Heart rate variability as a prognostic tool in cardiology. A contribution to the problem from a theoretical point of view

Abstract

Recent clinical studies have proposed standard deviation of heart rate as a diagnostic tool for the outcome of cardiac infarction. Mathematical analysis of heart rate variability shows that heart rate is influenced by different frequency components derived from different parts of the autonomous nervous system. In the experimental part of this study, we investigated the possibility of calculating a variable describing the parasympathetic branch of the autonomous nervous system exclusively.
In 60 healthy volunteers, heart rate was measured to 1 millisecond during two different conditions: 5 minutes of rest, and 5 minutes of intermittent handgrip dynamometry; the latter is known to increase sympathetic arousal selectively. Heart rate was found to be lower at rest (65.9 +/- 9.7 beats per minute) than during dynamometry (72.8 +/- 10.4 beats per minute, P < .001). Respiratory sinus arrhythmia (RSA) calculated from the mean absolute differences between successive heart beats showed no significant change (3.01 +/- 1.62 beats per minute at rest versus 2.97 +/- 1.30 beats per minute during dynamometry). In contrast, standard deviation increased from 5.19 +/- 1.98 to 9.22 +/- 3.56 beats per minute (P < .001).
It can be concluded from these data as well as from other plots presented in this article that RSA is a measure of the parasympathetic vagal tone, whereas standard deviation is increased by both sympathetic and parasympathetic arousal. Clinical evidence and data from physiological experiments are presented to show that a selective measure of vagal tone like RSA may offer advantages over standard deviation as a prognostic tool in cardiology.

Full text

Noordergraaf
M Moser, M Lehofer, A Sedminek, M Lux, HG Zapotoczky, T Kenner and A
problem from a theoretical point of view
Heart rate variability as a prognostic tool in cardiology. A contribution to the
ISSN: 1524-4539
Copyright © 1994 American Heart Association. All rights reserved. Print ISSN: 0009-7322. Online
72514
Circulation is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX
1994, 90:1078-1082Circulation
http://circ.ahajournals.org/content/90/2/1078
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
http://www.lww.com/reprints
Reprints: Information about reprints can be found online at
journalpermissions@lww.com
410-528-8550. E-mail: Fax:Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050.
Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters
http://circ.ahajournals.org//subscriptions/
Subscriptions: Information about subscribing to Circulation is online at
by guest on July 14, 2011http://circ.ahajournals.org/Downloaded from

1078
Special
Report
Heart
Rate
Variability
as
a
Prognostic
Tool
in
Cardiology
A
Contribution
to
the
Problem
From
a
Theoretical
Point
of
View
Maximilian
Moser,
PhD;
Michael
Lehofer,
MD,
PhD;
Andrea
Sedminek,
MD;
Manfred
Lux,
BA;
Hans-Georg
Zapotoczky,
MD;
Thomas
Kenner,
MD;
Abraham
Noordergraaf,
PhD
Background
Recent
clinical
studies
have
proposed
stan-
dard
deviation
of
heart
rate
as
a
diagnostic
tool
for
the
outcome
of
cardiac
infarction.
Mathematical
analysis
of
heart
rate
variability
shows
that
heart
rate
is
influenced
by
different
frequency
components
derived
from
different
parts
of
the
autonomous
nervous
system.
In
the
experimental
part
of
this
study,
we
investigated
the
possibility
of
calculating
a
variable
describing
the
parasympathetic
branch
of
the
autonomous
nervous
system
exclusively.
Methods
and
Results
In
60
healthy
volunteers,
heart
rate
was
measured
to
1
millisecond
during
two
different
conditions:
5
minutes
of
rest,
and
5
minutes
of
intermittent
handgrip
dynamometry;
the
latter
is
known
to
increase
sympathetic
arousal
selectively.
Heart
rate
was
found
to
be
lower
at
rest
(65.9±9.7
beats
per
minute)
than
during
dynamometry
(72.8±10.4
beats
per
minute,
P<.001).
Respiratory
sinus
arrhythmia
(RSA)
calculated
from
the
mean
absolute
differ-
ences
between
successive
heart
beats
showed
no
significant
change
(3.01+±
1.62
beats
per
minute
at
rest
versus
2.97±1.30
beats
per
minute
during
dynamometry).
In
contrast,
standard
deviation
increased
from
5.19±1.98
to
9.22±3.56
beats
per
minute
(P<.001).
Conclusions
It
can
be
concluded
from
these
data
as
well
as
from
other
plots
presented
in
this
article
that
RSA
is
a
measure
of
the
parasympathetic
vagal
tone,
whereas
standard
deviation
is
increased
by
both
sympathetic
and
parasympa-
thetic
arousal.
Clinical
evidence
and
data
from
physiological
experiments
are
presented
to
show
that
a
selective
measure
of
vagal
tone
like
RSA
may
offer
advantages
over
standard
deviation
as
a
prognostic
tool
in
cardiology.
(Circulation.
1994;90:1078-1082.)
Key
Words
*
myocardial
infarction
*
heart
rate
a
nervous
system,
autonomous
*
prognosis
*
Special
Reports
T
he
prognostic
value
of
heart
rate
variability
(HRV)
with
respect
to
survival
from
and
out-
come
of
myocardial
infarction
has
attracted
increasing
interest.
With
the
Holter
ECG,
a
large
amount
of
patient
data
can
be
collected
easily
and
noninvasively.
In
addition
to
the
shape
of
the
ECG,
irregularities
in
the
cardiac
rhythm
have
received
atten-
tion.
Because
HRV
mirrors
autonomic
equilibrium,
it
is
not
surprising
that
recent
studies
revealed
prognostic
possibilities
for
HRV
measurements
following
myocar-
dial
infarction.
Although
the
influence
of
the
autono-
mous
nervous
system
(ANS)
on
HRV
was
recognized
early
in
this
century,1
only
recently
have
several
stud-
ies2-5
made
it
clear
that
parasympathetic
and
sympa-
thetic
nervous
activities
influence
HRV
at
different
parts
of
the
frequency
spectrum.
At
first
glance,
standard
deviation
(SD),
widely
ac-
cepted
as
a
statistical
measure
of
dispersion,
would
appear
to
be
the
most
obvious
measure
of
HRV,
and
it
has
gained
popularity
for
this
purpose.
For
example,
Casolo
et
a16
proposed
a
relation
between
prognosis
subsequent
to
myocardial
infarction
and
SD
based
on
Received
December
6,
1993;
revision
accepted
March
30,
1994.
From
the
Physiological
Institute
(M.M.,
T.K)
and
the
Depart-
ment
of
Psychiatry
(M.
Lehofer,
A.S.,
M.
Lux,
H.-G.Z.),
Univer-
sity
of
Graz
(Austria);
and
the
Cardiovascular
Studies
Unit
(A.N.),
University
of
Pennsylvania
(Philadelphia).
Correspondence
to
Maximilian
Moser,
PhD,
Physiological
Insti-
tute,
Harrachgasse
21/5,
A-8010
Graz,
Austria.
©
1994
American
Heart
Association,
Inc.
24-hour
recordings
of
RR
intervals.
The
SD
of
heart
rate
appears
to
be
significantly
decreased
during
the
early
phase
of
myocardial
infarction,
and
this
observa-
tion
was
related
to
clinical
and
hemodynamic
indexes
of
severity.
Although
the
prognostic
value
seems
convinc-
ing
in
this
study,6
the
pathophysiological
basis
remains
unclear,
and
therefore
the
method's
reliability
is
questionable.
There
are
several
effects
that
contribute
to
total
HRV
and
therefore
to
SD.
Besides
patient
activity,
a
multitude
of
endogenous
physiological
rhythms
influence
heart
rate
during
a
24-hour
period.
Beginning
with
the
fastest,
the
following
main
periods
may
be
distinguished:
(1)
Respiratory
sinus
arrhythmia
(RSA)
is
mediated
by
respiration
and
is
strongly
controlled
by
parasympa-
thetic
activity.
The
short
response
time
typical
of
the
parasympathetic
system
is
responsible
for
this
relation.
This
is
supported
by
a
wide
range
of
studies
on
the
dose
response
to
atropine.24,7
Depending
on
respiratory
rate,
RSA
is
usually
observed
in
the
narrow
band
of
heart
rate
variations
ranging
from
2
to
5
seconds.
Sympathetic
nervous
system
activity
appears
to
be
too
slow
to
influence
this
frequency
band.5
In
a
clinical
study,8
RSA
and
therefore
parasympathetic
activity
were
found
to
be
significantly
reduced
in
patients
2
weeks
after
myocar-
dial
infarction.
(2)
"Medium"
waves
of
HRV
(period
length,
7
to
15
seconds):
Parasympathetic
as
well
as
sympathetic
influ-
ences
are
apparent
in
this
frequency
band,
which
in-
by guest on July 14, 2011http://circ.ahajournals.org/Downloaded from

Moser
et
al
Heart
Rate
Variability
1079
cludes
the
adrenergically
mediated
10-second
rhythm
of
blood
pressure
regulation.
Slow
respiration
contributes
RSA
components
in
this
band.
(3)
"Slow"
waves
of
HRV
(period
length,
>20
sec-
onds)
are
considered
to
be
sympathetic
in
origin.
Their
amplitude
is
primarily
inhibited
by
sympathetic
block-
ers.9
Augmentation
of
the
sympathetically
mediated
low-frequency
variations
of
heart
rate
was
found
in
cardiomyopathy.10
Slow
waves
are
also
augmented
by
handgrip
dynamometry,
which
is
known
to
selectively
increase
sympathetic
arousal."
In
hypertensive
patients,
the
contribution
of
slow
waves
compared
with
RSA
was
found
to
be
significantly
higher,
indicating
a
shift
toward
sympathetic
dominance.12
(4)
HRV
variations
with
a
period
in
the
minute
range
are
the
consequences
of
neurohumoral
oscillations
in
the
epinephrine-norepinephrine
and/or
angiotensin
lev-
els
in
the
circulating
blood.
Little
attention
has
been
devoted
to
these
oscillations
in
cardiac
disease.
(5)
Hour
rhythms
(3-,
4-,
or
6-hour
period
lengths)
can
be
considered
submultiples
of
the
circadian
rhythm
and
are
found
in
body
temperature.
They
can
be
drastically
increased
by
stresses
such
as
night
and/or
shift
work.13
Here,
the
circadian
system
reacts
by
mul-
tiplying
its
frequency,
like
a
flute
that
vibrates
with
its
overtone
if
blown
too
forcefully.
(6)
During
rapid
eye
movement
(REM)
stages
of
sleep,
additional
HRV
patterns
can
emerge
that
reflect
sympathetic
and
parasympathetic
arousal.14
These
pat-
terns
normally
follow
the
1.5-hour
rhythms
of
REM
sleep
periodicity.
(7)
Circadian
rhythms
appear
in
almost
every
Holter
ECG
and
reflect
alterations
in
ANS
equilibrium
throughout
the
day.
Studies
on
shift
workers
have
shown
that
the
circadian
amplitude
of
heart
rate
is
significantly
reduced
by
this
stress.
Heckmann
and
Busch'5
emphasized
that
the
circadian
variation
is
sig-
nificantly
reduced
immediately
after
myocardial
infarc-
tion
but
regenerates
over
a
period
of
weeks.
The
circadian
variation
may
also
be
influenced
by
antihyper-
tensive
agents
as
described
in
rats.16
Because
so
many
effects,
originating
from
different
neurohumoral
systems,
may
contribute
to
HRV,
the
diagnostic
and
prognostic
values
of
total
variability
measured
by
a
parameter
like
SD
may
be
questionable.
A
parameter
describing
just
one
branch
of
the
ANS
may
be
more
reliable.
Frequency
analysis
of
HRV
offers
differentiation
of
the
rhythms
described
above.
If
only
parasympathetic
arousal
is
of
interest,
RSA
would
ap-
pear
to
offer
the
best
parameter.
This
study
reports
a
comparison
between
heart
rate
SD,
a
quantity
used
in
cardiological
prognosis,
and
RSA,
a
quantity
known
to
reflect
cardiac
vagal
activity.
Both
quantities
can
be
calculated
easily
from
RR
inter-
val
data
obtained
with
a
Holter
ECG.
Methods
The
experiments
were
performed
on
60
healthy
supine
volunteers
in
three
parts.
In
part
1,
after
volunteers
rested
for
20
minutes
while
supine
in
a
quiet
room
under
constant
conditions,
heart
rate
during
5
minutes
of
rest
was
recorded
and
used
for
further
calculations.
In
part
2,
the
period
with
quiet
respiration
(part
1)
was
expiratory
breath
holding.
Respiratory
maneuvers
are
known
to
alter
the
activity
of
the
parasympathetic
branch
of
the
ANS.
This
part
of
the
experiments
is
used
only
for
Fig
1.
In
part
3,
5
minutes
of
intermittent
hand
dynamometry
followed.
The
dynamometry
consisted
of
5
seconds
of
maximal
and
1
minute
each
of
70%
and
40%
maximal
voluntary
handgrip
contraction
strength
interspaced
by
periods
of
rest.
Handgrip
dynamometry
is
known
to
activate
the
sympathetic
branch
of
the
ANS."1
Chest
wall
ECGs
were
recorded
and
stored
in
digital
form
during
the
experiment.
RR
intervals
were
determined
offline
to
1
millisecond.
A
computer
program
that
uses
matched
filtering
of
the
ECG
data
was
developed
to
recognize
the
R
peaks.
As
a
check,
all
QRS
complexes
were
plotted,
synchro-
nized
for
the
R
peak.
False
R
peaks
could
readily
be
recog-
nized
visually.
Experiments
with
>1%
false
R
peaks
were
excluded
from
further
processing.
Premature
beats
and
the
first
following
heartbeat
were
excluded
by
deleting
all
values
>20%
below
or
above
the
previous
intervals.
The
interbeat
intervals
were
converted
to
heart
rate.
Mean
and
SD
values
of
heart
rate
were
calculated
separately
for
each
part
of
the
experiment.
RSA
was
calcu-
lated
using
the
mean
absolute
difference
between
each
heart-
beat
interval
and
the
successive
one.17
This
procedure
is
even
simpler
than
the
calculation
of
SD.
RSA
was
calculated
separately
for
the
two
parts
of
the
experiment.
To
display
data
from
each
individual
experiment
(Fig
1),
successive
1-minute
periods
of
heart
rate
data
were
used
for
the
computation
of
SD
and
RSA.
After
each
calculation,
the
frame
was
moved
by
10
seconds,
and
another
minute
of
heart
rate
data
was
investigated.
This
process
gives
a
moving
frame
time
series
of
RSA
and
SD.
SD
of
all
experiments
was
plotted
against
RSA,
and
the
respective
regression
lines
were
computed
separately
for
the
two
experimental
conditions.
The
two
regression
lines
were
tested
for
parallelism
using
a
technique
described
by
Klein-
baum
and
Kupper.'8
All
other
regression
coefficients
were
tested
for
statistical
significance
by
a
one-sided
t
test.
Results
The
Table
shows
the
mean
values
of
heart
rate,
SD,
and
RSA
during
rest
and
during
dynamometry
for
all
subjects.
It
can
be
seen
that
heart
rate
increases
to-
gether
with
SD,
whereas
RSA
does
not
alter
signifi-
cantly
during
dynamometry.
A
plot
of
SD
of
heart
rate
versus
RSA
was
made
for
each
experiment. Fig
1
shows
one
experiment
in
which
the
effects
of
parasympathetic
and
sympathetic
activity
can
clearly
be
seen.
During
rest
(+),
parasympathetic
activity
is
at
a
medium
level.
During
inspiration
(I),
a
lower
level
of
parasympathetic
activity
was
achieved,
whereas
during
expiration
(E),
parasympathetic
activa-
tion
increased.
During
rest
as
well
as
during
respiratory
maneuvers,
expiration,
and
inspiration,
SD
displays
a
more
or
less
linear
relation
with
RSA.
Increasing
para-
sympathetic
activity
also
increases
RSA
and
SD
simultaneously.
During
the
dynamometric
part
of
the
experiment,
the
data
display
a
different
slope.
Handgrip
dynamometry
at
40%
strength
(4
in
Fig
1)
increases
SD
much
more
than
RSA.
This
effect
is
even
more
pronounced
at
70%
strength
(7
in
Fig
1),
where
RSA
does
not
increase
at
all.
In
Fig
2,
a
similar
diagram
is
shown
for
all
60
subjects.
Here,
RSA
and
SD
were
calculated
for
the
entire
5
minutes
of
rest
and
for
the
following
5
minutes
of
dynamometry
separately.
A
positive
nonlinear
correla-
followed
by
15
seconds
of
breath
holding
during
deep
inspira-
tion,
another
30
seconds
of
quiet
respiration,
and
15
seconds
of
tion
can
be
recognized
during
rest
(R=.78,
P<.001).
During
dynamometry,
SD
exhibits
a
distinct
increase
by guest on July 14, 2011http://circ.ahajournals.org/Downloaded from

1080
Circulation
Vol
90,
No
2
August
1994
10
5
11T-
4-
V)
0
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1
2
RSA,
beats/min
compared
with
RSA
and
strongly
increased
scatter
(R=.31,
P<.05).
Internal consistency
of
the
two
variables
is
shown
in
Fig
3
for
RSA
and
in
Fig
4
for
SD:
RSA
during
dynamometry
is
plotted
versus
RSA
at
rest.
It
can
be
seen
that
RSA
during
the
dynamometric
part
of
the
experiment
correlates
highly
with
RSA
in
the
same
volunteer
during
rest
(R=.88,
N=60,
P<.001,
Fig
3).
For
SD
(Fig
4),
the
correlation
between
dynamometry
and
rest
is
not
significant
(R=.15,
N=60,
P=NS).
Discussion
In
this
study,
the
prognostic
value
of
HRV
was
analyzed
from
a
physiological
point
of
view.
Parasym-
pathetic
and
sympathetic
tones
are
known
to
influence
the
total
HRV.
Although
empirical
evidence
shows
a
prognostic
value
of
total
HRV6
measured
by
SD,
a
differentiated
view
of
the
influences
coming
from
the
Heart
Rate,
Respiratory
Sinus
Arrhythmia,
and
Intraindividual
Standard
Deviation
of
Heart
Rate
Parameter
Rest
Dynamometry
P
HR,
bpm
65.9+9.73
72.8+10.44
<.001
RSA,
bpm
3.01+1.62
2.97+1.30
NS
SD,
bpm
5.19+1.98
9.22+3.56
<.001
N
60
60
HR
indicates
heart
rate;
bpm,
beats
per
minute;
RSA,
respira-
tory
sinus
arrhythmia;
and
SD,
standard
deviation.
FIG
1.
Standard
deviation
(SD)
of
heart
rate
(HR)
versus
respiratory
sinus
arrhythmia
(RSA)
during
two
different
experimental
conditions
recorded
in
one
subject:
5
min-
utes
of
rest:
the
patient
rests
quietly
during
the
recording
of
heart
rate
(+)
or
performs
deep
inspiration
(I)
or
expiration
(E);
and
5
minutes
of
a
dynamometric
test:
the
subject
presses
a
hand
dynamometer
at
70%
(7)
and
40%
(4)
of
maximal
strength.
ANS
holds
the
promise
of
a
more
precise
prognostic
tool.
It
also
offers
possible
understanding
of
the
patho-
physiological
background.
It
therefore
was
our
aim
to
25
SD
of
heart
rate
beats/min
20
15
10
0
0
0~~~~~~~
.
*
*
o°
000
000
-
~0
0
0c
0O
5
0
.
0o
RSA,
beats/min
0
0
1
2
3
4
5
6
7
8
FIG
2.
Standard
deviation
(SD)
of
heart
rate
versus
respiratory
sinus
arrhythmia
(RSA)
during
two
different
experimental
condi-
tions:
5
minutes
of
rest
(c):
the subjects
rest
quietly
during
the
recording
of
heart
rate;
and
5
minutes
of
a
dynamometric
test
(A):
the
patients
press
a
hand
dynamometer
at
70%
of
maximal
voluntary
contraction
force
(MVC)
for
1
minute
and
at
40%
of
MVC
for
another
minute.
Heart
rate
is
recorded
during
contrac-
tion
and
during
the
adjacent
resting
phases
of
1
minute
each.
For
rest,
SD=2.74
*
RSA05"',
R=.78,
N=60
(P<.001).
For
dyna-
mometry,
SD=6.46-
RSAo2aa,
R=.31,
N=60
(P<.001).
4-
7./
4
1
/
1<
by guest on July 14, 2011http://circ.ahajournals.org/Downloaded from

Moser
et
al
Heart
Rate
Variability
1081
8
,,,,...swl
U'''''''''
7
RSA
line
of
identity
6
-(beats/min)
/ /
5
4
X
regresion6_H
39
2
RSA
dyn
=
0,842
+
0,707
x
RSA
t
R=
0,882
0
..
..
,
I , ,
,
,
...
0
1
2
3
4
5
6
7
8
RSA
rest
(beats/mmn)
FIG
3.
Plot
of
respiratory
sinus
arrhythmia
(RSA)
during
dyna-
mometry
versus
RSA
during
rest
in
60
subjects.
Note
that
the
relation
is
nearly
independent
of
the
experimental
condition
and
especially
characteristic
for
a
person
at
low
RSA
values.
take
a
look
at
the
two
branches
of
the
ANS
using
SD
and
RSA.
Healthy
subjects
were
submitted
to
two
simple
exper-
imental
conditions:
rest
was
used
to
emphasize
para-
sympathetic
tone,
whereas
sympathetic
tone
was
in-
creased
by
hand
dynamometry.
The
significant
increase
of
SD
and
heart
rate
in
the
Table
indicates
enhanced
sympathetic
drive
during
hand
dynamometry,
accompa-
nied
by
no
significant
change
in
RSA.
It
is
well
established
that
RSA
is
a
measure
of
parasympathetic
(vagal)
tone.2,57
It
is
less
well
known
which
influences
constitute
the
SD
of
heart
rate.
There-
fore,
the
correlation
between
SD
and
RSA
was
calcu-
lated
in
Fig
2;
a
significant
positive
nonlinear
correlation
between
SD
and
RSA
was
found
during
rest.
The
25
.
*
1
'
''A
SD
dyn
line
of
Identity
20
(beatsimin)
*
0
10
,
i.
5
*/
regression
line
/SD
dyn
=
7,8
+
0,274*SD
test
R=
0,152
0
5
10
15
20
25
SD
rest
(beats/min)
FIG
4.
Plot
of
standard
deviation
(SD)
during
dynamometry
versus
SD
during
rest
in
60
subjects.
Note
that
the
quantity
is
highly
dependent
on
the
experimental
condition
and
has
a low
self-correlation.
scatter
present
in
the
correlation
indicates
that
even
during
rest,
SD
is
not
a
pure
measure
of
parasympa-
thetic
activity.
Sympathetic
influences
therefore
are
considered
to
be
responsible
for
the
variability
in
SD.
Accordingly,
during
dynamometry
the
regression
coef-
ficient
is
significantly
reduced
compared
with
rest
(P<.001)
due
to
the
increased
sympathetic
influences.
As
shown
in
the
Table,
it
turned
out
that
RSA
was
insensitive
and
SD
was
highly
susceptible
to
sympathetic
effects.
Because
both
sympathetic
and
parasympathetic
effects
influence
HRV,
the
SD
of
heart
rate
is
unable
to
distinguish
between
the
two
effects.
The
two
branches
of
the
ANS
exert
a
well-established
antagonistic
influence
on
the
threshold
of
ventricular
fibrillation.'9
Although
activation
of
the
sympathetic
ANS
at
every
level
leads
to
significant
lowering
of the
threshold
for
ventricular
fibrillation,
parasympathetic
activity
is
known
to
increase
this
threshold
with
preex-
isting
sympathetic
arousal.20
Ventricular
fibrillation
was
nearly
absent
even
with
ligation
of
the
left
anterior
descending
coronary
artery
in
the
absence
of
sympa-
thetic
activity
in
dogs.21
It
is
also
well
known
that
psychological
stress
accompanied
by
a
high
level
of
circulating
adrenergic
hormones
predisposes
to
sudden
cardiac
death.
It
has
been
shown
that
conditions
increasing
the
parasympathetic
tone,
such
as
meditation,
are
able
to
reduce
premature
ventricular
beat
frequency
in
cardiac
patients.19
The
beneficial
effect
of
vagal
stimulation
was
also
found
during
experimental
acute
myocardial
infarc-
tion
in
which
71%
of
dogs
with
vagal
stimulation
survived
after
30
minutes
of
coronary
artery
occlusion
compared
with
10%
of
the
control
animals.22
Therefore,
a
detrimental
effect
of
sympathetic
arousal
on
ischemic
heart
muscle
can
be
expected,
whereas
a
cardioprotective
effect
is
ascribed
to
para-
sympathetic
arousal.
Under
these
circumstances,
a
strict
separation
of
parasympathetic
and
sympathetic
influ-
ences
is
considered
a
prerequisite
for
a
prognostic
tool.
To
check
the
stability
of
the
parameters,
the
internal
consistency
was
investigated
for
the
RSA
(Fig
3)
and
the
SD
(Fig
4).
RSA
proved
to
be
the
more
stable
param-
eter,
being
less
susceptible
to
disturbing
effects.
SD
(Fig
4)
revealed
less
stability,
showing
no
significant
corre-
lation
between
dynamometry
and
rest.
SD
of
heart
rate
alone
should
be
used
cautiously
for
prognostic
purposes;
parasympathetic
as
well
as
sympa-
thetic
activity
will
influence
SD,
and
the
connection
between
high
values
of
SD
and
good
prognosis
de-
scribed
recently6'23
can
be
expected
only
if
no
disturbing
sympathetic
effect
interferes.
SD
is
a
measure
of
the
deviation
of
a
parameter
around
its
mean
value.
If
the
mean
value
is
the
24-hour
heart
rate,
then
its
SD
is
influenced
by
all
variations
occurring
in
the
range
of
seconds
to
24
hours.
Therefore,
a
pulse
frequency
of
70
during
the
day
and
60
during
the
night
without
any
other
variation
could
produce
the
same
SD
as
a
strong
RSA
during
the
day,
although
the
patients'
autonomic
states
could
be
completely
different
in
the
two
cases.
The
amount
of
circadian
variation
may
be
the
main
factor
influencing
the
24-hour
SD
of heart
rate.
HRV
is
a
quantity
that
can
be
obtained
with
any
Holter
ECG.
Its
obvious
clinical
importance
justifies
a
careful
mathematical
analysis
separating
the
influences
of
the
different
parts
of
the
ANS.
RSA
calculated
from
by guest on July 14, 2011http://circ.ahajournals.org/Downloaded from

1082
Circulation
Vol
90,
No
2
August
1994
mean
absolute
interval
differences
is
easy
to
calculate
and
could
complement
SD
of
heart
rate.
If
both
SD
and
RSA
are
evaluated
dynamically
from
the
same
data,
the
sympathetic
as
well
as
the
parasympathetic
part
of
the
HRV
can
be
displayed
(Fig
1).
These
two
simple
measures
of
variability
offer
insights
into
the
tonic
state
of
the
ANS
during
rest
as
well
as
into
the
dynamic
variations
during
exercise.
Additional
clinical
data
will
have
to
prove
the
value
of
RSA
as
a
prognostic
tool
in
cardiology.
Acknowledgments
We
would
like
to
thank
Prof
Gunther
Hildebrandt,
Univer-
sity
of
Marburg
(Germany),
UD
Dr
Karl-Peter
Pfeiffer,
Uni-
versity
of
Graz,
and
Prof
Max
Pichler,
Grossgmain,
Austria,
for
valuable
suggestions
and
Mag
Magdalena
Voica
for
help
during
the
measurements.
This
study
was
supported
in
part
by
the
Austrian
Ministry
of
Science
and
Research
(Project
Puls-
trans)
and
by
the
Austrian
Academy
of
Sciences.
References
1.
Wiersma
ED.
Der
Einflul
von
Bewui3tseinszustinden
auf
Puls
und
die
Atmung.
Z
Ges
Neurol
Psych.
1913;1-24.
2.
Fouad
FM,
Tarazi
RC,
Ferrario
CM,
Fighaly
S,
Alicandri
C.
Assessment
of
parasympathetic
control
of
heart
rate
by
a
nonin-
vasive
method.
Am
J
Physioi.
1984;246:H838-H842.
3.
Hayano
J,
Sakakibara
Y,
Yamada
A,
Yamada
M,
Mukai
S,
Fujinami
T,
Yokoyama
K,
Watanabe
Y,
Takata
K.
Accuracy
of
assessment
of
cardiac
vagal
tone
by
heart
rate
variability
in
normal
subjects.
Am
J
Cardiol.
1991;67:199-204.
4.
Billman
GE,
Dujardin
JP.
Dynamic
changes
in
cardiac
vagal
tone
as
measured
by
time-series
analysis.
Am
J
Physiol.
1990;258:
H896-H902.
5.
Akselrod
S,
Gordon
D,
Madwed
JB,
Snidman
NC,
Shannon
DC,
Cohen
RJ.
Hemodynamic
regulation:
investigation
by
spectral
analysis.
Am
J
Physiol.
1985;249:H867-H875.
6.
Casolo
GC,
Stroder
P,
Signorini
C,
Calzolari
F,
Zucchini
M,
Balli
E,
Sulla
A,
Lazzerini
S.
Heart
rate
variability
during
the
acute
phase
of
myocardial
infarction.
Circulation.
1992;85:2073-2079.
7.
Hayano
J,
Sakakibara
Y,
Yamada
A,
Yamada
M,
Mukai
S,
Fujinami
T,
Yokoyama
K,
Watanabe
Y,
Takata
K.
Accuracy
of
assessment
of
cardiac
vagal
tone
by
heart
rate
variability
in
normal
subjects.
Am
J
Cardiol.
1991;67:199-204.
8.
Lombardi
F,
Sandrone
G,
Pernpruner
S,
Sala
R,
Garimoldi
M,
Cerutti
S,
Baselli
G,
Pagani
M,
Malliani
A.
Heart
rate
variability
as
an
index
of
sympathovagal
interaction
after
acute
myocardial
infarction.
Am
J
Cardiol.
1987;60:1239-1245.
9.
Pagani
M,
Lombardi
F,
Guzzetti
S,
Rimoldi
0,
Furlan
R,
Pizzinelli
P,
Sandrone
G,
Malfatto
G,
Dell'Orto
S,
Piccaluga
E,
Turiel
M,
Baselli
G,
Cerrutti
S,
Malliani
A.
Power
spectral
analysis
of
heart
rate
and
arterial
pressure
variabilities
as
a
marker
of
sym-
pathovagal
interaction
in
man
and
conscious
dogs.
Circ
Res.
1986;
59:176-193.
10.
Binkley
PF,
Nunziata
E,
Haas
GJ,
Nelson
SD,
Cody
RJ.
Parasym-
pathetic
withdrawal
is
an
integral
component
of
autonomic
imbalance
in
congestive
heart
failure:
demonstration
in
human
subjects
and
verification
in
a
paced
canine
model
of
ventricular
failure.
J
Am
Coll
Cardiol.
1991;18:464-472.
11.
Smolander
J,
Louhevaara
V.
Effect
of
heat
stress
on
muscle
blood
flow
during
dynamic
handgrip
exercise.
J
Appl
Physiol.
1992;65:
215
-220.
12.
Guzzetti
S,
Piccaluga
E,
Casati
R,
Cerutti
S,
Lombardi
F,
Pagani
M,
Malliani
A.
Sympathetic
predominance
in
essential
hyper-
tension:
a
study
employing
spectral
analysis
of
heart
rate
vari-
ability.
J
Hypertens.
1988;6:711-771.
13.
Hildebrandt
G.
Chronobiologische
Grundlagen
der
Leistungsfa-
higkeit
und
Chronohygiene.
In:
Hildebrandt
G,
ed.
Biologische
Rhythmen
und
Arbeit.
Wien/New
York:
Springer;
1976:1-19.
14.
Raschke
F.
The
respiratory
system-features
of
modulation
and
coordination.
In:
Haken
H,
Koepchen
HP,
eds.
Rhythms
in
Physi-
ological
Systems.
Berlin:
Springer;
1991:155-164.
15.
Heckmann
C,
Busch
M.
Changes
in
circadian
pattern
of
heart
rate
in
patients
after
acute
myocardial
infarction.
In:
Hildebrandt
G,
Moog
R,
Raschke
F,
eds.
Chronobiology
and
Chronomedicine.
Frankfurt/Main:
Springer;
1987:299-306.
16.
Janssen
BJA,
Tyssen
CM,
Struyker
BHAJ.
Modification
of
cir-
cadian
blood
pressure
and
heart
rate
variability
by
five
different
antihypertensive
agents
in
spontaneously
hypertensive
rats.
J
Cardiovasc
Pharmacol.
1991;17:494-503.
17.
Eckoldt
K.
Probleme
und
Ergebnisse
der
Analyse
des
Sinus-
rhythmus.
In:
Zwiener
U,
Michalik
M,
Eckoldt
K,
Klossek
H,
eds.
Herzfrequenzvariabilitat-MMoglichkeiten
zur
Diagnostik
neurolo-
gischer
Erkrankungen.
Leipzig:
Hirzel;
1990:53
-
63.
18.
Kleinbaum
DG,
Kupper
LL.
Applied
Regression
Analysis
and
Other
Multivariable
Methods.
Duxbury,
Mass:
Duxbury
Press;
1978.
19.
Lown
B,
Desilva
RA,
Reich
P,
Murawski
BJ.
Psychophysiologic
factors
in
sudden
cardiac
death.
Am
J
Psychiatr.
1980;137:
1325-1335.
20.
Lown
B,
Verrier
RL.
Neural
activity
and
ventricular
fibrillation.
N
Engl
J
Med.
1976;294:1165-1170.
21.
Satinsky
J,
Kosowsky
B,
Lown
B.
Ventricular
fibrillation
induced
by
hypopthalamic
stimulation
during
coronary
occlusion.
Circulation.
1971;93(suppl
II):II-60.
Abstract.
22.
Myers
RW,
Pearlman
AS,
Hyman
RM,
Goldstein
RA,
Kent
KM,
Goldstein
RE,
Epstein
SE.
Beneficial
effects
of
vagal
stimulation
and
bradycardia
during
experimental
acute
myocardial
ischemia.
Circulation.
1974;49:943-947.
23.
Kleiger
RE,
Miller
JP,
Bigger
JT,
Moss
AJ,
for
the
Multicenter
Post-Infarction
Research
Group.
Decreased
heart
rate
variability
and
its
association
with
increased
mortality
after
acute
myocardial
infarction.
Am
J
Cardiol.
1987:57:256-262.
by guest on July 14, 2011http://circ.ahajournals.org/Downloaded from