) carried by each dog in their protective canvas vests. For each
dog, the pulse current provided was 25% more than the minimum needed for
complete electrical capture of the ventricles. Typically, this would range between 5
and 7 mA although there was a distinct tendency for this requirement to increase
during the 3 week period of pacing. Although rarely needed, the pacing rate could be
adjusted to modify the severity of heart failure. For the purposes of this study, heart
failure was arbitrarily deﬁned as a 25% or greater reduction in resting cardiac output
and a left atrial pressure greater than 12 mm Hg. Characterization of the failure state
was also accomplished using M-mode echocardiographic measurements of cardiac
dimensions (Electronics For Medicine, Mark V; 2.5 MHz non-focused transducer). Left
ventricular end-diastolic and end-systolic minor axis dimensions were determined as
well as left atrial and aortic root diameter. Left ventricular wall thickness and left

ventricular fractional shortening were also recorded.

Expeﬂmenialﬂeiacals

Baroreﬂex sﬁmulus-response relationships and the derived curve parameters were
determined during three phases of study. Phase I served as the control period to allow
characterization of the normal reﬂex for each dog before the induction of heart failure.
Phase lI comprised the heart failure portion of the study and began with the initiation of
chronic ventricular pacing and extended for 3 weeks. Phase III provided information

regarding recovery from heart failure. It began at the end of the 3 weeks of Phase II

25

when continuous pacing was discontinued, and extended as long as the experimental
models remained operative (approximately 2 weeks). Each dog was studied at
selected times during the ﬁrst two phases of study including: Phase l, day 1 (H), "-3,
"-7, "-14, and "-21. An attempt was made to study each dog through two additional
phase III days, Ill-7 and Ill-14 although only 3 dogs successfully completed this Phase.

CentmLExperjmems - It was important to be able to distinguish the effects of heart
failure per se on baroreﬂex function from any potential effects of chronic, repetitive,
current-pulse administration (pacing). Since heart failure could not be totally
dissociated from pacing, a truly rigorous test of this hypothesis could not be made.
However, two control experiments were designed to address this issue. The ﬁrst
experiment involved rapid pacing (240-260 bpm) for brief periods (30 minutes) during
Phase I. Preliminary experience suggested that such a short pacing duration was not
sufﬁcient to cause overt heart failure in dogs. No-pace curves were generated
immediately before and 10 minutes after this transient Phase I pacing period.
Baroreﬂex characteristics for these pre-pace and post-pace curves were compared to
identify any effects of pacing on reﬂex function that would be independent of heart
failure itself. For example, if pacing alone had some depressant effect on reﬂex
function, the gain and/or range of the baroreﬁex might be depressed in the post-pace
curve relative to the pre-pace curve. This experiment was performed on each of the 8

dogs in the principal study.

It could be argued though, that any effect of exogenous current administration
might be time dependent and therefore might not be evident within the short 30 minute
period of Phase I pacing described above. In two additional dogs, pacing was
continued for 7 days. Dissociation of pacing from heart failure in this longer control
experiment required that the pacing rate be reduced to 170 bpm which also avoided
overt evidence of circulatory failure. No-pace baroreﬂex characteristics were then

determined before initiation of low-frequency pacing (LFP) and then again after 7 days

26

of continuous LFP pacing. Using the same argument, if pacing had some direct effect

on reﬂex function, the 7 day LFP curves should be different than those determined
before LFP.

 

hypothesis was made using a comparison of open-loop baroreﬂex characteristics for
control of heart rate and mean arterial blood pressure between phase I, and the 4 days
of Phase ii. In order for natural reﬂex control of heart rate to be included in the
analysis, rapid ventricular pacing was temporarily discontinued for approximately 20
minutes in the Phase lI experiments to allow collection of the open-loop raw data
(no-pace curves). Although the primary focus was on these data, reﬂex characteristics
were also determined during a brief period of pacing in phase I and with the pacer on
in Phase ll (pace-curves). Since pacing could'be expected to attenuate the cardiac
output contributions to overall reﬂex control of blood pressure, this protocol might
enable an assessment of the dependency of the reﬂex on the cardiac component (and
conversely the vascular component, see below).

... :11": ... .. ; .. ... ; ...; ;, .. ... . .. . .. . ..; ;. ... , If
heart failure associated abnormalities in the baroreﬂex reside in the afferent arm or the
central processing sites, evidence of impairment in both efferent limbs (cardiac and
vascular) would be expected. As an initial search for the major site of impairment,
separate comparisons were made of baroreﬂex control of the two principle effectors,
cardiac output and TPR. Baroreﬂex characteristics for both variables were determined
on all Phase l and II experimental days, and for all available Phase III data. If, as
hypothesized, baroreﬂex control of cardiac output is selectively impaired while TPR
control remains intact, an efferent site of impairment would be suggested. in contrast,

if reﬂex control of both cardiac output and TPR are similarly affected, a generalized

27

depression of the reﬂex would exist. Assuming a single 'lesion', such a result would
implicate a deﬁcit In those parts of the reﬂex are common to both efferent limbs;
namely the afferent and/or central sites. Admittedly, this distinction is a simple one,
and more easily defended if selective Impairment ls identiﬁed. Nevertheless, this
analysis may provide important clues as to the mechanism of reﬂex impairment in

heart failure.

 

MW. In order to determine if observed
baroreﬂex impairments are quickly reversible, chronic rapid ventricular pacing was
discontinued at the end of the 3rd Phase Ii week, and the dogs allowed to recover from
heart failure. Baroreﬂex characteristics were again determined on day Ill-7 and Iii-14
for comparisons with Phase i and II results. Normalization of any changes in the
arterial baroreﬁex observed during Phase ll would indicate rapid reversal of baroreﬁex

impairment.

Sli'i' IE Ii

Statistical analysis involved comparisons between control and heart failure for
resting hemodynamic data, echocardiographic data, and the derived physiologic
baroreﬂex parameters. When 2 treatment conditions were compared, as with the
Phase l pre-pace and post-pace reﬂex characteristics, a paired ttest was used. When
comparisons involved several days, such as the repeated measures comparisons
across the 5 experimental days of Phase I and II, a randomized block, model Ill
analysis of variance was used. The exception to this involved the comparisons for the
pace-curves in Phase l and II. Missing data unavoidably occurred for one dog on two
days, and one dog on one day for a total of 3 out of the total 30 data cells of this

series. Rather than estimate these missing cells, a one way unbalanced analysis of

28

variance was performed. Post hoc tests for the blocked ANOVA were performed if the
overall F ratio reached the stated level of statistical signiﬁcance. These post hoc
comparisons were limited to contrasting each Phase ll day with the Phase I data. This
maintained the number of comparisons at or below the number of degrees of freedom
for the ANOVA, and thereby sustained the nominal alpha statistical value of the overall
comparison. (Bernhardson, 1975) Post hoc comparisons for the one-way ANOVA
were performed using the Tukey-Kramer method for unbalanced designs. Statistical
power (1 -beta) calculations performed on the comparison between Phase I-1 and
Phase "-21 no-pace curves were based on the non-central F distribution for paired t

tests deﬁned by Kendall and Stuart (1973) and implemented by Dallal (1988).

Analysis of the recovery data in Phase III presented a statistical challenge in the
sense that only 3 dogs successfully completed this phase. Moreover, only two of
these dogs had functional aortic ﬂow probes for Phase III experimental days. Clear
directional trends were evident in most parameters as the dogs were allowed to
recover from heart failure. Use of the common alpha level of .05, however, could
potentially result in an inordinate bias toward type I error protection with high
vulnerability to type ll errors. Since no physiologic basis could be advanced to justify
such a bias, a more balanced degree of protection was sought. To achieve this,
statistical power (1 -befa) values were calculated for each of the recovery phase
comparisons using the method described above. For interpretive reasons, an alpha
level of .30 was chosen for the level of statistical signiﬁcance instead of the traditional,
but arbitrary, value of .05. This assignment provided what was considered a minimum
degree of conﬁdence in protecting against both type I and type II errors, with a beta
level of approximately .60. Statistical signiﬁcance (alpha) for all other tests, including

the major comparisons between Phase I and Phase ll, was assigned at the .05 level.

RESULTS

Throughout this chapter, results will be presented in customary units unless
otherwise speciﬁed. Hence, units of pressure are mmHg, heart rates are beats per
minute (bpm), cardiac output is liters/min and total peripheral resistance (T PR) as
mmHg*min/liters. Units of gain are the quotient of the dependent variable unit and the
independent units (mmHg). Tabulated values are presented as group means and

standard deviations.

Stimulus-response ﬁgures provide a visual summary of the results and were
generated by averaging representative points from the individual dogs” ﬁtted curves. A
best-ﬁt sigmoidal curve was then applied to these averaged data points to provide a
composite group curve. Because of the two-stage averaging procedure used for these
composite curves, the apparent range and gains for these graphic ﬁgures may not
coincide precisely with the average of the individual curve parameters for each dog.

All comparisons and interpretations were derived from the latter, a group average of
the individual values for each dog. The group composite curves are provided only for

general visual comparisons.

EﬂectsoLEacingAlanean
Bl'll I . IBIIE I'

The effects of short periods of rapid ventricular pacing (RVP) are summarized in
Table 1. it was not possible to completely prevent hemodynamic changes
associated with RVP in this Phase I control experiment; even when limited to short 30
minute periods. Brief pacing at 255 bpm resulted in an increase in resting mean
arterial pressure from 95: 12 to 107: 12. Left atrial pressure (2:2 to 8:2) and TPR
(27:5 to 31 :5) were also increased. In contrast, right atrial pressure was not
signiﬁcantly affected by pacing (1 :2 vs 2:1). Despite these modest changes in
resting hemodynamics, pacing had very little effect on baroreﬂex function as illustrated
in Figure 3. The stimulus-response curves for MAP and HR before and after brief RVP
are nearly superimposable. Although not statistically signiﬁcant, a modest reduction in
both the range (p= .07, power= .54) and average gain (p= .37, power= .54) for cardiac
output is evident after brief RVP. The increase in left, but not right atrial pressure
suggests that even brief RVP can have deleterious effects on left ventricular function
which could have contributed to a reduction in the cardiac output reﬂex responses. No

signiﬁcant changes were evident in reﬂex control of TPR after brief RVP.

Since it was not possible to completely dissociate RVP from hemodynamic
changes that could potentially induce impairment in the arterial baroreﬂex, two dogs
were paced at lower rates (LFP) continuously for seven days. The duration of 7 days
was chosen since a reduction in the range and gain of baroreﬂex control of mean
arterial pressure was evident in all 8 dogs of the principal study by day 3 and 7 of
continuous RVP. The results of this control experiment are shown in Table 1 and
Figure 4. Since only two dogs were studied, and only one of these had a functional
aortic ﬂow probe, statistical analyses were not performed on these data. As evident in

Figure 4, though, LFP for 7 days had either no visible effect, or an effect opposite in

31

Table 1. Control Experiment: Effects of Brief Periods of Rapid Ventricular Pacing and
Chronic Low Frequency Pacing on Resting Hemodynamics and Baroreﬂex
Parameters. (* p<.05, NS = not significant for comparisons with control data).

Resting Values
HAP
HR
CO
TPR
LAP

Reflex Parameters

HAP
R09
HxGn
Aan

HR
Rho
MxGn
Aan

Rne

MxGn

Aan
TPR

Rns

chn

Aan

 

Control

96:12
99:13
3.5:.6
27:5
2:2

PrePace

151:23
-5.2:2.5
-3.3 1.5

122:33
-4.0:1.6
-2.5:.9

2.27:.9
-.13:.08
-.07:.05

26:7
-.86:.4
-.51:.2

Pacing

107:12
255
3.4:.8
32:5
8:2

PostPace

162:20
-4.6:3.0
-2.7:1.0

120:24
-3.4:1.1
-2.1:.6

1.59:.7
-.09:.05
-.05:.03

33:12
-.79:.7
-.44:.2

NS
NS
NS

NS
NS
NS

NS
NS
NS

7 days of

 

Control

110:4
110:16
5.9
17
2:0

PrePace

132:50
-1.70:.8
-1.0:.5

48:18
-1.12:.9
-.679:.5

1.3
-.012
-.008

13
-.130
-.086

Pacing

95:29
170
7.4
11
3:2

PostPace

129:7
-1.7:.8
-1.1:.4

61:25
-.732:.03
-.466:.05

3.3
-.041
-.028

11
-0091
-.073

leanArterialPreamllntlgl

Heart llate (bpm)

Figure 3.

32

 

250 I I V I U

zoo ‘

150

100

 

50

 

. 0 P081 PAC!
0 1 n a n l 7 . PREFACE

0 BO 1 OO 1 50 200 250 800

 

Carotid Sinus Preemie

250 ’

 

200

150

100

 

 

5° 1- .1

- O POSTPACE
O a n l 1 1 . PREFACE

0 BO 100 1 50 200 250 300

 

 

 

Carotid Sinus Pressure

Reﬂex stimulus response curves before and after short periods (30 minutes) of rapid
ventricular pacing. The ranges and gains of baroreﬂex control of mean arterial
pressure, heart rate, cardiac output and total peripheral resistance were not
signiﬁcantly affected by this short duration of pacing.

Cardiac Output (L/alai
o

80

CO

40

TPR (mile-111W

2O

 

 

 

 

 

 

 

 

 

 

 

i- «ll
. O POOTPACE
1 1 1 1 1 . PREFACE
O CO 100 150 200 250 BOO
Carotid Olnua Praaaura
ﬂ 1 I I I
. O POSTPACE
1 1 1 l l . PREFACE
O 50 100 1 50 200 250 300

Carotid Sinua Praaata'a

Figure 3. (cont’d.)

lean Arterial Pressure (math)

Heart Rate (boll)

Figure 4.

 

 

 

 

 

 

 

 

 

 

 

 

 

250 I U U T U
200 - q
1 BO .-
1 OO 1-
OO -
. O POS'I’PACB
O 1 1 1 1 I , O PREFACE
0 BO 1 OO 1 50 200 2 50 800
Carotid Sinua Praaaura
280 I I r ' I I
200 1- -1
1 50 :
1 OO -
BO -
. O POSTPACE
0 1 1 1 1 1 . PREFACE
O 50 1 OO 1 50 200 250 300

Carotid Sinua Praaataa

Reﬂex stimulus response curves before and after 7 days of low frequency
ventricular pacing (LFP, 170 bpm) illustrating an absence of effect of pacing alone
on baroreﬂex control of blood pressure, and heart rate. The effects of LFP on reﬂex
control of cardiac output and total peripheral resistance are opposite in direction to
those seen with chronic RVP and heart failure.

CordlecDutput (L/mln)

TPR (MW)

Figure 4.

1O

4O

30

20

1O

(cont’d.)

 

 

 

l l I l l

 

 

 

50 100 150 200 250

Carotid Sinua Praaauo

300

 

 

 

l l l l I

GO 1 OO 1 50 200 250

Carotid Sinua Praaatra

 

300

O POSTPACE
O PREFACE

O POSTPACE
O PREFACE

36

direction to that seen with congestive heart failure associated with chronic RVP. These
results suggest that pacing alone, independent of heart failure, does not signiﬁcantly

impair baroreﬂex function.

W
0i iBlllil'IE'

Phase ll of this study was initiated by programming a portable pulse generator to
pace the left ventricle at rates between 240 and 280 bpm (group mean, 255 bpm).
The effects of this pacing on resting hemodynamics are shown in Table 2. After three
days, resting cardiac output was diminished by 31% and left atrial pressure increased
from 3:2 to 12:4 mmHg. Cardiac output remained depressed by approximately
25-35% throughout phase II while left and right atrial pressure increased throughout
the pacing period. As previously reported for this model of heart failure, mean arterial
pressure decreases by approximately 15 mmHg, while unpaced heart rate and total
peripheral resistance generally increase. All of these changes were statistically
signiﬁcant by day 21 of Phase II. These ﬁndings conﬁrm the ability of this experimental

model to consistently produce low-output, bi-ventricular failure in dogs.

W1lson (1987) reported that as much as a 4-fold increase in coronary blood ﬂow
occurs after 3 weeks of rapid ventricular pacing. Since the aortic ﬂow probes were
placed distal to the coronary arteries in these dogs, total cardiac output was
underestimated by the magnitude of this coronary ﬂow. Because coronary ﬂow
increases substantially in this model of heart failure, the severity of the reduction in
cardiac output might be overestimated. For example, if coronary blood ﬂow normally

represents approximately 5% of total left ventricular ﬂow (total cardiac output), and

Table 2.

37

Effect of chronic rapid ventricular pacing on resting hernodynamlcs illustrating the
rapid development of congestive, low-output heart failure. MAP = mean arterial
pressure, HR = unpaced heart rate, CO = cardiac output, TPR = total peripheral

resistance. LAP = left atrial pressure, RAP = right atrial pressure. (* p <.05 for

Heart Failure

 

comparisons with control data)
Control
I-1 Ii-3
Resting Heaodynamics
HAP 95:12 77:8 *
HR (unpaced) 97:22 92:12 HS
CO 3.51:.6 2.39:.6 *
TPR 27:5 33:9 us
LAP 3:2 12:4 *
RAP 1:2 5:1 *
Table 3.

II-7

80:11

100:16 HS
2.39:.5 *

30:6
15:5
8:1

HS
a

*

II-14

81:16
104:14
2.51:.7
28:12
18:4
9:3

*

IS
a

HS

II-21

79:11
123:22
1.91:.6
36:9
21:4
12:3

Echocardiographic data illustrating cardiac enlargement and systolic dysfunction

resulting from chronic ventricular pacing. (n=5, * p< .05 compared to control

values)

Echocardiographic Data
LV End Dias Dian (mm)
LV End Sys Dian. (mm)
Frx Shortening (X)
Left Atrial Dian. (ma)
Aortic Root Dial. (ma)
Saptal Thickness (II)
LV Hall Thickness (an)

Control

41:3
31:2
25:2
21:3
24:2
7:1
8:1

Heart Failure

 

42:3
33:4
21:3
24:3
23:2
8:1
8:1

HS
HS

II-7

43:4
34:4
22:3
24:3
23:2
7:3
7:1

HS

II-14

46:5
38:5
16:3
27:3
22:3
8:3
7:1

IS
NS
NS

II-ZI

52:4
42:4
19:3
28:5
21:2
8:3
7:1

NS
NS
NS

38

then increases 4-fold by day 21, total cardiac output would have been estimated at
3.69 L/min in Phase I and approximately 2.64 L/min on day 21 of Phase II; still a
signiﬁcant (28%) reduction.

The echocardiographic data presented in Table 3 and Figure 5 also show typical
ﬁndings of congestive heart failure such as an increase in end-diastolic left ventricular
and left atrial diameters. Although a deﬁnitive characterization of the cause of heart
failure cannot be made from such data, a deﬁcit in contractility is suggested by the
decrease in fractional shortening (25:2 to 16:3 percent (p < .05) by day 21).
End-systolic diameter was also signiﬁcantly inereeseg, a change usually induced by
either a decrease in contractility and/or an increase in cardiac afterload. Crude

estimations of left ventricular circumferential wall stress by the equation:

(MAP) (Avg. LV Radius)

Stress (afterload) - LV wall Thick

 

suggest that afterload was not changed in this model (231 :23 vs 235:9 mmHg)

providing more support for a contractile deﬁcit interpretation.

Subjective signs of heart failure were also evident. Dogs experienced moderate
decreases in muscle mass and a mild decrease in body weight despite significant
ascitic ﬂuid accumulations. The dogs were generally not distressed at rest but
developed a mufﬂed cough with moderate activity. Appetites remained normal.
Systolic cardiac murmurs suggestive of mitral regurgitation developed during Phase il

in 3 of the 8 dogs.

 

LVESD

 

 

 

 

Percent of Control

 

 

 

 

 

 

Genital ( —l) "—3 “—7 II- M “-31

Perc ent of Control

Fx Shortening

 

0
Control (I—i) Il~3 11-7 “-14 “-21

Day of Study

Figure 5. Time course for the development of echocardiographic changes associated with
congestive heart failure. (n=5, * p<.05 compared to control, day l-1)

ll IE'I E ilDi B'IEIIIM
EacinngDiscantinued

To allow reﬂex control of heart rate to participate in baroreﬂex control of arterial
pressure, RVP was temporarily discontinued in Phase II. Figure 6 compares the
changes in resting hemodynamics induced by RVP and the stability of these changes
during the 20 minute period when pacing was discontinued for stimulus-response
determinations. Brief discontinuation of pacing resulted in mild recovery of the cardiac
output changes induced by RVP. On day 21 of pacing, cardiac output recovered
immediately from 1.91 : .6 to 2.40: .7 and ﬁnally to 2.62: .8 at the end of the 20 minute
no-pace period. in contrast, left atrial pressure did not change immediately after
turning the pacer off (21 :4) and actually increased to 24:6 during the ensuing
no-pace period. This increase might be related to the modest increase in venous

return during the no-pace period.

Statistical notations in Figure 6 are for comparisons between the no-pace values
and the Phase l, control data. All hemodynamic changes remained statistically

different from control and fulﬁlled the deﬁnition of heart failure for this study.

WW
I 'Iii: I' ll IE'I

The effects of pacing-induced congestive heart failure on baroreﬂex function are
summarized in Table 4 and Figure 7. The results of this study conﬁrm the ﬁndings of
others, that heart failure impairs reﬂex control of heart rate. This was evident at all days
of Phase II as a reduction in the range of reﬂex heart rate changes (decreased by
28%), the maximum heart rate attained during carotid sinus unloading (decreased by
20%), and the aueragegain for reﬂex control of heart rate (25% reduction). Maximum

Figure 8.

-Percent of Control
41 O3 03
O O O

N
O

41

 

MAP CO TP R

[J Control. Pacing Temp. No Pace

Comparison of resting hemodynamics in the control (Phase I) period, during pacing
of Phase II, and during a brief no-pace period In Phase lI. Congestive, low-output
heart failure Is maintained during transient Phase II periods when pacing is turned
off. (* p< .05 for comparisons with control)

42

Table 4. Effects of congestive heart failure on arterial baroreﬂex characteristics. (* p< .05

compared to control, NS = not signiﬁcant)

 

Control Heart Failure
1-1 "-3 "-7 ”-14 ”-21
Reflex Pararrl.
MAP
MAX 220:20 157:24 * 149:21 * 134:16 * 126:21 *
Mill 57:5 56:4 Its 51:6 * 48:9 * 48:3 *
ItllG 161:18 99:24 * 98:24 * 86:13 * 76:21 *
THRESII 95:15 75:13 * 78:20 * 83:11 * 81:15 *
SAT 153:16 135:17 Its 145:28 IIS 154:23 Its 157:39 Its
MAX GAIII -4.60:2.8 -2.42:.8 * -2.43:1.2 * -1.87:.8 * -1.73:.8 *
AVG GAIN -2.7:.9 -1.5:.4 * -1.6:.7 * -1.2:.4 * -1.1:.5 *
flit
MAX 198:21 156:18 * 156:19 * 165:16 * 159:20 *
Mill 73:21 78:21 Its 76:7 us 65:16 Its 74:19 its
ltIIG 125:28 78:32 * 79:19 * 99:20 * 89:23 *
THRESII 95:12 73:16 * 79:22 85:19 lls 91:17 Its
SAT 154:15 144:23 * 137:32 * 156:23 * 163:36 *
MAX GAIII -3.25:1.0 -1.68:.9 its -2.93:1.9 Its -2.26:1.1 lls -2.57:1.6 HS
AVG GAIII -2.05:.6 -1.16:.6 * -1.60:.7 Its -1.47:.7 * -1.53 :.9 *
CO
MAX 4.56:.9 3.30:.8 * 3.09:1.2 * 2.98:1.0 * 2.59:.7 *
Mill 2.98:.5 2.73:.9 Its 2.37:.8 * 1.79:.5 * 1.94:.6 *
ltlIG 1.74:.4 .46:.4 * .71:.4 * 1.18:.5 * .52:.5 *
MAX GAIII -.116:.05 -.040:.03 * -.041:.03 * -.076:.04 * -.046:.04 *
AVG GAIN -.081:.05 -.027:.02 * -.032:.03 * -.047:.03 * -.027:.02 *
TPR
MAX 53:12 57:23 lls 56:19 its 49:24 Is 55:21 lis
Mill 19:3 22:10 its 23:11 its 21:8 lls 21:9 lls
ltIIG 33:11 35:13 Iis 33:11 Its 28:17 us 35:14 its
THRESII 76:12 68:13 Its 71:17 us 81:15 Its 63:5 its
SAT 153:21 142:18 lls 153:24 Its 150:23 IIS 170:55 ils
MAX GAIII -.67:.4 -.67:.3 Its -.82:.7 Its -.TS:.7 Its -.53:.4 its
AVG GAIlI -.38:.1 -.44:.2 IIS -.45:.3 its -.46:.5 lls -.37:.3 its

IaanArterlalPreaaudW

Heart Rate (bpm)

Figure 7.

 

 

 

 

 

 

 

 

 

 

 

250 I I t I I
200 - a
1 50 - '-
1 OD '- -'
50
O FAILURE
o 1 1 1 1 l 7 C CONTROL
0 BO 1 00 1 60 200 2 BO 300
Carotid Clnua Praaata'a
250 I W— 1 I U
200 -
1 50 P .1
1 oo - -i
‘0 i- d
. O FAILURE
o . . n n . a couraor.
0 GO 1 OO 1 50 200 250 300

Carotid Sinua Praaaura

Heart faﬂure reduces the range and average gain of baroreﬂex control of both mean
arterial pressure and heart rate.

44

gain for heart rate control was decreased (3.25:1 to -2.57: 1.6) by heart failure
although this difference was not statisﬁcally signiﬁcant (control vs day 21, p= .277;
power= .529). The minimum heart rate that could be achieved during isolated carotid
sinus hypertension was not changed by heart failure. interestingly, the bradycardic

response to baroreceptor loading appeared to remain largely intact.

Reﬂex control of mean arterial pressure was also signiﬁcantly impaired as early as 3
days after the initiation of RVP and low-output failure, and continued to be depressed
throughout the 21 days of Phase II. MAP range decreased by 52%, primarily the result
of a decrease in the maximum blood pressure attained during carotid sinus
hypotension (220:20 vs 126: 16). Auerage and maximaLgain of reﬂex control of
MAP were similarly decreased by 59% (-2.7: .9 to -1.1 : .5) and 62% (46:28 to
-1 .7: .8) respectfully. All of these changes were statistically signiﬁcant (p < .05).

These results conﬁrm the ﬁrst hypothesis of this study, that congestive heart failure

impairs baroreﬂex control of blood pressure in conscious subjects.

WWW
lSll'l'liBﬂDlltCl'Qil

The mechanism for impaired reﬂex control of blood pressure in this model of heart
failure was due primarily, if not entirely, to impairment in baroreﬂex control of cardiac
output (T able 4 and Figure 8) since TPR control was not affected by heart failure (range
p=.88, power= .88; max gain p= .59, power=.64; avg gain p= .91, power= .90).

As for MAP and HR, reﬂex control of cardiac output was depressed as early as 3
days after induction of heart failure and persisted throughout Phase II. A negative
ordinal shift in cardiac output (approximately 1 L/min) was evident as a signiﬁcant
reduction in both the maximum (4.56: .9 to 2.5: .7) and minimum (2.98: .5 to

1.94: .6) reﬂex changes in cardiac output. Heart failure was also associated with a

45

 

 

 

 

 

 

 

 

 

 

 

7 I I I I I
O - ..
‘E
E 5 - ..
\
:5
E 4i
8 s
o
8
E 2
0
1 P- -
. O iFAlilLiiJilRlE
O 1 1 1 1 l C CONTINGE-
O 50 100 150 200 250 300
Carotid Sinus Praeawa
80 U I 1 I 1
CO
3
g 40
20
. O FAILURE
O n n L 1 n O CONTROL
O BO 100 150 200 260 300

Carotid Sinua Praaawa

Figure 8. Heart failure related lmpairrnents In the arterial baroreﬂex are due to selective
impairment in reﬂex control of cardiac output. Baroreﬂex control of total peripheral
resistance (T PR) is unaffected by pacing-Induced heart failure.

46

reduction in the range (70%, 1.74: .4 to 52:5), ayeragegaln (66%, -.081:.05 to
-.027: .02), and maximumgain (60%, -.116:.05 to -.046: .04) of reﬂex cardiac output
control. For every dog, the degree (expressed as percent of control) of reﬂex
impairment in cardiac output was greater than for heart rate. This difference, however,

was not statistically signiﬁcant by a Wilcoxon non-parametric comparison (p = .06).

These data conﬁrm the second hypothesis which predicted that impairment in
reﬂex control of blood pressure would occur primarily because of a selective
impairment in cardiac responses. Although not conclusive, this ﬁnding is supportive of
a peripheral or efferent site of reﬂex dysfunction since an afferent or central site would

be expected to affect both efferent limbs of the reﬂex.

Eiilill iE'l BIIE i"E' D

As originally designed, stimulus-response curves generated during RVP in Phase I
and if were to be used as a means to eliminate the cardiac reﬂex responses. if pacing
was effective in eliminating reﬂex changes in cardiac output, reﬂex control of TPR
could be examined in relative isolation. In addition, this might have provided an
additional means to address the primacy of cardiac impairment as the cause for
overall reﬂex depression. The changes in reﬂex parameters during pacing associated
with heart failure are shown in Figure 9. Despite the imposition of RVP, a considerable
degree of reﬂex cardiac control was sustained in the normal dogs (Phase I) due to
changes in stroke volume. As in the no-pace curves, heart failure partially abolishes
this residual cardiac reﬂex response. The results of this experiment, therefore, were
similar to those of the no-pace data, with a signiﬁcant depression in the range of blood
pressure responses, entirely attributable to a reduction in cardiac output range. Both

of these changes in range were due to a signiﬁcant decrease in the maximum value of

47

 

250 l I I I I

200 - -

150

100

 

50

Mean Arterial Pressure (mmHg)

. O FAILURE
O 1 1 1 1 1 O CONTROL

0 50 100 150 200 250 300

 

 

 

Carotid Sinus Pressure

Figure 9. Pacing Curves: Heart failure reduces the range of reﬂex control of blood pressure
and cardiac output during pacing. Reﬂex control of TPR was not affected by heart
failure. Stimulus response curves were generated during rapid ventricular pacing.

Cardiac Output IL/mlrii
a

CO

40

TPR (mmHg'mIn/LI

Figure 9.

 

 

O 50

I I I I I

1100 150 200 250

Carotid 81mins Pressure

 

300

 

 

I I I I I

 

 

O 80

(cont’d.)

100 1 50 200 280

Carotid Sinus Pressure

300

O IFAIIILILIIIRIE
C CONTROL

0 FAILURE
C CONTROL

49

MAP (39%, 168:32 to 102:22) and CO (51%, 3.78: 1.1 to 1.84:.6) attained during
carotid sinus unloading. Reﬂex control of TPR during RVP was not affected by heart

failure.

HeartﬁailuraandﬂeneximnalrmemmEartially.
B 'IIEIIE' '0' I' I

Figures 10 and 11 and Table 5 show the changes in resting hemodynamics and
reﬂex parameters that occurred when rapid ventricular pacing was discontinued and
the dogs were allowed to recover from heart failure. As previously stated, only three
dogs successfully completed phase III and only two of these had functional ﬂow
probes up to day Ill-7. As a result, statistical comparisons were not made for cardiac
output and TPR recovery data. Comparisons for mean arterial pressure and heart rate
were performed with adjustments for the small sample size. Justiﬁcations for these
comparisons was based on the appearance of clear trends in resting and reﬂex
parameters during the early phase of recovery. The adjustment was a relaxation of the
alpha level for statistical signiﬁcance to .30 to maintain a beta (power) level of at least
.60. To prevent over-interpretations that might result from this relatively liberal
statistical approach, comparisons were limited to data at the peak of heart failure (day

"21) with data from day Ill-7.

Resting mean arterial pressure and heart rats exhibited signiﬁcant recovery
following cessation of RVP. Both variables actually 'overshot' the control data in the
early stages of recovery. Similar recoveries were also strongly suggested at day Ill-7
for cardiac output and TPR. This is consistent with previous reports on the magnitude
and time course of recovery for this model of heart failure. The reﬂex impairments in
heart rate range and ayeragegain show a similar recovery pattern, with values that

actually slightly exceed the control values.

Table 5.

Resting Haaodynaaics

Recovery Data: Effects of discontinuing rapid ventricular pacing on resting

hemodynamics and baroreﬂex characteristics for the 3 dogs studied in Phase III.
Only two of these 3 dogs had a functional ﬂow probe and left atrial catheter during
Phase III. (* p<.3 for comparisons of day Ill-7 with lI-21)

HAP
HR

CO (n22)
TPR (nsZ)
LAP (m2)

Reflex Parameters

Control

I-1

93:10
108:19
3.39:.2
26:5
3:2

CHF

II-ZI

76:17
was
1.72:.2
33:12
22:3

 

Recovery
Iii-7 iii-14
101:18 * 91 14
94:37 * 92 16
4.02:.2
20:2
7:.7 3 .7

MAP

Range
Max Gain
Avg Gain

Range
Max Gain
Avg Gain

155:33
-4.00:1.0
-2.49:.5

105:24
-2.56:1.0
-1.58:.5

60:21
-1.39:1.0
-.89:.6

67:22
-1.53:1.1
-1.07:.7

125:28
-3.39:3.3
-2.09:1.9

115:27
-3.18:2.9
-1.92:1.6

*

119 27
-2.19 1.9
-1.35 1.1

106 32
-4.89 4
-1.57 1.3

51

Percent of Control

 

140

 

 

120

 

100

 

 

 

8° ------ -—

 

5:3:
E' '-‘-'2- if!)
if: “25 $17;
60 ; , ............ 1.. .... 33‘:
' ' ' . :y‘.
;. 1 - - “-111
'1, 52;“, $163
- , »2_ .1}:

 

40 ... °*

 

 

 

{If
22:
iii

MAP HR CO
n=3 n=3 n=2

 

 

 

 

 

 

TPR

n=2

Control. n-mﬂ ni—rﬁ 111-14

 

Figure 10. Phase III reversal of resting hemodynamic abnormalities.

1.400
1.200
1 ,000
800

(LAP)
500

400

200

 

 

 

 

 

 

120
100
80
60
40
20

Percent of Control

140 _

 

 

 

Heart Rate

(‘ p<-3)

 

 

 

 

 

   

 

 

 

 

 

Range Max am Avg Gain

Mean Arterial Pressure

E] Control- “—21% iii—7E 111-14

Figure 11. Phase iii reversal of arterial baroreﬂex abnormalities.

53

Reﬂex control of blood pressure also shows a rapid recovery pattern, but without
evidence of an 'overshoot'. These ﬁndings suggest that at least part of the baroreﬂex
dysfunction associated with heart failure is mediated by some rapidly adapting
mechanism; perhaps a neuro-endocrine mechanism. This is also suggested by the
rapidity with which the impairments develop in Phase II (Figure 12). Most of the

changes were evident within 3 days of failure.

140

 

120

 

100

 

 

 

 

 

 

 

 

 

 

Percent of Control

 

 

 

 

 

 

 

 

 

 

Range Max Gain Avg Gain
Mean Arterial Pressure

Control- n-sl 11-7- 11-14. 11-21

Figure 12. Time course for the development of abnormalities in the arterial baroreﬂex as a
result of Phase ll pacing. (* p<.05 compared to control values, day l-1)

DISCUSSION

60.813121101311111!

The primary aim of this study was to determine if baroreﬁex control of [2|ng
pressure in senseieuesubjems was impaired by the presence of congestive heart
failure. Two subordinate aims were 1) the characterization of reﬂex abnormalities as
either generalized or selective for cardiac or vascular responses, and 2) to determine if

reﬂex impairments were reversible.

With respect to the ﬁrst aim, the results conﬁrmed the ﬁndings of others, that
baroreﬂex control of heart rate is depressed, but provided the important additional
information that the regulated output of the reﬂex is also poorly controlled. In addition,
the cause for the impairment was localized entirely to abnormalities in baroreﬂex
control of cardiac output. Although cardiac unresponsiveness might have been
expected, the distinction lies in the absence of any effect of heart failure on reﬂex
control of total peripheral resistance. Although not conclusive, this provides strong
support for an efferent site of reﬂex dysfunction, since an afferent or central site would
have been expected to cause a generalized depression in both effector arms of the
reﬂex. In addition, the results of this study provide additional evidence supporting the
recent ﬁndings that the impairments in baroreﬂex function can be quickly reversed if

the heart failure state is resolved.

56

Each of the goals set forth for this study was therefore met, although the level of
conﬁdence for the reversibility data was diminished by the limited number of dogs who

were available for Phase III evaluations.

C . iiiJ'IlB II [E 1 Si l'

The majority of pre-existing evidence for arterial baroreﬁex dysfunction in heart
failure has been based on demonstrations that the slope of the heart interval-systolic
pressure relationship (HI-SAP) is markedly reduced. The results of this study conﬁrm
the general hypothesis that baroreﬂex chronotropic control is abnormal in heart failure,
yet with noticeable differences from the previous reports. The magnitude of the
depression in heart rate control observed in this study was moderate with a 28%
reduction in range and a 25% reduction in average gain. In contrast, the slopes of the
Hl-SAP relationship were reduced by as much as 95% in previous studies (avg.
reduction, 82%). (White, 1981a; White, 1981b; Higgins, 1972; Ellenbogen, 1989) in
addition, the major effect of heart failure in this study was to blunt the tachycardic
response to baroreceptor unloading while the bradycardic response to baroreceptor
hypertension was not signiﬁcantly affected. Many of the previous studies were limited
to Hl-SAP analyses of hypertensive stimuli but those that also included hypotensive
stimuli found the more profound effect to be an attenuation of the bradycardic reﬂex
response. (White, 1981a; White 1981 b) The results of this study would therefore
appear to be at odds with those of previous reports. At least three possibilities exist
that could potentially explain these discrepancies. First, different forms and durations
of heart failure were present in these studies including a pulmonic stenosis-tricuspid
avulsion model (White, 1981b; Higgins, 1972), high-output failure from an aorta-caval
ﬁstula (White, 1981a), and unspeciﬁed causes of naturally occurring heart failure in
human patients (Ellenbogen, 1989). Secondly, the subjects with heart failure in these

57

studies tended to have higher heart rates (shorter heart intervals) than the normal
controls. Since the relationship between heart rate and heart interval is nonlinear,
higher resﬁng heart rates can cause a small reduction in the sensitivity of heart rate
control to appear as a large decrease in the sensitivity of heart inienrai control. Finally,
the transient beat-to-beat analysis used in generating the Hl-SAP slopes is inherently
biased to fast acting parasympathetic inﬂuences with a relative insensitivity to the
slower sympathetic responses. While the reﬂex bradycardic response in normal dogs
is almost entirely parasympathetically mediated, Vatner (1974) has shown that
sympathetic inﬂuences assume a greater role in dogs with heart failure. This shift in
the relative roles of the two autonomic inﬂuences during heart failure could artifactually
reduce the slope of the Hl-SAP analysis and accentuate the apparent severity of reﬂex

impairment.

Only White’s study (1981b) also included an assessment of baroreﬂex control of
blood pressure. In that report, he examined the range of the reﬂex change in blood
pressure resulting from changes in isolated right carotid sinus pressures in
anesthetized dogs with low-output heart failure. Although the normal ranges he
observed were less than those of this study (probably due to the effects of anesthesia
and the opposing inﬂuences of intact aortic baroreceptors), the magnitude of the
change in reﬂex range for blood pressure associated with heart failure was comparable

to that observed in this study (48% reduction vs 52% reduction respectively).

Signiﬁcanceimteﬁesults

The large magnitude of the changes in blood pressure control suggest that the
effects of heart failure on reﬂex function are likely to have physiologic consequences.
This interpretation is also consistent with the clinical observations that patients with
heart failure are unable to effectively attenuate changes in blood pressure resulting

58

from physical stresses such as postural changes. Using a simple control systems
analysis (Appendix A), the effect of a change in open-loop reﬂex gain on the blood
pressure buffering capacity can be estimated by the equation:

AMAth (1+Go)norm
AMAPmm (1+Go),,,

 

 

where delta-MAP is the change in blood pressure caused by a standardized
cardiovascular disturbance and Go is the open-loop gain of the reﬂex. Accordingly,
the effect of the 59% reduction in average gain for blood pressure would cause a given
disturbance in blood pressure to be magniﬁed by a factor of 1.8. A physical stress that
would normally cause only a 10 mmHg fall or rise in blood pressure would be

magniﬁed to an 18 mmHg disturbance in the presence of heart failure.

Another predicted consequence of abnorrhal reﬂex control of blood pressure is a
greater lability of blood pressure. This of course, assumes that the frequency of
occurrence and severity of spontaneous disturbances in blood pressure are similar in
patients with heart failure, and that alternative stabilizing mechanisms do not
compensate; assumptions yet to be refuted or conﬁrmed. Blood pressure instability, if
it exists in heart failure, could potentially aggravate the primary disturbances in blood
ﬂow to peripheral tissues and possibly accentuate the morbidity associated with CH F.
Olivari (1983) suggested this after observing that those heart failure patients with
blunted reﬂex responses to orthostatic stress experienced a higher mortality than
others with intact reﬂex responses. Parati et al (1986) proposed a similar hypothesis
for patients with hypertension. He separated them according to the magnitude of their
average 24 hour mean blood pressure. He then sub-stratiﬁed these groups into high
and low blood pressure variability groups. For almost all degrees of hypertension, the
high variability patients experienced a greater incidence of target organ damage. it

remains unclear whether this association involves a causal relationship. Nevertheless,

59

results such as these continue to raise questions regarding the importance of moment-
to-moment blood pressure regulation in maintaining health, or at least limiting the

progression of disease.

Another interesting question is whether the changes in baroreﬂex control of blood
pressure are truly deleterious, or might they be some means of compensation? For
example, clinical experience has clearly shown the beneﬁts of reducing afterload in
restoring the circulatory function of patients with heart failure. If so, why doesn’t
sempensaiem afterload reduction, through a mild reduction in blood pressure, occur
naturally in heart failure? If a mechanism did exist for such compensation, it might also
be important to prevent its attenuation by reﬂexes that regulate blood pressure, hence
the ‘value' of a blunted baroreﬂex. The asymmetrical effects of heart failure on
baroreﬂex function are consistent with such a theory since the reduction in reﬂex range
and gain were evident primarily in response to baroreceptor unloading. As a result,
the ability of the baroreﬂex to oppose a fall in blood pressure was decreased while the

ability to attenuate an increase in pressure was largely unaffected by heart failure.

While perhaps a plausible strategy, the data in this study would argue against reﬂex
changes being adaptive. This is because the impairments in reﬂex function were
entirely manifested in cardiac responses. It would be counter-productive to limit
cardiac responsiveness as a means to preserve adaptations designed to support
cardiac performance. For similar reasons it seems unlikely that arterial baroreﬂex
impairment represents an adaptive mechanism to increase sympathetic drive to a

failing heart and cardiovascular system. (Hirsch, 1987)

The ﬁnding that the impairments in reﬂex function after 3 weeks of heart failure are
partially reversible is potentially important from a clinical perspective. The reduced
tolerance to physical stress and increased blood pressure lability predicted from a
depression in arterial baroreﬂex function could contribute to the morbidity of
congestive heart failure. Though true reuersai of heart failure is usually limited to heart

60

transplant recipients or those recovering from acute, non-progressive forms of heart
failure (post myocarditis or myocardial ischemia for example), the hemodynamic
abnormalities associated with heart failure can often be improved substantially by
medical therapy. It so, the results of this study would suggest that substantial
improvement in baroreﬂex function would be expected to accompany the
hemodynamic improvements. Whether restoration of reﬂex function under these
circumstances would be entirely related to improved hemodynamics or possibly some
direct effect of medications is a topic of current debate. For example, acute adminis-
traﬁons of digitalis glycosides in normal dogs increases the afferent activity of arterial
baroreceptors. (Quest 1971, 1974) A similar enhancement of receptor activity is
assumed for chronic therapy as well since the bradycardic and hypotensive response
to baroreceptor loading remains enhanced. (Reing, 1973) These direct sensitizing
effects of digitalis are thought to play an important role in mediating its therapeutic
effect in the treatment of patients with heart failure. (Thames, 1982)

Ell'lS'i [Bill . |

An analysis of the type of baroreﬁex changes induced by heart failure provides
some insight regarding likely sites of impairment. The major observations included a
diminished reﬂex gain and a reduction in the MAX values for HR, CO, and MAP during
carotid sinus hypotension. Consideration of a simple control system model of the

arterial baroreﬂex facilitates the analysis (Figure 13).

Barereeenters - The 'feedback' function of the reﬂex is provided by the carotid
baroreceptors, characterized by their own sensitivity or gain and their own range of
responsiveness. Could, as commonly proposed, an isolated defect in receptor
function explain the observed impairment in reﬂex control of blood pressure? The
reduction in HRrm COM and MAPM during heart failure would not support such

Command Signal

A

 

61

parasympathetic

 

 

p

 

 

 

 

 

sympathetic
Venous Return

Contractility

 

 

 

i

 

 

 

 

 

 

 

 

Heart Rate

 

J

 

 

Stroke Volume

 

 

i

 

 

Cardiac Output

 

 

 

 

Arterioles (TPR)

 

 

l

‘—— External

‘— External

‘— External

—— baroreceptors ‘——// //—— Blood Pressure

Figure 13. Functional schematic of the arterial baroreﬂex control system. Effects designated
as 'External' represent inﬂuences that are independent of the reﬂex loop. Their
effects are simply added to the reﬂex inﬂuences.

62

an explanation. The maximal increase in HR, CO, and BP should be insensitive to a
change in baroreceptor function since they are inactive at subthreshold pressures and
therefore exert no negative feedback restraint on the reference or command signal.
The MAX values for the cardiovascular variables is determined by non-receptor
inﬂuences which includes the sum of the command signal effects and any external
(reﬂex-independent) modiﬁers. Using this model, the reduction in HR“, COM and

BP.m must be due to a mechanism other than baroreceptor dysfunction.

Could the baroreceptors also be adversely affected by heart failure and contribute
to the decrease in reﬂex gain? Although reduced aortic baroreceptor sensitivity has
actually been reported for this model of heart failure (Dibner—Dunlap, 1988), the results
of this study do not support this as a signiﬁcant contributing factor. Since the
baroreceptors are common to both cardiac and vascular efferent limbs, any change in
their gain or range, should become evident in both efferent components. A statistically
different reduction in the reﬂex range and gainm for TPR control was not evident in
this study. Moreover, asserting the alternative hypothesis, that TPR control is m;
altered by heart failure, can be made with reasonable conﬁdence (power = .88 and .91
respectfully). Thus, a significant contribution from a change in receptor range and/or
gain is generally incompatible with the findings of these experiments. The exception
would require a concomitant increase in reﬂex vascular control at some other site to

counterbalance a receptor abnormality (see below).

QantraLSites - The reduction in HR“, COrm and 8P“m could conceivably be
due to a reduction in the command signal analog of the control system model.
Although the existence of an anatomic or even functional equivalent of this 'baroreﬂex
command center' has never been conclusively demonstrated, it would presumably
exist proximal to the medullary autonomic efferent centers through which it acts.
Because this hypothetical site is 'common' to both efferent limbs, the lack of a

significant change in TPRM would tend to exclude this as a source of reﬂex

63

dysfunction. The differential or selective reﬂex impairment observed in heart failure is
much more suggestive of an efferent source of reﬂex dysfunction. But at what level
does this efferent dysfunction exist, brainstem or spinal autonomic centers, efferent
nerve transmission, or an end-organ insensitivity to autonomic stimulation? The data
provided in this study do not allow a deﬁnitive answer to this question. The fact that
reﬂex control of cardiac output (presumably via reﬂex control of the heart, see below) is
selectively impaired might suggest a parasympathetic lesion because of its preferential
cardiac distribution. Presumably, depressed parasympathetic withdrawal could
account for the reduced capacity to increase HR, CO, and ultimately BP in response to
baroreceptor unloading. Alternatively, the observed baroreﬁex abnormalities could be
due to a decrease in the range and gain of cardiac sympathetic nerve stimulation.
Dibner-Dunlap, in a preliminary report, asserted that the 'central" gain of baroreﬂex
control of renal sympathetic nerve activity remains normal in this model of heart failure.
In addition, current evidence strongly suggests that baroreﬂex control of sympathetic
efferent activity is not end-organ specific. Arterial baroreﬂex control of sympathetic
stimulation to both the heart and vasculature appears to arise from a common
medullary site, the C1 area of the rostral ventro-Iateral medulla. (Loewy, 1986; Refs,
1986) This is also supported by neurophysiologic studies of central baroreﬂex
modiﬁcation. Interventions that increase or decrease reﬂex sympathetic activity,
characteristically do so to both the heart and vasculature simultaneously, suggesting a
common origin. (Fleis, 1986) As a result, a central sympathetic efferent 'lesion' would
require a second 'lesion', an increase in vascular sympathetic sensitivity, to maintain

the normal TPR responses observed in this study.

Forster’s (1989) results would seem to support such a “two-lesion" hypothesis since
she reported an enhanced vascular responsiveness to alpha agonists in vitro.

However, when in viva vascular responsiveness has been examined in this model of

s4

heart failure, the results indicate either no change (Kaiser, 1989), or even a slight
depression (\Mlson, 1988) in vascular sensitivity to sympathetic nerve stimulation,
norepinephrine infusions, and nitroglycerin applications.

The unlikellhood of a central site of reﬂex dysfunction would tend to exclude those
potential inhibitory inﬂuences thought to act via central interactions. Thus, Increased
circulating levels of atrial natriuretic peptides, vasopressin, and angiotensin II as well
as enhanced cardiopulmonary receptor activity would be unlikely mechanisms for the
reﬂex impairment observed. Evidence against a major role for ventricular receptor
activation in heart failure is also suggested by the lack of the generalized pattern of
reﬂex impairment (both cardiac output and TPR) characteristically seen when

ventricular receptors are stimulated chemically. (Denison, 1987)

Was - For the reasons presented, the preferential impairment in reﬂex
cardiac output control during heart failure is most supportive of a cardiac site as the
major cause for reﬂex dysfunction. it should be pointed out though, that a depression
in reﬂex control of the bean is not essential, nor necessarily totally responsible, for
defective cardiac output control. Other reﬂex effects on preload, for example, could be
contributory. That the heart is at least a component of this effect is clearly shown by
the decreased range and gain of heart rate control. Defective control of stroke volume
is also suggested. Although not statistically signiﬁcant (p= .06), the magnitude of the
impairment in reﬂex control of heart rate was always less than the magnitude of
impairment for cardiac output, implicating defective stroke volume responses. This is
also shown by the stimulus-response curves generated during pacing in which heart
failure depresses the residual range of cardiac output and ultimately blood pressure.
Any change in reﬂex cardiac output responses under these circumstances must be
due to changes in stroke volume. The ability to increase stroke volume in response to
baroreceptor hypotension is mediated, at least partially by an increase in contractility.

Shimizu (1980), for example demonstrated an increase in left ventricular dp/dt in

65

response to bilateral carotid occlusion in anesthetized cats that was partially abolished
by cardiac sympathetectomy and completely abolished by B-adrenergic blockade.
Hainsworth, and others have reported similar ﬁndings in both anesthetized and
conscious animals and have also excluded a cerebral ischemic response as a cause

for the observed inotropic response. (Hainsworth, 1972; Aylward, 1985)

The importance of reﬂex-mediated changes in venous compliance and venous
return to the changes in cardiac output are more controversial. Shoukas (1973)
continues to assert, primarily on theoretical grounds, that the increase in venous return
accompanying carotid sinus hypotension is important in maintaining the cardiac
output response. Browse and Donald (1966) however, convincingly showed that the
actual magnitude of this increase in venous return is relatively small in dogs and that
the reﬂex cardiac output and pressor responses are not changed when this increase in
central blood volume is experimentally diverted. Nevertheless, blunted passive or
active (adrenergic) changes in venous compliance might assume greater relative
importance in subjects with heart failure and therefore cannot be totally excluded as a

possibility.

I“ D IEIIiIliI' i
Badmdﬁanﬂacﬂesmnsivenass

The reﬂex impairment associated with this model of congestive heart failure
develops rapidly and resolves quickly if heart failure is reversed. The capacity for
partial reversal of reﬂex dysfunction within days is in contrast to that reported by White
(1981) in dogs with high-output heart failure. He found sustained depressions in the
Hl-SAP slopes for up to 8 months after surgical closure of an aorta-caval ﬁstula. He
proposed that structural damage to the baroreceptors and/or vessel walls (aorta,

carotid sinus) was the most likely explanation for this ﬁnding. Others have speculated

66

that the increased stroke volume and pulse pressure in this model of high-output
failure may have damaged aortic receptors, a situation unlikely to occur in low-output

failure.

it could be argued that the relatively short duration of failure in this study was not
sufﬁcient to induce long-lasting reﬂex impairments. However, Ellenbogen (1989)
recently reported a similar time course for recovery of baroreﬁex heart rate responses
(atrial remnant) after cardiac transplantation in patients with chronic end-stage heart
failure. Structural changes, particularly in the heart, may still be contributory but the
rapidity with which the major reﬂex deﬁcits apparently develop and resolve suggests
an important rapidly adapting mechanism, such as a neural endocrine or biochemical

change.

This study does not address these issues although existing evidence suggests
several possibilities. As mentioned, the observed abnormalities in heart rate control
are potentially due to either defective parasympathetic control, ineffective sympathetic
nerve activity, or some combination of both. A reduction in the cardiac
responsiveness to these neural inputs must also be considered. The possibility that
autonomic neurotransmitter release is reduced in failing hearts has been suggested in
studies documenting a reduction in the cardiac response to experimental sympathetic
nerve stimulation. (Covell, 1966) The reduction in myocardial norepinephrine content
and relevant biosynthetic enzyme activity has also fostered the idea that sympathetic
neurotransmitter preparation and/or release is impaired in heart failure. (Chidsey 1964,
1965; Pool, 1967; Sole, 1977) There may also exist, a reduction in the myocardial
cellular response to released norepinephrine. Down regulation and/or a reduction in
the number of beta-adrenergic receptors in failing hearts has been well publicized.
(Bristow, 1982; Vatner, 1985) Whether post-receptor responses are also abnormal
remains uncertain. Erne (1988) documented attenuated cardiac responses to

isoproterenol infusions in men with advanced heart failure of different etiologies. The

67

heart rate responses to intravenous forskolin infusions however, were normal
suggesting an intact post-receptor mechanism for cardiac stimulation. Newman
(1978) found a similar reduction in responses to beta-agonists but also found a
depression in the length-tension relationship in response to glucagon, a

non-adrenergic stimulant of cAM P.

Evidence for other humoral inﬂuences on cardiac function is also slowly starting to
emerge. Exogenous vasopressin can depress cardiac output in sino-aortic
denervated animals. This effect appears to be a direct effect on the myocardium rather
than just a change in afterload since the magnitude of cardiac depression is much
greater than for equivalent pressor doses of phenylephrine. (Webb, 1986)

Conceivably, the high circulating levels of vasopressin present in heart failure might
have a similar depressant effect. While this might exert a depressant effect on resting
cardiac performance, it would not be a likely explanation for the abnormalities in reﬂex
function seen in heart failure since vasopressin sensitizes reﬂex function in normal
subjects and acts as sites (receptors and central) that are inconsistent with the

selective impairment observed in this study.

Exogenous atrial natriuretic peptides are also increased in congestive heart failure.
interestingly, exogenous administrations of ANP has been observed to reduce the
tachycardic but not the bradycardic response to baroreﬁex manipulations, a ﬁnding
similar to that observed in this study. (Ebert, 1988) Nevertheless, the presumed central
site of action of ANP on baroreﬂex function mentioned earlier would tend to exclude it

as a signiﬁcant contributor to baroreﬁex dysfunction in this model.

Heart failure can also be associated with increased circulating levels of natural
opiates. The evidence for a cardiovascular regulatory role of these opiates is also
slowly increasing. Administration of opiate receptor antagonists such as nalmefene to
dogs with right heart failure for example, results in a signiﬁcant increase in cardiac

output and heart rate that is sympathetically mediated. (Liang, 1987) Another recent

68

brief report describes partial restoration of baroreﬂex control of heart rate in heart
failure after the administration of the mu-opiate antagonist, naloxone. (Sakamoto,
1988) The mechanism may involve an opiate-mediated presynaptic inhibition of
sympathetic norepinephrine release although central mechanisms have also been

postulated. (Dubocovich, 1980)

Angiotensin II enhancement of baroreﬂex function has already been mentioned
although these effects may be species-dependent. Somewhat surprisingly, several
studies have suggested that captopril and other converting enzyme inhibitors can
actually increase the sensitivity of baroreﬂex control of heart rate in humans with
naturally occurring heart disease. (Mancia, 1982; lbsen, 1983;) This would suggest a
depressant effect of angiotensin ll. These observations are not universal though, since

other investigators have reported opposing results. (Heavy, 1978)

Changes in the biochemistry of cardiac ceilqu function could also cause cardiac
depression with a time course similar to that observed in this study. In response to an
increase in load, myocardial protein can increase by as much as 50% within the ﬁrst 48
hours. Mitochondrial cytochrome proteins initially increase, but within 3 days, their
concentrations per gram of cardiac protein and cardiac weight are reduced.
(Rabinowitz, 1973) Although dogs and humans apparently have a limited capacity to
shift to different myosin isoenzymes in response to cardiac load, a reduction in total
myoﬁbrillar protein has been reported in naturally occurring forms of myocardial
failure. (Swynghedauw, 1986; Pagani, 1988) Changes in intracellular calcium kinetics
are yet another potential mechanism for rapid development of cardiac dysfunction and
perhaps reﬂex function. Calcium uptake by the sarcoplasmic reticulum for example,
can be severely attenuated within minutes to hours of ischemic damage to the

myocardium.

69

Whether these neural, endocrine, or biochemical changes are reversible and
whether they play any role in the baroreﬁex Impairments observed in this model of

heart failure remains speculative at this point.

Whatever the mechanism involved, the results of this study suggest that it
originates in the heart itself rather than some other site in the baroreﬂex arc. Although
not deﬁnitive, the results do not support the common assertion that heart failure-
related impairments in baroreﬂex function are related to some fundamental
derangement in the neural elements of the reﬂex. If so, is the reﬂex really impaired or
does this merely represent end-organ dysfunction? Can the two be considered
separate? Hypothetically, if there were no end-organ response to the neural reﬂex
components, there could be no homeostatic reﬂex adjustments of the circulation. This
illustrates the necessity of including the end-organs as components in the overall
reﬂex. As a result, even in the absence of signiﬁcant neural impairments, a depression

in end-organ responsiveness can be interpreted as a cause for baroreﬂex dysfunction.

SUMMARY

Impairment in arterial baroreﬂex regulation of blood pressure has been assumed for
years based on evidence that baroreﬂex control of heart rate was depressed.
However, heart rate responses alone can be unreliable in predicting overall reﬂex
function, and in predicting the characteristics of reﬂex control of the other efferent
components such as cardiac output and total peripheral resistance. The results of this
study using an open-loop experimental design conﬁrm previous reports of a
depression in arterial baroreﬂex control of heart rate in conscious dogs with heart
failure. The results of this study also provide additional information not previously

available:

1) Congestive heart failure rapidly induces a dramatic depression in the
open-loop gain and range of carotid baroreﬂex control of mean arterial

pressure in conscious dogs.

2) This depression in reﬂex control of blood pressure is due to a depression in
reﬂex control of cardiac output. Baroreﬂex control of total peripheral

resistance is not affected in this model of heart failure.

70

71

3) The changes in arterial baroreﬁex function associated with heart failure
develop quickly and resolve rapidly as dogs recover from heart failure. This
suggests a rapidly adapting mechanism such as a neural, endocrine, or
biochemical change (as opposed to a structural change) as the predominant

cause for reﬂex impairment.

APPENDIX

APPENDIX A

 

The capacity of the arterial baroreﬂex to attenuate the magnitude of a disturbance in
blood pressure is a function of the range and gain of the reﬂex. If the magnitude of the
disturbance is relatively small compared to the range of the reﬂex, less than 30% for
example, then only the gain of the reﬂex is important in determining the degree of
attenuation. Since the gain of the reﬂex changes as a function of the prevailing blood
pressure in the closed-loop state, the average gain will be used in this example for
simplicity.

In the closed-loop state, any cardiovascular disturbance that alters blood pressure
is sensed by the baroreceptors which initiate rapid reﬂex compensations to attenuate
the magnitude of this disturbance. The degree to which a relatively small disturbance
(deﬁned above) is attenuated is related to the open-loop gain (6,) of the reﬂex

according to the following equation. (Houk, 1980)

M A p dist Equation 1

10 =—
( + J AMAP

72

73

where MAP“ refers to the magnitude of the disturbance (the change in blood
pressure that would occur in the absence of any baroreﬂex compensation) and delta
MAP refers to the change in blood pressure actually observed (including the effects of

baroreﬂex compensation). Rearranging Equation 1 gives:

(l+G,,)(AMAP) = MAP“, Equation2

If a standard blood pressure disturbance (MAPdM) is applied to a representative dog in

both normal (avg. gain = -2.7) and heart failure (avg. gain = -1.1) condition then:

(1+Gonorm)(AM/1Pnorm) " MAPam ' (1+GO.;)(AMAPM) Equation3

A ﬁnal rearrangement gives:

(“00,”) _ (AM/1PM) Equation4
(1 +00”) (AM/rpm.)

Substituting the known values for the group averaged day "-21 gains in this study

gives:

-3.7 __ (AM/1PM) Equations

—2.1 ' (AMAPWm)

 

or (rounding to 2 signiﬁcant digits)

(AM/1PM)
(AMAPROI'M)

 

11.8 =

Thus, the observed change In blood pressure in response to a standard disturbance

would be 1.8 times greater in heart failure than under normal conditions.

LIST OF REFERENCES

LIST OF REFERENCES

Abraham A: The structure of baroreceptors in pathologic conditions In man. in
Baroreceptors and Hypertension. P. Kezdi, ed. Pergamon Press Ltd.,
Oxford England, 1967.

Angeli-James JE, George MJ: Carotid sinus baroreceptor reﬂex control of the
circulation in medial sclerotic and renal hypertensive rabbits and its modiﬁcation by
aortic baroreceptors. Circ Res, 47:890. 1980.

Armstrong PW, Stopps TP, Ford SE, DeBold AJ: Rapid ventricular pacing In the dog:
Pathophyslologic studies of heart failure. Circ, 74:1075, 1986.

Aylward P, McR'rtchie R, West M, Chalmers J: Relative roles of vagal and sympathetic
effector mechanisms in the baroreﬂex control of myocardial contractility in conscious
rabbits. Pﬂugers Arch, 403:21, 1985.

Bagshaw RJ, Peterson LH: Sympathetic control of the mechanical properties of the
canine carotid sinus. Am J Physiol, 222:1462, 1972.

Bernhardson CS: Type l error rates when multiple comparison procedures follow a
signiﬁcant F test of ANOVA. Biometrics, 31:229. 1975.

Bolter C, Ledsome J: Inﬂuence of cervical sympathetic nerve stimulation on carotid
sinus baroreceptor afferents. Experentia 36:1301, 1980.

Bristow MR, Ginsburg MR, et ai: Decreased catecholamine sensitivity and
B-adrenergic receptor density in failing human hearts. N Eng J Med, 307:205, 1982.

Browse N, Donald D, Shepherd J: Role of the veins in the carotid sinus reﬂex. Am J
Physiol, 210:1424, 1966.

Campese VM, Romoff M, DeQuattro V, Massry 86: Relationship between plasma
catecholamines, plasma renin activity, aldosterone, and arterial pressure during
postural stress in normal subjects. J Lab Clin Med, 95:927. 1980.

Chidsey C, Braunwald E, Morrow A: Catecholamine excretion and cardiac stores in
congestive heart failure. Am J Med, 39:442, 1965.

74

75

Chidsey C, Kaiser G, Sonnenblick E, Spann J, Braunwald E: Cardiac norepinephrine
stores in experimental heart failure in dogs. J Clin Invest, 43:2386, 1964.

Chidsey CA, Braunwald E, Morrow AG, Mason DT: Myocardial norepinephrine
concentrations in man: Effects of reserpine and congestive heart failure. N Eng J Med,
269:653, 1963.

Cody RJ, Franklin KW, Kluger J, Laragh JH: Mechanisms governing the postural
response and baroreceptor abnormalities In chronic congestive heart failure: Effects of
acute and long term converting enzyme inhibition. Circ, 66:135. 1982.

Coleman HN. Taylor RR, Pool PE, Whipple GH, et al: Congestive heart failure
following chronic tachycardia. Am H J, 81:790. 1971.

Cottle MK: Degeneration studies of primary afferents of IX and Xth cranial nerves in
the cat. J Comp Neurol, 122:329. 1964.

Covell JW, Chidsey CA, Braunwald E: Reduction in cardiac response to
postganglionic sympathetic stimulation in experimental heart failure. Circ Res, 19:51,
1966.

Cowley AW, Merrill D, Osborn J, Barber BJ: Inﬂuence of vasopressin and angiotensin
on baroreﬂexes in dogs. Circ Res, 54:163. 1984.

Cox RH, Bagshaw RJ: Inﬂuence of anesthesia on the response to carotid hypotension
in dogs. Am J Physiol, 237:H424, 1979.

Dallal GE: Survival: A Supplementary Module for SYSTAT. Evanston lL, Systat inc.,
1988.

Denison AL, Hull SS, Stephenson RB, Cornish KG, Zucker IH: Left ventricular receptor
stimulation attenuates baroreﬁex control of peripheral resistance in conscious dogs.
Circ, 72:lIl-224, 1985.

Denison AL: interaction of the left ventricular receptor reﬂex with the carotid
baroreceptor reﬂex In conscious dogs. Doctoral Dissertation. Michigan State
University. Department of Physiology, 1987.

Dibner-Dunlap M, Thames M: Arterial baroreﬂex control of renal sympathetic nerve
activity is preserved in heart failure. Circ, Supp 78:Il-365, 1988.

Dubocovich M, Langer S: Pharmacological differentiation of presynaptic inhibitory
alpha-adrenoceptors and opiate receptors in the cat nictitating membrane. Br J
Pharmac, 70:383. 1980.

Ebert TJ, Cowley AW: Atrial natriuretic factor attenuates carotid baroreﬂex mediated
cardloacceleration in humans. Am J Physiol, 254:R590, 1988.

Ebert TJ, Kotrly KJ, et al: Halothane anesthesia attenuates cardiopulmonary baroreﬂex
control of peripheral resistance in humans. Anesthesiology, 63:668. 1985.

76

Eilenbogen KA, Mohanty PK, Szentpetery S, Thames M: Arterial baroreﬂex
abnormalities in heart failure. Circ. 79:51 , 1989.

Eme P, Lipkin D. Maseri A: Impaired beta-adrenergic receptor and normal
postreceptor responsiveness in congestive heart failure. Am J Cardiol, 61:1132, 1988.

Forster C. Carter 8, Armstrong P: Vascular smooth muscle responsiveness to
noradrenaline and phenylephrine following experimental heart failure in dogs.
Cardiovas Res, 23:489. 1989.

Francis 6: Neuroendocrine manifestations of congestive heart failure. Am J Cardiol,
62:9A. 1988.

Goldsmith SR, Francis GS. Levine TB. Cowley AW. Cohn JN: impaired response of
plasma vasopressin to orthostatic stress in patients with congestive heart failure. J Am
Coll Card, 2:1080. 1983.

Greenberg TT. Richmond WH, Stocking RA, et al: impaired atrial receptor responses
in dogs with heart failure due to tricuspid insufficiency and pulmonary artery stenosis.
Circ Res, 32:424. 1973.

Guo GB. Sharabl FM. Abboud FM. Schmid PG: Vasopressin augments baroreﬂex
inhibition of lumbar sympathetic nerve activity in rabbits. Circ. 66:Supp "-35. 1982.

Guo GM. Thames MD, Abboud FM: Arterial baroreﬂexes in renal hypertensive rabbits:
selectivity and redundancy of baroreceptor inﬂuence on heart rate, vascular resistance,
and lumbar sympathetic nerve activity. Circ Res, 53:224. 1983.

Hainsworth R. Karim F: Left ventricular inotropic and peripheral vasomotor responses
from independent changes in pressure in the carotid sinuses and cerebral arteries in
anesthetized dogs. J Physiol, 228:139, 1973.

Higgins CB. Vatner SF, Eckberg DL. Braunwald E: Alterations in the baroreceptor
reﬂex In conscious dogs with heart failure. J Clin Invest, 51:715. 1972.

Hirsh AT, Dzau V. Creager MA: Baroreceptor function in congestive heart failure:
Effect on neurohumoral activation of regional vascular resistance. Circ Supp IV: lV-36,
1987.

Houk JC: Homeostasis and control principles. In: Medical Physiology. Fourteenth
edition. ed. Mountcastle. V.B., CV Mosby Co., St. Louis, pg. 246. 1980.

Hull SS, Stephenson RB: Central angiotensin II augments baroreﬁex gain in
conscious dogs. Circ. 72:lll-18, 1985.

Ibsen H. Egan B. Osterziel K. Vander A, Julius S: Reﬂex hemodynamic adjustments
and baroreﬂex sensitivity during converting enzyme inhibition with MK-421 in normal
humans. Hypertension. 5:I184, 1983.

lshikawa N. Kaliman CH, Sagawa 1c Rabbit carotid sinus reﬂex under pentobarbital,
urethane. and chloralose anesthesia. Am J Physiol, 246:H696, 1984.

77

Kassis E: Cardiovascular response to orthostatic tilt in patients with severe congestive
heart failure. Cardiovas Res, 21:363. 1987.

Keith iC. Kidd C, Malpus CM, Penna PE: Reduction of baroreceptor impulse activity
by sympathetic nerve stimulation. J Physiol, 238:61P, 1974.

Kendall MG. Stuart A: The Advanced Theory of Statistics, Vol 2, 3rd ed, New York:
Hafner Publishing Co, 1973.

Lam RL, Tyler HR: Electrical responses evoked in the visceral afferent nucleus of the
rabbit by vagal stimulation. J Comp Neurol. 97:21, 1952.

Levine TB. Francis GS. Goldsmith SR. Cohn JN: The neurohumoral and
hemodynamic response to orthostatic tilt in patients with CHF. Circ, 57:1070, 1983.

Liang C, imai N. Stone C, Woolf P, Kawashima S. Tuttle R: The role of endogenous
opioids In congestive heart failure: effects of nalmefene on systemic and regional
hemodynamics in dogs. Circ, 75:443. 1987.

Mancia G, Parati G. Pomidossi G, Grassi G, et al: Modiﬁcation of arterial baroreﬂexes
by captopril in essential hypertension. Am J Cardiol, 49:1415, 1982.

Moe G, Stopps T, Howard R, Armstrong PW: Early recovery from heart failure:
Insights into the pathogenesis of experimental chronic pacing-induced heart failure. J
Lab Clin Med, 172:426. 1988.

Mohanty PK, Arrowood JA, Ellenbogen KA, Thames MD: Neurohumoral and
hemodynamic effects of lower body negative pressure in patients with congestive heart
failure. Am H J. 118:78. 1989.

Motulsky H, Ranshas LA: Fitting curves to data using nonlinear regression: a practical
and nonmathematical review. FASEB J, 1:365, 1987.

Munch P. Thoren P, Brown A: Dual effects of norepinephrine and mechanism of
baroreceptor stimulation. Circ Res, 61:409. 1987.

Newman WH: A depressed response of left ventricular contractile force to
isoproterenol and norepinephrine in dogs with congestive heart failure, Am H J,
93:216. 1977.

Newman WH: Volume overload heart failure: length-tension curves, and response to
B-agonists. calcium. and glucagon. Am J Physiol, 735:H690. 1978.

Olavari MT, Levine T, Cohn J: Abnormal neurohumoral responses to nitroprusside
infusions in congestive heart failure. J Am Coll Card. 2:411, 1983.

Pagani E. AIousi A, Grant A, Older T, Dziuban S. Allen P: Changes in myoﬁbrillar
content and Mg-ATPase activity In ventricular tissue from patients with heart failure
caused by coronary artery disease, cardiomyopathy, or mitral valve insufﬁciency. Circ
Res. 63:380. 1988.

78

Parati G, Pomidossi G. Bertinieri G. DiRienzo M. Mancia G: Blood Pressure variability:
Its mechanisms and its clinical implications. in Neural Mechanisms and
Cardiovascular Disease. B. Lown. ed. Vol 5, Springer-Verlag, New York, 1986.

Pool PE. Covell J. Levitt M, Gibb J. Braunwald E: Reduction of cardiac tyrosine
hydroxyiase activity in experimental congestive heart failure. Its role in depletion of
cardiac norepinephrine stores. Circ Res. 20:349. 1967.

Quest JA, Gillis R: Carotid sinus reﬂex changes produced by digitalis. J Pharm Exp
Ther. 177:650, 1971.

Quest JA, Gillis R: Effect of digitalis on carotid sinus baroreceptor activity. Circ Res,
35:247. 1974.

Rablnowitz M: Protein synthesis and turnover in normal and hypertrophied heart. Am
J Cardiol, 31:202. 1973.

Rapundalo S. Briggs FN. Feher J: Effects of ischemia on the isolation and function of
canine cardiac sarcoplasmic reticulum. J Mol Cell Cardiol, 18:837. 1986.

Being C, Thibodeaux H. Gillis R. Kot P: Inﬂuence of baroreceptor reﬂexes in
cardiovascular responses of the dog to acetylstrophanthidin. J Pharm Exp Ther,
184:641. 1973.

Riegger GAJ, Eisner D. Krorner EP. et al: Atrial natriuretic peptide In congestive heart
failure in the dog: Plasma levels, cyclic guanosine monophosphate, ultrastructure of
atrial myoendocrine cells, and hemodynamic. hormonal, and renal effects. Circ,
77:398. 1988.

Riegger GAJ, Liebau G, Hoizchuh M. et al: Role of renin-angiotensin system in the
development of congestive heart failure in the dog as assessed by chronic converting
enzyme blockade. Am J Cardiol, 53:614. 1984.

Riegger GAJ, Liebau G: The renin-angiotensIn-aldosterone system. antidiuretic
hormone, and sympathetic nerve activity in an experimental model of congestive heart
failure in the dog. Clin Sci. 62:465. 1982.

Roweli LB: Adjustments to upright posture and blood loss. in Human Circulation,
Regulation During Physical Stress, Oxford Univ Press, New York, 1986.

Sakamoto S, Liang C: Partial restoration of baroreﬂex sensitivity by naloxone in heart
failure. Circ, Supp 78:ll-366, 1988.

Sampson S. Mills E: Effects of sympathetic stimulation on discharges of carotid sinus
baroreceptors. Am J Physiol, 218:1650. 1970.

Scher AM, Your AC: Reﬂex control of heart rate in the unanesthetized dog.
Am J Physiol, 218:780. 1970.

Schmid PG. Guo GB. Abboud FM: Different eﬂects of vasopressin and angiotensin ii
on baroreﬂexes. Fed Proc, 44:2388, 1985.

79

Shimizu T. Bishop V: Role of carotid sinus and cardiopulmonary reﬂexes on left
ventricular dP/df in cats. Am J Physiol, 238:H93. 1980.

Shoukas A, Sagawa K: Control of total systemic vascular capacity by the carotid sinus
baroreceptor reﬂex. Circ Res, 33:22. 1973.

Sinoway U. Minott JR. Davis D. et al: Delayed reversal of impaired vasodilation in CHF
after heart transplantation. Am J Cardiol, 61 :1076, 1988.

Sleight P. Wrddicombe JG: Action potentials in ﬁbers from receptors in the epicardium
and myocardium of the dog’s left ventricle. J Physiol, 181 :235, 1965.

Smyth HS. Sleight P. Pickering GW: Reﬂex regulation of arterial pressure during sleep
in man: A quantitative method of assessing baroreﬂex sensitivity. Circ Res, 24:109.
1969.

Sole M, Kamble A, Hussain M: A possible change in the rate limiting steps for cardiac
norepinephrine synthesis in the cardiomyopathic syrian hamster. Circ Res, 44:814.
1977.

Stephenson RB. Donald DE: Reﬂexes from isolated carotid sinuses of intact and
vagotomized conscious dogs. Am J Physiol, 238:H815. 1980.

Stephenson RB, Donald DE: Reversible vesicular isolation of carotid sinuses in
conscious dogs. Am J Physiol, 238:H809, 1980.

Swynghedaun B: Adaptation of muscle contractile proteins. Physio Rev, 66:710.
1986.

Teo KK, Man GC, Kappagoda CT: Interaction of left atrial receptors and carotid sinus
baroreceptors on heart rate in the dog. Am J Physiol, 248:H631. 1985.

Thames MD, Miller B, Abboud FM: Sensitization of vagal cardiopulmonary
baroreﬂexes by chronic digoxin. Am J Physiol, 243:H815. 1982.

Thoren P: Characterization of left ventricular receptors with non-medullated vagal
afferents In the cat. Circ Res, 40:231. 1977.

Tomomatsu E, Nishi K: increased activity of carotid sinus baroreceptors by
sympathetic stimulation and norepinephrine. Am J Physiol, 240:H650, 1981.

Undesser KP, Hasser E, Johnson AK, Haywood JR, Bishop VS: Lesions of the area
postrema prevent enhancement of arterial baroreﬂex inhibition of sympathetic activity
by vasopressin in conscious rabbits. Fed Proc, 43:1067. 1984.

Vatner DE. Vatner SF, Fuji AM. Homcy CJ: Loss of high afﬁnity cardiac B-adrenergic
receptors in dogs with heart failure. J Clin invest, 76:2259. 1959.

Vatner SF. Higgins CB, Braunwald E: Sympathetic and parasympathetic components
of reﬂex tachycardia induced by hypotension in conscious dogs with and without heart
failure. Cardiovasc Res. 8:153, 1974.

80

Whipple GH, Shefﬁeld LT, Woodman LT. Theophilis EG, Friedman S: Reversible
congestive heart failure due to rapid ventricular stimulation of the normal heart. Proc
New Eng Cardiov Soc, 20:39. 1962

White CW: Abnormalities in baroreﬂex control of heart rate in canine heart failure. Am
J Physiol, 204:H793, 1987.

Wilkinson L: Nonlinear Module, in SYSTAT: The System For Statistics. Evanston IL,
Systat Inc. 1988.

Wilson JR. Douglas P, Hickey WF. Lanoce V, Ferraro N: Experimental congestive
heart failure produced by rapid ventricular pacing in the dog: Cardiac effects. Circ.
75:857. 1987.

Wilson JR. Matthai W. Lanoce V, Frey M, Ferraro N: Effect of experimental heart failure
on peripheral sympathetic vasoconstriction. Am J Physiol, 254:H727, 1988.

Wood JE, Litter J, Viﬁlkins RW: Peripheral venoconstriction in human congestive heart
failure. Circ, 13:524. 1956.

Zelis R, Delea CS. Coleman HN. Mason DT: Arterial sodium content in experimental
congestive heart failure. Ciro. 41:213. 1970.

Zelis R: The contribution of local factors to the elevated venous tone of congestive
heart failure. J Clin Invest, 54:219. 1974.

Zucker lH, Niebauer MJ, Holmberg MJ: Aortic baroreceptor characteristics in dogs
with chronic volume overload. The Physiologist, 26:A-103, 1983.