Failure of propranolol to prevent tilt-evoked systemic vasodilatation, adrenaline release and neurocardiogenic syncope.

Abstract

In patients with neurocardiogenic syncope, head-up tilt often evokes acute loss of consciousness accompanied by vasodilatation, increased plasma adrenaline and systemic hypotension. Since hypotension increases adrenaline levels and adrenaline can produce skeletal muscle vasodilatation by activating beta2 receptors, adrenaline might induce a positive feedback loop precipitating circulatory collapse. We hypothesized that propranolol, a non-selective beta-blocker, would prevent adrenaline-induced vasodilatation and thereby prevent syncope. Eight subjects with recurrent neurocardiogenic syncope and previously documented tilt-induced syncope with elevated plasma adrenaline levels participated in the present study. Subjects underwent tilt table testing after receiving oral propranolol or placebo in a double-blind randomized crossover fashion. Haemodynamic and neurochemical variables were measured using intra-arterial monitoring, impedance cardiography, arterial blood sampling and tracer kinetics of simultaneously infused [3H]noradrenaline and [3H]adrenaline. The occurrence of tilt-induced neurally mediated hypotension and syncope, duration of tilt tolerance, extent of the decrease in SVRI (systemic vascular resistance index) and magnitude of plasma adrenaline increases did not differ between the propranolol and placebo treatment phases. SVRI was inversely associated with fractional increase in plasma adrenaline during both phases. One subject did not faint when on propranolol; this subject's response is discussed in the context of central effects of propranolol. In this small, but tightly controlled, study, propranolol did not prevent tilt-induced vasodilatation, syncope or elevated plasma adrenaline.

Full text

Page 1

Clinical Science (2006) 111, 209–216 (Printed in Great Britain) doi:10.1042/CS20060017
209
Failure of propranolol to prevent tilt-evoked
systemic vasodilatation, adrenaline release
and neurocardiogenic syncope
Basil A. ELDADAH∗, Sandra L. PECHNIK∗, Courtney S. HOLMES∗, Jeffrey P. MOAK†,
Ahmed M. SALEEM∗and David S. GOLDSTEIN∗
∗Clinical Neurocardiology Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health,
Bethesda, MD 20892-1620, U.S.A., and †Department of Cardiology, Children’s National Medical Center,
Washington DC 20010-2970, U.S.A.
ABSTRACT
In patients with neurocardiogenic syncope, head-up tilt often evokes acute loss of consciousness
accompanied by vasodilatation, increased plasma adrenaline and systemic hypotension. Since
hypotension increases adrenaline levels and adrenaline can produce skeletal muscle vasodilatation
by activating β2receptors, adrenaline might induce a positive feedback loop precipitating circu-
latory collapse. We hypothesized that propranolol, a non-selective β-blocker, would prevent
adrenaline-induced vasodilatation and thereby prevent syncope. Eight subjects with recurrent
neurocardiogenic syncope and previously documented tilt-induced syncope with elevated plasma
adrenaline levels participated in the present study. Subjects underwent tilt table testing after
receiving oral propranolol or placebo in a double-blind randomized crossover fashion. Haemo-
dynamic and neurochemical variables were measured using intra-arterial monitoring, imped-
ance cardiography, arterial blood sampling and tracer kinetics of simultaneously infused
[3H]noradrenaline and [3H]adrenaline. The occurrence of tilt-induced neurally mediated hypo-
tension and syncope, duration of tilt tolerance, extent of the decrease in SVRI (systemic vascular
resistance index) and magnitude of plasma adrenaline increases did not differ between the
propranolol and placebo treatment phases. SVRI was inversely associated with fractional increase
in plasma adrenaline during both phases. One subject did not faint when on propranolol; this
subject’s response is discussed in the context of central effects of propranolol. In this small,
but tightly controlled, study, propranolol did not prevent tilt-induced vasodilatation, syncope or
elevated plasma adrenaline.
INTRODUCTION
Neurocardiogenicsyncope(alsoknownasvasovagalsyn-
cope, neurally mediated hypotension and the common
faint) is the most common cause of acute reversible
loss of consciousness [1,2]. Neurocardiogenic syncope
has been classified into three types: cardioinhibitory,
vasodepressor and mixed, based on changes in heart rate
at the time of syncope during upright tilt table testing [3].
All three types feature acute large-magnitude decreases
in blood pressure leading to loss of consciousness.
Mechanisms underlying the hypotension attending
neurocardiogenic syncope have been considered for sev-
eral decades [4]. It is now clear that vagally induced
bradycardia cannot account completely for the decreased
blood pressure. Neither atropine nor cardiac pacing
consistently prevents hypotension or syncope [5,6].
Moreover, syncope can occur in heart transplant
Key words: adrenaline, β-blocker, neurally mediated hypotension, tilt testing, vasovagal syncope.
Abbreviations: MAP, mean arterial pressure; SVRI, systemic vascular resistance index.
Correspondence: Dr Basil A. Eldadah (email eldadahb@ninds.nih.gov).
C ?2006 The Biochemical Society

Page 2

210
B. A. Eldadah and others
Table 1
∗Syncope type is according to the modified VASIS (Vasovagal Syncope International Study) classification [3] as follows: 1, mixed; 2B cardioinhibition with asystole;
3, vasodepressor. Types are listed for placebo and propranolol phases respectively.
Characteristics of the subjects
Propranolol Tilt duration (min)
Subject numberAge (years) Gender Weight (kg)Height (cm)PhaseDoseSyncope type∗
PlaceboPropranolol
#1
#2
#3
#4
#5
#6
#7
#8
45
36
63
31
18
46
50
25
Female
Female
Female
Female
Male
Female
Female
Female
76
64
73
60
70
54
92
52
163
168
157
165
177
166
162
157
A
B
B
A
A
A
B
B
20 mg every 5 h
40 mg every 6 h
60 mg every 6 h
20 mg every 6 h
20 mg every 8 h
40 mg every 8 h
60 mg every 6 h
40 mg every 6 h
3, 1
1, 3
3, 1
?, none
3, 1
1, 1
3, 1
2B, 2B
36
31
20
2
19
13
13
19
16
16
7
40
18
7
32
28
recipients in whom cardiac innervation has been severed
[7].
Systemic vasodilatation without a compensatory in-
crease in cardiac output may explain the hypotension
attending syncopal reactions in at least some patients
[8–12]. Such vasodilatation probably accounted for the
observation in 1793 by the English surgeon John Hunter
that the colour of venous blood from a syncopal patient
was ‘a fine scarlet’ [13], as skeletal muscle vasodilatation
probably arterialized the patient’s venous blood. Several
mechanisms probably underlie vasodilatation, including
withdrawal of sympathetic nervous system outflow
[14–17] and high circulating adrenaline levels [18–22].
Adrenalinestimulatesvasodilatoryβ2receptorsonskele-
tal muscle vascular cells, and inhibition of adrenaline at
these receptors might prevent hypotension and syncope.
Consistent with this theory, a meta-analysis of 24 trials
of β-blockers for neurocardiogenic syncope showed
that non-selective blockers were more effective than β1-
selectiveblockersinpreventingsyncope[23].Conversely,
hypotension is well known to increase circulating
adrenaline levels [24]. Therefore inhibition of adrenaline
might prevent a neurocirculatory positive feedback
loop, where adrenaline would cause vasodilatation and
hypotension that, in turn, would lead to further adren-
aline release. In the present study, we tested the hypo-
thesis that oral administration of the non-selective
β-blocker propranolol would attenuate decreases in
systemic vascular resistance and thereby preserve sys-
temic blood pressure and prevent tilt-induced neurocar-
diogenic syncope.
METHODS
Study subjects
The subjects were eight patients with a history of re-
current syncope and previously documented tilt-induced
neurocardiogenic syncope associated with elevated
plasma adrenaline levels. Prior evaluation excluded pri-
mary cardiac or neurological causes of syncope. Sym-
pathetic neurocirculatory failure was excluded by the
absence of persistent consistent orthostatic hypotension
and by a normal pattern of beat-to-beat blood pressure
associated with performance of the Valsalva manoeuvre
[25]. Table 1 lists other clinical characteristics.
Caffeine-containing beverages, cigarettes and alcohol
were not allowed during the study. All potentially inter-
fering medication, such as non-investigational β-block-
ers, midodrine, and certain antidepressants, were tapered
and discontinued prior to entering the study; all other
medication continued without changes in dosing.
This study was approved by the Institutional Review
BoardoftheNationalInstituteofNeurologicalDisorders
and Stroke and was in accordance with the principles
outlinedintheDeclarationofHelsinki.Eachsubjectgave
informed written consent prior to participation.
Study protocol
Subjects first underwent a dose-finding period with
escalating doses of propranolol administered orally.
Supinebloodpressureandheartrateweremeasuredprior
to, and 2 h after, each dose of propranolol. Dosage was
increased after each 24 h period to allow enough time to
achieve steady-state plasma concentrations (mean half-
life of propranolol in plasma is 3–4 h [26]). The target
dosing schedule was determined by a 10% decrease in
supine resting heart rate from baseline (Table 1).
After a washout period lasting at least 2 days, subjects
began receiving propranolol or placebo in randomized
double-blind fashion at the target dosing schedule.
Administration continued for 72 h. At 1–2 h after the
last dose, subjects underwent upright tilt table testing
(described below). After a washout period of at least
2 days, subjects received propranolol or placebo in a
crossover fashion, and administration of the second
treatment continued for another 72 h at the target dosing
schedule. At 1–2 h after the last dose, subjects underwent
a second tilt table test.
C ?2006 The Biochemical Society

Page 3

Propranolol and syncope
211
Tilt table testing
Subjects attended the clinical laboratory in the morning
after fasting for ?6 h. Peripheral intravascular catheters
were placed in a brachial or radial artery and two ante-
cubital veins. Baseline blood was sampled through both
arterial and intravenous catheters at least 15 min after
catheter insertion for analysis of catecholamine concen-
trations. In seven subjects, l-[ring-2,5,6-3H]noradrenal-
ine and l-[N-methyl-3H]adrenaline (PerkinElmer) were
infused intravenously at approx. 0.4 µCi/min for the
duration of testing to calculate catecholamine clearance
and spillover [27]. In the present study, spillover
calculated from arterial plasma values is termed ‘total
body spillover’.
Subjects were tilted upright to 70◦on a motorized
tilt table (Colin Medical Instruments) for 40 min or
until syncope or pre-syncope developed. Blood pressure
was measured continuously through the arterial catheter.
Arterial and venous blood was sampled approximately
every 4 min and at the time of syncope or pre-syncope.
ECG and heart rate were recorded through an analogue-
to-digital converter (PowerLab 16SP; ADInstruments).
Forearm vascular resistance was measured by strain
gauge plethysmography (Hokanson). Stroke volume was
continuously measured non-invasively by impedance
cardiography (CardioDynamics) [28], which has been
validated previously [29]. Systemic vascular resistance
was calculated according to the formula: (MAP−central
venous pressure)/(stroke volume×heart rate), where
MAP is mean arterial pressure. Central venous pressure
was assumed to be 6 mmHg. SVRI (systemic vascular
resistance index) was calculated by multiplying systemic
vascular resistance by body surface area [30].
Study variables
Study variables included haemodynamic measurements,
such as blood pressure, heart rate, stroke volume and
forearm blood flow, and neurochemical measurements,
including plasma adrenaline and noradrenaline concen-
trations.Derivedvariablesincludedcardiacoutput,SVRI,
forearm vascular resistance, catecholamine clearance and
catecholamine spillover. Assays of plasma catecholamine
concentrations were performed by HPLC with electro-
chemical detection after batch alumina extraction [31].
[3H]-Labelled catecholamines were quantified by as-
saying
corresponding to the retention time with noradrenaline
and adrenaline standards [32,33]. Noradrenaline and
adrenaline spillover values were estimated according to
methods described previously [32].
3H content in the chromatographic fractions
Data analysis and statistics
Each subject was his/her own control in this crossover
study. Haemodynamic and neurochemical data were
analysedintermsofmeanabsolutevalues(Table2)andas
meannormalizeddata(Figures1,2and4).Normalization
was performed for each parameter by expressing each
subject’s results during upright tilt in either phase as a
percentage of that subject’s results at baseline on placebo.
Values are means+−S.E.M. Statistical significance was
assessed using paired two-tailed Student’s t tests.
RESULTS
Of the eight subjects, seven had tilt-induced syncope
or pre-syncope during both propranolol and placebo
treatments. One subject (#4) had a negative tilt table test
when on propranolol (Table 1). Across all subjects, the
mean duration of tolerated tilt did not differ between
placebo and propranolol phases (19+−4 min compared
with 21+−4 min respectively).
MAP, SVRI and stroke volume were significantly
lower at the end of tilt table testing than at baseline
(Figure 1 and Table 2). The extents of these decreases
weresimilarbetweentheplaceboandpropranololphases.
Cardiac output at end-tilt on propranolol was a smaller
percentage of the baseline value than that on placebo
(63+−10% compared with 81+−13% respectively; P=
0.03). Baseline and peak heart rates were significantly
lower on propranolol (Figure 1 and Table 2).
Plasma adrenaline and noradrenaline increased sig-
nificantly at end-tilt compared with baseline in both
phases (Figure 2 and Table 2). There was a trend towards
a greater increase in adrenaline during propranolol
comparedwithplacebo,whichwasstatisticallysignificant
whenonlysubjectswithpositivetilttestswereconsidered
(3317+−624%comparedwith1332+−472%respectively;
P=0.04).
With the exception of subject #4 (discussed in more
detail below), there was an inverse relationship between
SVRI and fractional change in plasma adrenaline during
both placebo and propranolol phases (Figure 3).
Baseline clearance from arterial plasma of both
adrenaline and noradrenaline significantly decreased
during propranolol (Figure 4, upper panels). There were
further significant decreases in clearance at end-tilt on
propranolol (Figure 4, upper panels). During placebo
treatment, clearance remained stable during and at the
end of upright tilting (Figure 4, upper panels).
Total body spillovers of adrenaline and noradrenaline
had an upward trend from baseline during upright tilt
in both phases (Figure 4, lower panels). Noradrenaline
increases were statistically significant while upright in
both placebo and propranolol treatments.
Patient #4 was the only one who did not develop
syncope during upright tilt while taking propranolol
(Table 1). Her MAP was maintained during upright
tilt, heart rate increased by 33 beats/min, SVRI and
respiratory rate were stable, and there was a steady
increase in arterial plasma adrenaline from 0.45 nmol/l
at baseline to 1.44 nmol/l at end-tilt. In contrast,
C ?2006 The Biochemical Society

Page 4

212
B. A. Eldadah and others
Figure 1
all of the subjects
Data for each subject were normalized to the baseline placebo condition of that
subject. Values are means+S.E.M. at supine rest (baseline), at 3–5 min prior to
end-tilt (tilt), and at end-tilt, except in (C), where the middle bar represents peak
heart rate during tilt.∗P <0.05,∗∗P <0.01 and∗∗∗P <0.001 compared
with baseline within phase; †P <0.05, ††P <0.01 and †††P <0.001
compared with placebo at the respective time point.
Haemodynamic results of tilt table tests among
while taking placebo, her heart rate increased by
90 beats/min and respiratory rate increased by more than
70 breaths/min during the first minute of tilting. She
became unresponsive after 2 min; however, her blood
pressure remained stable and SVRI exceeded the baseline
value. Despite the absence of hypotension, her plasma
adrenaline increased from 0.42 to 12.43 nmol/l.
Patient #8 had asystole leading rapidly to syncope
during both treatment phases (Table 1). While she was
on placebo, asystole lasted 20 s, whereas, on propranolol,
asystole lasted 44 s. During the screening tilt table test,
this subject did not have bradycardia or asystole.
Forearm blood flow was measured by strain gauge
plethysmography.Becauseofartifactualeffectsofsubject
movement and rapid changes in directly recorded intra-
arterial pressure at the time of syncope, compared
with the relatively long time required for blood flow
measurements, we did not have sufficient confidence in
the calculated values of forearm vascular resistance to
report them in the present study.
DISCUSSION
In the present study, oral propranolol had no effect on
systemic vasodilatation, hypotension, plasma adrenaline
levels or duration of tolerance to upright tilt in patients
with previously documented tilt-induced neurocardi-
ogenic syncope with elevated plasma adrenaline. The
findings did not support the hypothesis that propranolol
would attenuate tilt-induced acute vasodilatation and
thereby prevent syncope in such patients.
SVRI fell significantly at the time of syncope (Fig-
ure 1B), consistent with a vasodepressor mechanism, and
SVRI was inversely associated with a fractional change
in plasma adrenaline concentration (Figure 3). These
observations would be consistent with the notion that
circulating adrenaline causes vasodilatation; however, in
seven of the eight patients, inhibition of β-receptors
failed to attenuate either tilt-induced vasodilatation
or increased circulating adrenaline (Figures 1B, 2A
and 3B). The vasodilatation attending neurocardiogenic
syncope in this setting appears to reflect processes
independent of vasodilatory β2receptors. Studies have
evaluated other mediators of vascular tone in syn-
cope, including sympathoneural outflow, vasopressin,
endothelin-1, angiotensin II and nitric oxide [4]. None
of these mediators, however, can account completely for
the vasodilatation observed in syncope.
Withdrawal of sympathoneural outflow leads to
decreased vascular resistance, and propranolol would not
be expected to prevent sympathoinhibition. Indeed, in
healthy volunteers, propranolol does not affect peroneal
nerve sympathetic outflow [34]; however, it is not known
whether β-receptor inhibition acts similarly in patients
with neurocardiogenic syncope. Furthermore, as there
C ?2006 The Biochemical Society

Page 5

Propranolol and syncope
213
Table 2
propranolol phases
Data are means+−S.E.M. at baseline and end-tilt respectively, except for heart rate data, which are listed for baseline, peak while upright and at
end-tilt respectively.∗P <0.05 and∗∗P <0.01 compared with baseline within phase; ††P <0.01 and †††P <0.001 compared with placebo at the
respective time point.
Summary of the haemodynamic and neurochemical data from tilt table tests during placebo or
Placebo Propranolol
Baseline PeakEnd-tilt BaselinePeakEnd-tilt
MAP (mmHg)
Heart rate (min−1)
Stroke volume (ml)
SVRI (dyn·s−1·cm−5/m2)
Cardiac output (litres/min)
Noradrenaline (nmol/l)
Adrenaline (nmol/l)
Noradrenaline spillover (µmol/min)
Adrenaline spillover (µmol/min)
Noradrenaline clearance (ml/min)
Adrenaline clearance (ml/min)
101+−3
76+−5
73+−4
2588+−176
5.5+−0.4
1.31+−0.22
0.28+−0.07
3.10+−0.95
0.76+−0.32
2018+−307
2244+−356
67+−9∗∗
97+−23
47+−9∗∗
1524+−319∗
5.3+−0.8
4.54+−1.43∗
4.45+−1.53∗
8.50+−3.40
14.54+−7.85
1644+−161
2409+−615
97+−5
63+−4†††
75+−4
2771+−122
4.7+−0.4†††
1.65+−0.30
0.57+−0.27
3.40+−0.77
1.46+−0.80
1868+−142
2131+−192
62+−6∗∗
58+−10
64+−6∗
1838+−323∗∗
4.0+−0.3∗
5.16+−1.82
9.80+−5.45
6.68+−3.24
16.71+−11.36
1164+−163
1251+−153
124+−17∗
80+−6∗∗††
Figure 2
(bottom) in tilt table tests among all of the subjects
Data for each subject were normalized to the baseline placebo condition of that
subject. Values are means+S.E.M. at supine rest (baseline), at 3–5 min prior to
end-tilt (tilt) and at end-tilt.∗P <0.05 and∗∗P <0.01 compared with baseline
within phase.
Arterialadrenaline(top)and noradrenaline
is a lag between changes in skeletal muscle sympathetic
outflow and changes in circulating noradrenaline levels
[35], changes in arterial plasma noradrenaline may not
be an accurate reflection of sympathetic nerve activity,
particularly in such a dynamic condition as syncope [36].
It may be argued that the failure of propranolol
to prevent systemic vasodilatation resulted from suffi-
ciently high plasma adrenaline levels overcoming β2-
receptor inhibition by propranolol. Plasma adrenaline
concentrations were indeed remarkably elevated at the
timeofsyncope.Inaddition,increasesinadrenalinelevels
had a higher trend in subjects on propranolol than on
placebo, a relationship that was statistically significant
when only subjects who had positive tilt table tests were
considered. Nevertheless, the affinity of propranolol at
β2-receptors is over 900 times that of adrenaline [37].
Furthermore, we retrieved frozen plasma samples from
five of the eight subjects in the present study and had
the samples assayed for propranolol levels. Propranolol
ranged from 80–1100 nmol/l, which is several times the
mean peak concentration of adrenaline in the present
study. Moreover, across all subjects, heart rate, cardiac
output and catecholamine clearance were significantly
reduced in subjects on propranolol (Figures 1 and 4). As
such, propranolol in this study was probably sufficient
to confer significant β-antagonism.
The present findings are consistent with those in
a prospective randomized crossover study by Flevari
et al. [38], who evaluated the non-selective β-blockers
propranolol and nadolol. The results are inconsistent,
however, with other studies showing a relatively greater
benefit of non-selective β-blockers over selective β-
blockers [23].
Propranolol is known to reduce plasma clearance of
[3H]noradrenaline [39] and unlabelled adrenaline [40].
The present results fit with this literature, as propranolol
administration was associated with an approx. 20%
decreaseinclearanceofbothcatecholaminesfromarterial
C ?2006 The Biochemical Society