Journal
of
Physiology
(1994),
480.2
Cardiac
contraction
and
intramyocardial
venous
pressure
generation
in
the
anaesthetized
dog
Isabelle
Vergroesen,
Yves
Han,
Masami
Goto
and
Jos
A.
E.
Spaan
Department
of
Medical
Physics
and
Medical
Informatics,
Faculty
of
Medicine,
University
of
Amsterdam,
Amsterdam,
The
Netherlands
1.
Two
hypotheses
relating
to
the
influence
of
contraction
of
the
heart
on
coronary
venous
pressure
(Pv)
were
tested.
The
first
assumes
a
direct
transmission
of
left
ventricular
pressure
(PLv).
According
to
the
alternative
hypothesis
the
Pv
is
caused
by
cyclical
changes
in
the
elastance
of
the
surrounding
tissue.
2.
A
small
epicardial
vein
was
cannulated
retrogradely
in
eight
open-chest
dogs
deeply
anaesthetized
with
fentanyl.
The
duration
of
diastoles
was
varied
after
induction
of
a
heart
block
with
formaldehyde.
Coronary
arterial
inflow
and
perfusion
pressure
were
controlled
by
a
perfusion
system
connected
to
the
left
main
coronary
artery
by
a
Gregg
cannula.
Stopped-flow
Pv
was
studied
with
intrinsic
coronary
tone
(IT)
and
after
maximal
dilatation
with
adenosine.
3.
The
Pv
pulse
in
the
first
contraction
after
a
long
diastole
was
not
significantly
correlated
to
the
PLV
pulse,
with
a
slope
of
0
5,
in
any
dog,
either
with
IT
or
during
adenosine
treatment.
Comparing
the
first
contraction
after
the
long
diastole
with
the
last
beat
before,
systolic
Pv
pulse
decreased
significantly
in
seven
out
of
eight
dogs,
but
systolic
PLV
pulse
increased
in
five
dogs
and
was
unaltered
in
three
dogs
in
both
conditions.
In
contrast,
end-diastolic
Pv
was
significantly
correlated
to
the
systolic
Pv
in
each
individual
animal
under
either
condition.
4.
The
results
indicate
that
pressure
generation
in
the
small
coronary
veins
can
be
explained
on
the
basis
of
the
time-varying
elastance
hypothesis
and
that
a
direct
transmission
of
PLV
to
Pv
is
absent.
Progress
in
understanding
the
mechanics
of
the
coronary
circulation
has
been
hampered
by
a
lack
of
insight
into
the
events
that
take
place
in
the
intramyocardial
small
vessels.
Recent
studies
indicate
that
intramyocardial
vascular
blood
volumes
vary
during
the
heart
cycle
in
a
fashion
that
is
related
to
the
resistance
of
the
microvessels
(Hoffman
&
Spaan,
1990).
Moreover,
volume
variations
are
related
to
the
capacitive
flow
components
that
contribute
to
the
phasic
behaviour
of
coronary
arterial
and
venous
flow
(Spaan,
Bruinsma
&
Laird,
1983).
In
order
to
gain
a
better
understanding
of
the
factors
that
control
the
mechanism
of
blood
flow
into
and
out
of
the
intramyocardial
vessels
it
is
important
that
we
identify
the
forces
that
are
exerted
by
contraction
of
the
heart
on
the
vessels
and
how
these
are
modified
by
vascular
tone.
In
the
past
it
was
widely
accepted
that
contraction
forces
were
related
to
tissue
pressure,
which
in
turn
was
directly
determined
by
the
pressures
generated
within
the
left
ventricular
cavity
(Downey
&
Kirk,
1975).
This
hypothesis
assumes
a
direct
transmission
of
left
ventricular
pressure,
one
to
one
at
the
subendocardium
and
decreasing
to
zero
at
the
subepicardium.
If
this
hypothesis
is
correct
then
in
an
epicardial
vein,
cannulated
retrogradely,
the
pressure
pulse
should
essentially
equal
0
5
times
the
left
ventricular
pressure
pulse.
Contradicting
this
hypothesis,
Krams,
Sipkema
&
Westerhof
(1989)
showed
that
in
the
isolated
cat
heart
changes
in
the
elastance
of
the
myocardium
between
systole
and
diastole
could
be
a
major
factor
determining
systolic
to
diastolic
flow
differences
in
the
coronary
arteries.
According
to
this
alternative
hypothesis
the
magnitude
of
the
pressure
pulse
in
any
compartment
inside
the
heart
muscle
would
then
be
determined
by
the
volume
of
this
compartment
and
by
the
maximal
systolic
elastance
of
the
heart
muscle.
Krams
et
al.
(1989)
found
no
direct
transmission
of
the
left
ventricular
pressure
pulse
on
the
flow
pulsations.
However,
the
coronary
vessels
may
not
be
simply
affected
by
either
elastance
or
left
ventricular
pressure.
Kouwenhoven,
Vergroesen,
Han
&
Spaan
(1992)
showed
in
a
recent
study
that
left
ventricular
pressure
can
affect
considerably
coronary
arterial
pressure
and
early
systolic
flow,
depending
on
the
mode
of
perfusion
of
the
coronary
circulation.
The
two
modes
of
perfusion
can
be
denoted
as
pressure
clamping
or
constant
M$S
1860,
pp.
343-353
343
I.
Vergroesen,
Y.
Han,
M.
Goto
and
J
A.
E.
Spaan
pressure
perfusion
(pressure
during
the
heart
beat
is
constant)
and
as
flow
clamping
or
constant
flow
perfusion
(flow
during
the
heart
beat
is
constant).
During
constant
flow
perfusion,
coronary
arterial
pressure
is
affected
by
left
ventricular
pressure
during
the
whole
of
systole,
and
during
constant
pressure
perfusion
only
early
systolic
flow
is
influenced
by
the
left
ventricular
pressure.
Moreover,
coronary
venous
flow
is
pulsatile,
being
high
in
systole
and
low
in
diastole
(Spaan
et
al.
1983).
Hence,
myocardial
contraction
forces
venous
blood
to
leave
the
myocardial
wall.
However,
the
influence
of
contraction
on
the
intra-
myocardial
veins
has
been
much
less
studied
than
the
influence
on
arterial
flow.
The
purpose
of
the
present
study
was
to
distinguish
between
the
above-mentioned
two
mechanisms
in
their
effects
on
stopped-flow
venous
pressure.
A
measure
of
the
latter
was
obtained
by
retrograde
cannulation
of
a
small
epicardial
vein.
Induction
of
diastoles
of
variable
duration
allowed
intramyocardial
veins
to
be
filled
to
a
variable
degree
(Vergroesen,
Noble
&
Spaan,
1987).
During
the
following
contraction,
the
aortic
valves
opened
at
the
existing
end-diastolic
aortic
pressure,
leading,
in
general,
to
a
lower
left
ventricular
pressure
relative
to
the
last
contraction
before
the
longer
diastole.
The
venous
pressure
generation
during
the
first
contraction
after
a
long
diastole
was
used
to
distinguish
between
the
two
mechanisms
that
may
generate
intramyocardial
venous
pressures.
According
to
the
direct
transmission
of
left
ventricular
pressure
hypothesis,
venous
pressure
pulses
after
a
longer
diastole
should
be
decreased
to
an
extent
that
is
related
to
the
decrease
in
the
left
ventricular
pressure
pulse
itself
by
a
factor
of
0
5.
However,
according
to
the
time-varying
elastance
hypothesis,
the
systolic
venous
pressure
should
be
increased
after
a
long
diastole
due
to
an
increased
filling
of
the
venous
compartment.
METHODS
Animal
preparation
Eight
mongrel
dogs
weighing
24-34
kg
were
anaesthetized
by
intramuscular
injection
of
a
mixture
of
4
ml
ketamine
hydrochloride
(100
mg
ml-'
Aescoket;
Aesculaep
BV,
Boxtel,
The
Netherlands),
4
ml
Rompun
(20
mg
ml-';
Bayer,
Leverkusen,
FRG)
and
2
ml
atropine
sulphate
(0
5
mg
ml-').
Anaesthesia
was
maintained
by
intravenous
injection
of
50
ml
fentanyl
(0
05
mg
ml-')
and
2
ml
pancuronium
bromide
(2
mg
ml-1
Pavulon;
Organon,
Boxtel,
The
Netherlands),
a
paralytic
agent,
used
in
order
to
enable
ventilation
of
the
animal.
This
high-dose
fentanyl
anaesthesia
is
routinely
used
in
human
open-heart
surgery,
since
it
has
very
little
cardiac
effect.
After
intubation
the
dogs
were
ventilated
using
a
Harvard
respirator
with
a
2:1
nitrous
oxide-oxygen
mixture.
Arterial
blood
gases
were
measured
frequently,
and
arterial
oxygen
tension
was
maintained
in
a
physiological
range,
by
adjustment
of
either
ventilation
volume
or
frequency.
Every
4
h
an
additional
dose
of
5
ml
fentanyl
(0
05
mg
ml-')
was
given.
A
left
thoracotomy
was
performed
in
the
fourth
or
fifth
intercostal
space.
The
subelavian
artery
was
dissected
free
and
two
ligatures
were
placed
around
it.
The
pericardium
was
opened
and
the
heart
was
suspended
in
a
pericardial
cradle.
A
heart
block
was
induced
by
injection
of
formaldehyde
into
the
bundle
of
His,
close
to
the
arterio-ventricular
(AV)
node
(Steiner
&
Kovalik,
1968).
The
heart
was
paced
on
the
right
ventricular
free
wall
at
a
rate
of
90
beats
min-.
The
left
main
coronary
artery
was
dissected
and
a
ligature
placed
around
it.
A
small
epicardial
vein
was
selected
and
two
7-0
ligatures
were
placed
around
it
approximately
1
cm
apart
from
each
other.
A
pre-heparinized
cannula
(PE-90,
i.d.
0-58
mm,
o.d.
0-96
mm,
length
approximately
5
cm)
was
inserted
retrogradely
into
the
small
epicardial
vein
after
the
downstream
ligature
was
tied.
This
cannula
was
advanced
far
enough
to
be
fixed
into
place
by
the
upstream
ligature
and by
a
drop
of
Histoacryl®
(Braun,
Melsungen,
FRG),
a
tissue
adhesive
(Fig.
1).
A
successful
cannulation
resulted
in
venous
blood
flowing
through
the
cannula
predominantly
during
systole.
A
purse-string
suture
in
the
left
atrial
appendage
was
made
for
introduction
of
a
catheter-tip
manometer.
In
five
dogs
two
ultrasonic
crystals
were
sewn
to
opposite
sides
of
the
left
ventricle.
After
intravenous
administration
of
750
i.u.
kg-'
of
heparin,
the
left
carotid
artery
was
cannulated.
A
stainless
steel
cannula
(Gregg
cannula,
i.d.
3
mm,
o.d.
4
mm)
was
inserted
into
the
aorta
via
the
left
subclavian
artery.
Blood
from
the
cannulated
left
carotid
artery
was
circulated
for
10
min
through
the
perfusion
system
(Spaan
et
at.
1983)
and
returned
to
the
aorta
through
the
Gregg
cannula,
before
the
cannula
was
inserted
into
the
left
main
coronary
artery
and
secured
in
place
with
the
ligature
while
maintaining
a
constant
perfusion
pressure.
A
catheter-tip
manometer
was
inserted
through
the
purse-string
suture
in
the
left
atrial
appendage
into
the
left
ventricle
for
measurement
of
the
left
ventricular
pressure,
PLV.
Anticoagulation
was
maintained
by
continuous
infusion
of
heparin
(5000
i.u.
h-1).
After
the
experiment
a
mixture
of
gelatine
and
white
latex
paint,
with
a
particle
size
of
50
,sm,
was
injected
into
the
artery
to
outline
the
perfused
tissue.
After
cooling
in
ice
for
15
min
the
perfused
tissue
was
dissected
and
weighed.
Measurements
The
PLV
was
measured
using
a
catheter-tip
pressure
transducer
(5F,
Millar
SPC-350;
Millar
Instruments
Inc.,
Houston,
TX,
USA).
Coronary
perfusion
pressure
(Pp)
was
measured
at
the
tip
of
the
cannula
by
a
small
microtip
pressure
transducer
(5F;
Philips,
Best,
The
Netherlands).
Stop-flow
coronary
venous
pressure
(Pv)
was
measured
using
a
catheter-tip
pressure
transducer
with
a
lumen
(6F,
Philips)
connected
to
the
venous
cannula
by
a
special
connector
(Fig.
1).
Coronary
arterial
blood
flow
(Qa)
was
measured
by
an
electromagnetic
cannulating
flow
probe
(Statham
5
mm,
model
Sp2202;
Oxnard,
CA,
USA)
interposed
into
the
perfusion
tube.
In
five
dogs,
left
ventricular
dimensions
were
measured
with
ultrasonic
crystals,
diameter
4
mm,
sewn
to
either
side
of
the
ventricle.
The
measurement
cycle
was
dictated
by
a
1
kHz
clock.
After
a
clock
pulse
one
crystal
sent
a
burst
of
500
kHz,
this
being
the
resonance
frequency
of
both
crystals.
Time
measurement
started
when
the
burst
was
sent
by
the
sending
crystal,
and
stopped
when
the
receiving
crystal
received
a
burst
of
the
right
frequency
and
amplitude
above
a
pre-set
threshold.
The
digital-analog
conversion
used
an
344
J.
Physiol.
480.2
Cardiac
contraction
and
coronary
venous
pressure
integrator
and
a
sample-and-hold
circuit.
A
low-pass
filter
(-3
dB
=
200
Hz)
suppressed the
signal
of
the
1
kHz
clock
at
the
outlet.
Calibration
of
the
distance
measurement
was
done
by
putting
the
crystals
in
a
water-filled
container.
The
calibration
was
corrected
for
the
difference
in
sound
speed
in
water
and
tissue.
The
area
was
calculated
by
squaring
the
numbers
after
calibration
in
centimetres
and
area
changes
were
used
as
indication
of
volume
changes
in
the
left
ventricle.
The
ultrasonic
crystals
were
very
precise
(041
mm)
as
long
as
they
correctly
picked
up
the
signal
of
the
partner
crystal.
When
they
did
not
pick
up
this
signal,
no
distance
signal
was
generated,
and
no
pressure-area
loops
could
be
calculated.
All
signals
were
A/D
converted
on-line
at
a
sampling
rate
of
80
Hz
with
an
Olivetti
M24
PC
equipped
with
an
Analog
Devices
RTI-800
A/D
board
(Norwood,
MA,
USA),
using
software
which
was
developed
in
our
laboratory.
Analog
signals
were
filtered
at
32
Hz
(12
dB
octave-')
prior
to
A/D
conversion.
Digitized
data
were
stored
on
hard
disk
for
off-
line
data
analysis.
All
signals
were
recorded
on
a
Hewlett-
Packard
FM
instrumentation
recorder
(HP
3968A)
for
analog
backup.
Interventions
Reactive
hyperaemia
(RH)
was
induced
by
an
occlusion
of
the
perfusion
tube
for
15
s
to
check
the
presence
of
vascular
tone
and
this
was
routinely
followed
by
a
3
min
interval
without
intervention
in
order
to
resume
normal
tone.
Long
diastoles
(LD)
were
induced
by
switching
off
the
pacer
for
durations
between
1
and
14
s
in
random
order,
each
followed
by
at
least
twenty
normal
beats.
Protocol
Perfusion
pressure
(Pp)
was
set
at
100
mmHg
and
an
RH
was
recorded.
At
random,
Pp
was
set
at
a
high
(120
mmHg)
or
a
low
level
(80
mmHg).
At
least
eight
LDs
of
variable
length
and
random
order
were
induced
at
the
chosen
Pp.
At
this
Pp
a
RH
was
recorded.
Subsequently
the
Pp
was
set
at
the
other
level
(see
above)
and
at
least
eight
LDs
of
variable
length
were
induced
at
this
Pp
level
followed
by
a
RH.
The
adenosine
infusion
was
then
started
and
controlled
by
a
15
s
occlusion.
In
the
vasodilated
coronary
bed
Pp
was
randomly
set
at
high
(60
mmHg)
or
low
level
(40
mmHg),
followed
by
aproximately
eight
LDs
of
variable
length.
The
Pp
was
then
set
at
the
other
level
and
another
eight
LDs
were
induced.
The
protocol
was
extended
to
the
one
described,
after
work
with
the
first
three
dogs
(A,
B
and
C)
in
which
only
one
level
of
Pp,
was
used,
in
order
to
induce
more
variation
in
diastolic
venous
pressure.
Data
analysis
The
data
for
each
dog
were
grouped
according
to
the
coronary
tone,
i.e.
intrinsic
coronary
tone
and
vasodilatation.
One
dog
(dog
D)
had
a
peak
reactive
hyperaemic
flow
less
than
twice
the
control
level
and
from
this
dog
only
measurements
after
full
vasodilatation
were
included
in
the
analysis.
The
other
dogs
had
peak
reactive
hyperaemic
flows
of
at
least
4
times
the
control
level.
In
one
dog
(dog
C)
no
measurements
were
performed
during
adenosine
perfusion.
A
beat
was
defined
as
starting
at
the
beginning
of
diastole
and
including
the
following
systole.
Thus
a
'long
diastolic
beat'
starts
with
a
prolonged
diastole
and
includes
the
following
systole.
In
Fig.
2,
venous
pressure
during
a
typical
prolonged
diastole
is
shown.
In
the
beat
before
the
long
diastole
two
time
points
are
marked
(1,
end
diastole;
2,
peak
systole)
and
in
the
long
diastolic
beat
four
time
points
are
marked
(3,
onset
of
diastole;
4,
onset
of
the
plateau
in
P,;
5,
end
diastole;
and
6,
peak
systole).
The
onset
of
the
plateau
in
P,
was
determined
only
in
the
long
diastoles
that
lasted
long
enough
to
reach
a
plateau.
Statistical
tests
The
influence
of
left
ventricular
pressure
on
venous
pressure
generation
was
tested
in
two
ways:
(a)
by
correlating
within
each
dog
the
pulse
pressure
in
the
left
ventricular
cavity
to
coronary
venous
pulse
pressure
for
all
long
diastolic
beats
and
(b)
by
correlating
within
each
dog
the
change
in
LV
pulse
Tip
manometer
Upstream
tie
Downstream
tie
Coronary
sinus
Figure
1.
Schematic
drawing
of
the
cannulation
of
a
small
epicardial
vein
and
interpretation
of
stopped-flow
venous
pressure
measurement
Node
1
is
the
first
interconnection
upstream
in
the
venous
network
relative
to
the
position
of
the
occluded
vein.
Alternative
venous
outflow
routes
drain
into
the
coronary
sinus.
R
represents
the
resistance
of
each
part
of
the
venous
network
in
between
node
points.
Since
during
pressure
measurement
no
flow
is
present
between
node
1
and
the
cannulation
side,
the
venous
pressure
measured
is
equal
to
the
pressure
at
node
1.
J.
PhysioL.
480.2
345
I.
Vergroesen,
Y
Han,
M.
Goto
and
J:
A.
E.
Spaan
E
0-A
150
-
-
E
0-
J
-U
J J
-
-
0
5
10
15
Time
(s)
Figure
2.
Explanation
of
the
data
analysis
on
a
tracing
of
stopped-flow
venous
pressure
(top)
and
left
ventricular
pressure
(bottom)
during
a
long
diastole
The
vertical
lines
indicate
the
time
points
at
which
the
signals
are
analysed:
1,
end
diastole
in
the
last
beat
before
the
long
diastole;
2,
peak
systole
in
the
same
beat;
3,
early
minimum
in
long
diastole;
4,
onset
of
plateau
in
pressure;
5,
end
diastole;
and
6,
peak
systole
after
long
diastole.
At,
time
difference
between
time
4
and
time
3.
AP,
pressure
difference
along
the
tangent
between
time
4
and
time
3.
AP/At,
rate
of
pressure
change
at
a
point
between
time
3
and
time
4,
approximates
the
slope
of
the
tangent
at
that
point.
Intrinsic
tone
Vasodilatation
6
-
6-
-2u
-2
150-
150-
E
E
0-
0-
150
-
150
1-a
I
0E
0-Mx
ol--
E
E
M
oi
|
oj~~
~9
40
-
40
-
0
5
10
15
0
5
10
15
Time
(s)
Time
(s)
Figure
3.
Representative
tracings
of
measured
signals
during
a
prolonged
diastole
Left
panels,
recordings
made
during
intrinsic
coronary
tone;
right
panels,
recordings
made
during
vasodilatation
with
adenosine.
Qa
represents
coronary
arterial
flow,
PI,
is
coronary
perfusion
pressure,
Pv
is
stopped-flow
coronary
venous
pressure
and
PLV
is
left
ventricular
pressure.
Note
that
the
left
ventricular
pressure
pulse
decreases
after
a
prolonged
diastole,
while
coronary
venous
pressure
pulse
increases.
346
J.
PhysioL.
480.2
Cardiac
contraction
and
coronary
venous
pressure
pressure
to
the
change
in
venous
pulse
pressure
induced
by
the
long
diastole.
The
following
null
hypotheses
were
based
on
the
direct
transmission
of
left
ventricular
pressure
to
venous
pressure
with
an
average
transmission
over
the
myocardial
wall
of
0
5:
(a)
the
pulse
in
left
ventricular
pressure
is
directly
related
to
the
pulse
in
venous
pressure
(slope
of
regression
line
=
05),
and
(b)
the
changes
in
both
pressure
pulses
as
defined
above
are
related
to
each
other
with
a
slope
of
0
5.
The
influence
of
elastance
on
the
epicardial
venous
pressure
generation
was
tested
by
correlating
within
each
dog
end-
diastolic
venous
pressure
to
the
peak
systolic
venous
pressure
for
the
long-diastolic
beats.
Data
are
presented
as
means
+
S.E.M.
unless
indicated
otherwise.
Student's
t
test
was
used
to
evaluate
the
statistical
differences
of
the
estimated
regression
slopes
and
the
theoretically
expected
values.
P
<
0 05
was*
used
as
the
probability
level
for
statistical
significance.
RESULTS
Figure
3
depicts
two
typical
long
diastoles,
one
with
intrinsic
tone
(left)
and
one
during
maximal
vasodilatation
with
adenosine
(right).
The
duration
of
the
diastoles
varied
OU
I
Dog
A
40
80
E
E
X
40
v0
40-
.3
CLL
a)
:3
0
U)
CA
80
:3
0
'a
0CZ
LU
8
-
&
Intrinsic
tone
-052
±
0.08*t
Adenosine
-0-16
+
0.29*
Dog
C
<
0-01
±
004*
0
9---
-
'CW
-
Q-Ot
-',
-
-O
Dog
E
-010
±
0.11*
30
Dog
G
[0
~
~~
~~~~
'
°
--
0-22+±
0-65
012
±
006*t
0
I-el-
50
0-31
±
0
13t
100
150
347
between
0
45
and
13-53
s
with
a
mean
of
2-46
+
0-27
s
with
intrinsic
coronary
vascular
tone
and
between
0-46
and
1107
s
with
a
mean
of
2-59
+
0
24
s
during
adenosine
infusion.
In
the
diastoles
of
Fig.
3
venous
pressure
started
to
increase
at
the
onset
of
diastole
and
reached
a
steady
state
after
about
4
s.
With
intrinsic
tone,
only
34
out
of
161
diastoles
allowed
a
steady
state
in
Pv
to
be
reached,
while
with
adenosine
infusion
these
numbers
were
49
out
of
121.
The
average
diastolic
dPv/dt
with
vascular
tone
was
2-71
+
0-29,
and
during
adenosine
infusion
8-18
+
0-73
mmHg
s-i.
This
threefold
difference
in
rate
of
change
of
Pv
was
statistically
significant.
After
infusion
of
adenosine,
coronary
blood
flow
increased
about
threefold
in
the beating
heart
(mean
beat-'
+
S.E.M.
in
seven
dogs
was
1H
+
0
2
ml
s-5
(100
g)-'
with
intrinsic
tone
increasing
to
3-2
+
0-8
ml
s-5
(100
g)-'
after
vasodilatation
with
adenosine).
The
flow
increased
notwithstanding
a
decrease
in
perfusion
pressure
from
99+
7
to
50+
4
mmHg
due
to
pump
limitations.
Vasodilatation
had
no
significant
effect
on
left
ventricular
Dog
B
~
009±012*
005
±
0-09*
-1
_
-l
MB
Dog
D
-0.20
±
015*
lw_it-
F
-
-0
28
+
0.26*
]Dog
H
0
50
100
0-06
±
0Q09*
150
Left
ventricular
pressure
pulse
(mmHg)
Figure
4.
First
test
of
the
hypothesis
on
direct
transmission
of
left
ventricular
pressure
to
venous
pressure
The
figure
shows
the
linear
regression
between
the
coronary
venous
pressure
pulse
in
the
long
diastoles
and
the
left
ventricular
pressure
pulse
in
the
same
beat
for
each
dog
during
intrinsic
tone
and
during
vasodilatation
with
adenosine.
The
thick
dashed
lines
represent
the
expected
regression
lines
if
left
ventricular
pressure
directly
affects
the
venous
pressure
pulses.
The
numbers
on
the
left
side
of
each
panel
represent
the
estimated
regression
slope
(+
S.E.M.
of
the
estimate)
for
intrinsic
tone
while
those
on
the
right
side
indicate
the
regression
slope
during
vasodilatation.
*slope
significantly
different
from
0
5,
which
is
the
expected
average
transmission
of
left
ventricular
pressure
pulses;
tslope
significantly
different
from
zero).
In
none
of
the
dogs
did
the
expected
relation
for
the
direct
transmission
of
left
ventricular
pressure
fit
the
data.
J.
Physiol.
480.2
on
-0-04
±
0-13
*
10-!k
0--l
00-11,
0-11,
I.
Vergroesen,
Y
Han,
M.
Goto
and
J
A.
E.
Spaan
pressure
and
venous
pressure;
with
intrinsic
tone,
diastolic
PLV
=
9
+
2,
diastolic
Pv
(Pv,dia)
=
5
+
2,
systolic
PLV
=
95
+
6
and
systolic
P1
(Pv
sYs
)=
29
+
6
mmHg;
and
after
vasodilatation
with
adenosine,
diastolic
PLV
=
8
+
2,
Pv,
dia
=
4
+
1,
systolic
PLV
=
86
+
5
and
Pv=sys
33
+
7
mmHg).
Test
of
the
direct
transmission
of
left
ventricular
pressure
hypothesis
According
to
the
hypothesis
of
the
direct
transmission
of
left
ventricular
pressure
on
coronary
venous
pressure
generation,
the
pulses
of
the
PLV
and
Pv
should
be
related
by
a
factor
of
0
5.
The
ratio
of
the
venous
pressure
pulse
to
the
left
ventricular
pressure
pulse
was
in
the
long
diastolic
beats
044
+
022
(mean
+
S.D.)
with
intrinsic
tone
present.
After
vasodilatation
with
adenosine
this
ratio
was
053
+
024
(mean
+
S.D.).
The
regression
results
for
venous
and
ventricular
pressure
pulses
for
all
long
diastolic
beats
are
plotted
in
Fig.
4.
The
theoretically
expected
lines,
based
on
a
direct
transmission
of
left
ventricular
cavity
pressure
to
the
venous
pressure
pulse,
are
shown
by
dashed
lines
in
each
panel.
The
estimated
slopes
of
the
regression
lines
are
provided
in
each
panel,
with
the
value
for
intrinsic
tone
on
the
left-hand
side
and
with
vasodilatation
at
the
right-hand
side.
With
intrinsic
coronary
tone
all
slopes
of
the
regression
lines
were
significantly
different
from
the
theoretically
expected
value
of
0
5.
After
vasodilatation
with
adenosine
five
out
of
seven
slopes
of
regression
lines
were
significantly
different
from
05.
However,
in
none
of
the
dogs
did
the
expected
relation
fit
the
data.
When
we
compared
the
slopes
presented
in
Fig.
4
with
zero,
two
were
significantly
negative,
two
were
significantly
positive
and
all
others
were
not
significantly
different
from
zero.
In
addition
we
calculated
the
relation
between
systolic
venous
pressure
and
systolic
left
-
--
G
Intrinsic
tone
-*
Adenosine
40
Dog
A
'-054
+
0-25*t
°t~~
~ ~
~~
--
---.
--
--
-
...............
-0-29+±0-03*t
-20
_
40
Dog
C
20
0-00
±
008*
-20
_
40
Dog
E
0-18±
019
0....
...-...
...
-0-05
±
0-05*
3.i
.
-20
_
40
l
Dog
G
*
010±042
..................
p*
0.13_±0-10*
-30
40
20
0
40
20
0
401
20
0'
Dog
8
0-10
±
0-07*
.............................
;."
.
-0-1
6
±
0
13
*
Dog
D
,-0-14+±0-14*
..
.
.
1
.
_..
. .
Dog
F
0-03
±
0-16
0-2
..........................
..
..nqlc.............,
-0-26
±
0
04*t.-
_
:
401
Dog
H
20
o
-0.12
±
0.10*
0...................
...
i-0-30
±
0.13*t_
-30
0
30
pc3
0
-
6
0
-
L
eft
ventricular
pressure
pulse
change
(mmHg)
Figure
5.
Second
test
of
the
hypothesis
on
direct
transmission
of
left
ventricular
pressure
to
venous
pressure
The
figure
shows
the
linear
regression
of
the
coronary
venous
pressure
pulse
change
during
a
long
diastole
and
the
change
in
left
ventricular
pressure
pulse
for
8
animals
in
2
conditions,
intrinsic
tone
and
during
vasodilatation
with
adenosine.
The
thick
dashed
lines
represent
the
expected
regression
line
according
to
the
hypothesis
of
direct
transmission
of
left
ventricular
pressure
to
venous
pressure
generation.
The
thin
dotted
lines
indicate
zero
change
in
both
variables.
The
estimated
regression
slopes
(+
S.E.M.
of
the
estimate)
are
given
for
intrinsic
tone
in
the
left
bottom
corner
of
each
panel
and
during
vasodilatation
in
the
right
top
corner.
*
slope
significantly
different
from
05,
which
is
the
expected
average
transmission
of
left
ventricular
pressure
pulse
changes;
tslope
significantly
different
from
zero.
In
none
of
the
dogs
did
the
expected
relation
fit
the
data.
348
J.
Physiol.
480.2
'a
I
E
E
5,
CD
c
0
Co
0.
a)
C'
0
5)
a-
0
.-a
w
20
0-
I
_
%n
{1
1
-
4
e-
*
7.
-'V6
Cardiac
contraction
and
coronary
venous
pressure
ventricular
pressure.
Similar
results
were
found.
With
intrinsic
coronary
tone
five
slopes
were
significantly
negative,
one
significantly
positive
and
one
not
significantly
different
from
zero.
Compared
with
a
slope
of
05,
six
out
of
seven
slopes
were
significantly
different.
With
adenosine
six
slopes
were
not
significantly
different
from
zero
and
one
was
significantly
negative.
Compared
to
a
slope
of
05,
again
six
out
of
seven
slopes
were
significantly
different.
In
addition
we
calculated
the
relation
between
the
change
in
systolic
venous
pressure
and
the
change
in
systolic
left
ventricular
pressure
(Fig.
5).
Again
similar
results
were
found.
With
intrinsic
coronary
tone
four
slopes
were
significantly
negative,
none
were
significantly
positive
and
all
seven
were
significantly
different
from
0
5.
With
adenosine
one
was
significantly
negative,
none
significantly
positive
and
six
out
of
seven
were
significantly
different
from
05.
Test
of
the
elastance
hypothesis
According
to
the
elastance
hypothesis
a
long
diastole
would
induce
an
increase
in
vascular
volume
and
thus
in
end-diastolic
venous
pressure.
This
increase
in
volume
would
result
in
an
increase
in
systolic
venous
pressure.
Thus
as
a
test
of
the
elastance
hypothesis,
the
maximal
systolic
venous
pressure
(Pv,8Y)
in
the
first
contraction
after
the
prolonged
diastole
was
correlated
with
the
end
diastolic
venous
pressure
(Pv,dia).
The
regression
results
for
each
individual
dog
are
presented
in
Fig.
6.
The
estimated
slopes
and
standard
error
of
the
individual
regression
lines
are
given
in
the
figure.
Values
obtained
after
vaso-
dilatation
with
adenosine
are
placed
at
the
right-hand
side
of
each
panel
and
for
intrinsic
tone
at
the
bottom
middle.
All
regression
lines
were
significantly
different
from
zero.
According
to
the
elastance
hypothesis,
the
generation
of
venous
pressure
would
be
sensitive
to
a
change
in
inotropic
-
-E-
Intrinsic
tone
Dog
A
-_-
Adenosine
2-5
+
0Q03*
80'
3-6
±
0-3*
1
Dog
C
10.
O
1.2
±
01*
0-
.
Dog
E
1-2
+
0.2*
10.
0g
Dog
G
9
+
01*
1.0
+
0
1
*
20
40
8
Dog
B
2-0
±
0-1
*
2-1
±
0.5
*
o
0-
.
Dog
D
1
-7
±
0.1
*
0
.Dog
F
30
34
±
1.1
*
j.ogH
28
±+0
03*
.Dog
H1
30
-1-1
±
0-2*
*
8
+
2
O
..
.
_
.
.
...0
2
20
40
Diastolic
venous
pressure
(mmHg)
Figure
6.
Test
of
the
time-varying
elastance
hypothesis
as
an
explanation
for
the
generation
of
coronary
venous
pressures
According
to
this
hypothesis
the
systolic
venous
pressure
depends
on
the
filling
of
the
coronary
venous
compartment.
In
diastoles
of
variable
duration
end-diastolic
venous
pressure
is
used
as
indication
of
this
filling
and
is
related
to
the
systolic
pressure
in
the
following
contraction.
Each
panel
represents
1
animal
both
during
intrinsic
tone
and
during
vasodilatation
with
adenosine.
The
dotted
regression
lines
were
fitted
through
the
open
circles,
and
represent
intrinsic
tone
for
each
animal.
The
continuous
lines
were
fitted
to
the
circles
and
apply
to
the
vasodilated
coronary
bed.
The
slopes
of
the
regression
lines
and
the
standard
error
of
the estimates
are
shown
during
vaso-
dilatation
at
the
right
top
corner
of
each
panel
and
during
intrinsic
tone
in
the
middle
at
the
bottom
of
each
panel.
All
were
significantly
different
from
zero,
as
indicated
by
the
asterisk.
J.
Physiol.
480.2
349
80
E:
81
E
E
CD
b-
U)
a)
:O
U)
0
0
>
81
v)
u0
80-
0
8
8
I.
Vergroesen,
Y
Han,
M.
Goto
and
J
A.
E.
Spaan
140
-
0
20
Area
(a.u.)
,
Long
diastolic
beat
45
Figure
7.
Test
for
a
change
in
maximal
end-systolic
elastance
due
to
the
prolongation
of
diastole
Two
ultrasonic
crystals
were
sewn
on
either
side
of
the
left
ventricle.
The
continuously
measured
distance
between
these
crystals
was
squared
and
used
as
measure
for
the
'volume'
of
the
ventricle.
The
thus
obtained
measure
was
named
'area'
and
expressed
in
arbitrary
units
(a.u.).
Analogous
to
the
pressure-volume
loops
of
Suga,
Sagawa
&
Shoukas
(1973),
pressure-area
loops
were
constructed
of
several
normal
beats
and
one
beat
containing
a
long
diastole.
Maximal
elastance
of
each
beat
was
defined
as
maximal
left
ventricular
pressure
(PLV)
at
minimal
area
or
'volume'.
As
can
be
seen
here
maximal
elastance
in
the
long
diastolic
beat
is
not
different
from
the
maximal
elastance
in
the
normal
beats.
state
of
the
heart
muscle
due
to
a
change
in
end-systolic
elastance.
We
expected
no
change
in
inotropic
state
as
a
result
of
the
prolongation
of
diastole.
This
was
verified
in
five
dogs- equipped
with
ultrasound
crystals
by
(D
CO)
CO)
i
._
E0
.0
0
E
ITO
CS
construction
of
pressure-area
loops
of
each
long
diastolic
beat
and
several
normal
beats.
A
typical
result
is
depicted
in
Fig.
7.
The
end-systolic
points
of
the
normal
beats
were
used
as
an
indication
of
normal
maximal
elastance.
The
tan
(a(t))
Diastolic
Intramyocardial
blood
volume
Figure
8.
Schematic
explanation
of
the
predictions
of
the
time-varying
elastance
hypothesis
for
the
coronary
venous
compartment
During
normal
beating,
the
elastance,
AP/A
V=
tan
(a
(t))
of
the
ventricle
changes
from
the
diastolic
to
the
end-systolic
value.
According
to
the
time-varying
elastance
hypothesis,
this
change
in
elastance
affects
all
compartments
inside
the
heart
muscle.
In
the
venous
compartment
no
valves
are
present,
thus
an
increase
in
developed
venous
pressure
leads
instantaneously
to
a
decrease
in
venous
volume.
However,
if
the
end-diastolic
volume
increases,
as
indicated
by
the
arrow
along
the
diastolic
pressure-volume
curve
(for
instance
by
prolongation
of
diastole),
systolic
venous
pressure
will
be
increased
in
the
following
systole,
depending
on
the
amount
of
volume
displaced,
irrespective
of
the
developed
left
ventricular
pressure
(figure
modified
from
Westerhof,
1990).
A
change
in
inotropic
state
of
the
ventricle
will
affect
the
venous
pressure
pulse
by
changing
the
slope
of
the
end-systolic
elastance
relation.
350
J.
Phy8iol.
480.2
Cardiac
contraction
and
coronary
venous
pressure
end
systolic
elastance
of
the
beat
containing
the
longer
diastole
was
then
compared
to
the
end-systolic
elastance
of
the
normal
beats.
From
the
pressure-area
plots
we
concluded
that
maximal
elastance
of
the
long
diastolic
beats
was
not
substantially
altered
compared
with
that
pertaining
in
normal
beats.
DISCUSSION
In
situ
perfused
heart
preparation
Cannulation
and
perfusion
of
the
left
main
coronary
artery
is
a
classic
procedure
used
to
study
coronary
arterial
pressure-flow
relations
and
the
effect
of
contraction
on
myocardial
perfusion
(Gregg
&
Green,
1940).
It
is
unlikely
that
this
artificial
perfusion
influenced
our
results
on
venous
pressure.
The
waves
in
coronary
arterial
and
coronary
venous
flow
are
uncoupled
in
the
normal
beating
heart
due
to
a
large
intramyocardial
compliance
(Spaan
et
al.
1983).
An
advantage
of
artificial
perfusion
and
pacing
is
that
the
effects
of
reflexes
from
the
heart
on
the
coronary
circulation
are
minimized,
since
possible
reflex-induced
changes
in
left
ventricular
pressure
and
atrial
rate
cannot
affect
coronary
perfusion
pressure
and
ventricular
rate
in
this
preparation.
Finally,
possible
effects
of
reflexes
on
ventricular
contractility
were
unlikely
in
the
present
study
as
this
would
have
been
clear
from
our
pressure-area
loops.
We
did
not
measure
right
atrial
pressure
in
our
experiment,
due
to
the
limited
number
of
pressure
amplifiers
we
had
available.
Right
atrial
pressure
would
be
an
important
variable,
if
an
increase
in
diastolic
coronary
venous
pressure
could
be
caused
by
an
increase
in
atrial
pressure.
However,
from
earlier
studies
(Spaan,
1985)
we
know
that
after
coronary
arterial
occlusion
during
a
long
diastole,
intramyocardial
blood
volume
decreases
at
constant
epicardial
venous
pressure,
which
makes
it
unlikely
that
our
retrogradely
measured
increase
in
venous
pressure
was
dominated
by
right
atrial
pressure.
Moreover,
the
rate
of
change
of
coronary
venous
pressure
at
the
beginning
of
diastole
would
not
be
dependent
on
vasomotor
tone
if
right
atrial
pressure
were
the
dominant
factor.
Further,
prolonged
diastoles
occur
naturally
in
chronically
instrumented
conscious
dogs
and
were
used
in
the
study
of
coronary
mechanics
by
Bellamy
(1978).
We
found
that
right
atrial
pressure
(in
our
preparation
atrial
rate
was
between
90
and
120
beats
min-')
could
only
have
a
small
effect
on
coronary
venous
pressure
during
the
long
diastole,
as
is
clear
from
Fig.
2.
Thus,
we
feel
that
we
can
discount
the
possibility
that
changes
in
atrial
pressure
affected
our
results.
The
measurement
of
stopped-flow
venous
pressure
as
an
indication
of
pressure
in
venules
The
coronary
venous
bed
in
the
dog
is
rich
in
anastomoses
1988).
Cannulation
of
one
small
vein
near
the
place
where
it
emerges
out
of
the
heart
muscle,
and
measuring
stopped-flow
pressure,
allows
an
estimate
of
the
pressure
at
the
nearest
network
node
point
inside
the
muscle
(see
Fig.
1).
The
absolute
pressure
measured
depends
strongly
on
the
resistance
of
the
alternative
outflow
channels
and
the
blood
flow
level.
Since there
are
anastomoses
in
the
venous
bed
(Grayson,
Davidson,
Fitzgerald-Finch
&
Scott,
1974;
Cohen
et
al.
1988),
the
pressure
measured
is
an
underestimation
of
the
capillary
pressure
in
the
region
drained
by
the
cannulated
vein.
It
could
be
argued
that
this
pressure
might
have
been
increased
above
normal
due
to
flow
obstruction.
However,
Tillmanns,
Steinhausen,
Leinberger,
Thederan
&
Kiibler
(1981)
measured
coronary
pressures
in
small
epicardial
veins
in
the
rat
and
cat
using
the
servo-nulling
technique
described
by
Fox
&
Wiederhielm
(1973).
In
the
normal
beating
heart,
peak
systolic
pressures
were
around
25
mmHg
and
diastolic
pressures
around
5
mmHg.
These
values
are
very
similar
to
the
values
of
30
and
5-6
mmHg
found
by
the
present
method
in
the
last
beat
before
a
long
diastole.
Further,
Chilian,
Eastham
&
Marcus
(1986)
showed
that
small
venous
pressure
in
the
cat
heart
increased
with
vasodilatation,
up
to
32
mmHg
in
systole
and
5
mmHg
in
diastole.
This
again
is
quite
similar
to
our
values
for
the
last
beats
before
the
long
diastole
during
vasodilatation
with
adenosine:
34
mmHg
systolic
pressure
and
4-6
mmHg
diastolic
pressure.
These
similarities
support
the
idea
that
stopped-flow
pressures
in
the
veins
cannulated
by
us
do
reflect
normal
pressures
in
the
microcirculation
distal
to
the
capillary
bed.
The
measurement
of
stopped-flow
pressure
in
one
small
coronary
vein
is
not
harmful
to
the
myocardial
tissue,
since
many
venous
collaterals
exist
in
the
dog
myocardium
(Grayson
et
al.
1974;
Cohen
et
al.
1988).
However,
anatomical
differences
between
dogs
do
result
in
inter-animal
variability
of the
absolute
pressure
values.
These
differences
cannot
be
explained
by
invoking
either
of
the
hypotheses
we
set
out
to
test.
However,
because
these
were
differences
between
dogs
we
have
presented
the
data
for
each
dog
separately.
Generation
of
pressure
pulses
in
intramyocardial
veins
The
pulsations
that
occur
in
stopped-flow
venous
pressure
must
have
been
caused
by
cardiac
contraction.
The
hypothesis
on
the
direct
transmission
of
left
ventricular
pressure
predicts
a
ratio
between
the
two
pressure
pulsations
of
about
0
5
based
on
the
space
average
over
the
left
ventricular
wall
(Downey
&
Kirk,
1975;
Hoffman
&
Spaan,
1990).
In
this
study
in
the
long
diastolic
beats
an
average
pulse
ratio
of
0
44
+
0-22
(mean
+
S.D.)
was
found
with
intrinsic
coronary
tone
and
after
vasodilatation
with
adenosine
this
number
was
0
53
+
0-24
(mean
+
S.D.).
This
small
difference
from
0
5
would
have
been
explicable
by
between
veins
at
all
levels
(Cohen,
Matsuki
&
Downey,
J.
PhyioL.
480.2
351
the
influence
of
venous
anatomy.
However,
the
absence
of
I.
Vergroesen,
Y.
Han,
M.
Goto
and
J:
A.
E.
Spaan
the
expected
relationship
between
the
change
in
magnitude
of
the
pulsations
in
the
venous
and
left
ventricular
pressure
leaves
little
room
for
the
direct
transmission
of
left
ventricular
pressure
to
be
the
compressive
force
for
the
intramyocardial
veins
(see
Fig.
5).
An
alternative
hypothesis
to
explain
pressure
generation
in
intramyocardial
vessels
put
forward
by
Krams
et
al.
(1989)
and
Westerhof
(1990)
predicts
that
the
pressure
in
the
intramyocardial
vessels
depends
on
cyclical
elastance
variations
and
vascular
volume
(see
Introduction).
As
is
illustrated
in
Fig.
8,
if
this
were
the
case,
then
the
maximal
systolic
venous
pressure
at
constant
maximal
elastance
would
be
determined
by
the
end-diastolic
vascular
volume.
Krams
et
al.
(1989)
showed
in
an
isolated
cat
heart
preparation
the
dependence
of
the
coronary
arterial
flow
pulsations
on
the
inotropic
state
of
the
heart
muscle.
Thus,
a
change
in
end-systolic
elastance
of
the
heart
muscle
may
determine
the
pressure
pulsations
inside
any
compartment
in
the
myocardium.
From
those
five
animals
in
which
we
measured
pressure-area
loops,
no
indication
was
found
of
an
increase
of
end-systolic
elastance
in
the
first
contraction
following
a
long
diastole
(see
Fig.
7).
If
direct
transmission
of
left
ventricular
pressure
was
not
the
cause
(see
above)
and
time-varying
elastance
was
unaltered,
an
increase
in
intramyocardial
vascular
volume
is
the
only
explanation
left
for
the
increased
venous
pressure
pulse
during
a
long
diastolic
beat.
We
have
shown
previously
(Vergroesen
et
al.
1987)
that
cardiac
arrest
induces
a
blood
volume
increase
that
follows
a
time
course
quite
similar
to
the
venous
pressure
measured
in
this
study.
Further,
the
increase
in
blood
volume
was
quicker
after
vasodilatation
with
adenosine,
just
as
was
the
case
for
the
rate
of
increase
in
Pv
during
the
first
seconds
of
diastole
in
the
present
study.
Therefore,
the
correlation
seen
between
systolic
and
end-diastolic
pressure
can
be
readily
explained
on
the
basis
of
the
elastance
hypothesis.
The
fact
that
systolic
venous
pressure
peaked
at
the
end
of
systole
is
also
in
agreement
with
the
time
course
of
time-varying
myocardial
elastance,
in
that
maximal
elastance
is
found
at
the
end
of
systole
(Suga,
Sagawa
&
Shoukas,
1973).
Maximal
left
ventricular
pressure
occurs
early
in
systole
and
at
that
time
venous
pressure
is
still
rising.
Thus,
the
coincidence
of
maximal
elastance
and
maximal
coronary
venous
pressure
is
further
evidence
of
the
role
of
time-varying
elastance
in
the
generation
of
pressure
in
the
coronary
veins.
In
the
present
study
all
estimated
slopes
of
the
regression
relations
between
diastolic
venous
pressure
and
systolic
venous
pressure
(Fig.
6)
were
significantly
different
from
zero.
This
is
in
agreement
with
the
elastance
hypothesis.
However,
in
the
way
the
hypothesis
is
schematically
explained
in
Fig.
8
one
would
expect
that
a
change
in
diastolic
venous
pressure
would
induce
a
more
than
proportional
change
in
systolic
venous
pressure.
were
not
significantly
different
from
1
(see
Fig.
6).
This
implies
that
the
increase
in
end-diastolic
venous
volume
that
coincides
with
the
increased
diastolic
venous
pressure
does
not
induce
a
larger
pulse
in
venous
pressure.
However,
reality
may
deviate
from
the
schematic
in
Fig.
8
in
several
ways.
In
the
first
place
the
diastolic
elastance
relation
might
not
be
a
straight
line,
but
curvilinear
with
a
steeper
curve
at
higher
volumes.
It
is
then
possible
that
the
diastolic
and
systolic
elastance
relations
become
parallel
over
the
diastolic
venous
pressure
range
measured
in
these
dogs.
This
indeed
would
imply
a
venous
pressure
pulse
independent
of
diastolic
venous
filling.
Additionally
it
is
quite
possible
that
at
higher
diastolic
volumes
the
venous
compartment
is
emptied
more
readily.
Hence,
the
venous
pressure
generated
during
a
beat
will
be
reduced
due
to
the
higher
volumes
expelled
simultaneously.
According
to
the
elastance
hypothesis,
if
contraction
is
stopped
in
a
segment
of
the
myocardium,
diastolic
and
systolic
pressure
in
the
venous
compartment
of
this
segment
would
be
equal
(i.e.
no
pulses
would
be
generated)
the
regression
slope
would
be
1
and
the
intercept
would
be
zero.
However,
pulses
are
generated
in
all
our
experiments
(see
Fig.
4)
and
the
regression
lines
all
have
intercepts
well
above
zero
(see
Fig.
6).
Thus,
this
property
of
the
elastance
hypothesis
cannot
be
responsible
for
the
absence
of
a
relation
between
diastolic
venous
pressure
and
venous
pulse
pressure
in
six
out
of
fourteen
cases.
In
conclusion,
the
data
are
consistent
with
the
elastance
hypothesis,
but
a
quantitative
prediction
of
the
effect
of
contraction
on
venous
pressure
generation
requires
a
method
for
the
measurement
of
instantaneous
intramural
venous
volume.
Such
a
method
is
not
available
(Van
der
Ploeg,
Dankelman
&
Spaan,
1993)
Comparison
of
the
data
with
the
literature
Tissue
pressure
as
measured
by
Rabbany,
Kresh
&
Noordergraaf
(1989)
by
the
introduction
of
pressure
transducers
on
top
of
a
micro-tip
catheter
within
the
myocardium
also
exhibited
an
increased
systolic
pressure
in
the
first
beat
after
a
long
diastole.
It
is
most
likely
that
the
explanation
for
their
findings
is
similar
to
the
one
given
for
our
venous
pressure
measurement.
It
is
not
unlikely
that
interstitial
fluid
accumulated
around
their
catheter
tip
in
the
long
diastole,
resulting
in
the
increased
systolic
pressure
in
the
first
beat
thereafter.
Contraction
would
rapidly
expel
this
extra
fluid
and
hence
pressure
around
the
sensor
would
become
normal.
The
absence
of
a
direct
influence
of
left
ventricular
pressure
on
pressure
generation
in
the
venous
compartment
does
not
imply
that
it
could
not
be
true
for
other
intramyocardial
blood
compartments.
It
was
recently
shown
that
when
perfused
at
constant
flow
the
pressure
pulsations
in
the
coronary
arterial
pressure
closely
follow
the
left
ventricular
pressure
variations,
with
an
average
regression
slope
of
0
3
(Kouwenhoven
et
al.
1992).
However,
mid-systolic
However,
in
six
out
of
fourteen
cases
regression
slopes
352
J.
Physiol.
480.2
coronary
flow
at
constant
pressure
perfusion
was
insensitive
J.
Physiol.
480.2
Cardiac
contraction
and
coronary
venous
pressure
tO
PLV.
It
was
suggested
that
this
difference
in
dependency
on
PLV
is
due
to
the
ability
of
the
arteriolar
compartment
to
change
volume
at
constant
pressure
perfusion,
whereas
this
effect
is
prevented
at
constant
flow
perfusion.
This
explanation
is
in
agreement
with
the
present
finding
of
no
direct
transmission
of
left
ventricular
influence
on
venous
pressure
generation,
since
blood
should
be
easily
expelled
from
the
intramyocardial
veins
during
systole.
The
elastance
concept
is
also
consistent
with
the
generation
of
pressure
in
intramyocardial
lymph
vessels
(Han,
Vergroesen
&
Spaan,
1993
b).
However,
in
contrast
to
the
present
study,
recent
experiments
performed
by
our
group
(Han,
Vergroesen,
Goto,
Dankelman,
van
der
Ploeg
&
Spaan,
1993a)
showed
that
left
ventricular
pressure
affected
lymph
pressure
during
diastole
10
times
more
than
during
systole.
This
difference
in
sensitivity
was
not
found
in
the
venous
compartment.
As
yet,
we
do
not
have
an
explanation
for
this
difference
but
it
indicates
the
importance
of
considering
the
effect
of
heart
contraction
on
each
intramural
compartment
seperately.
Implication
of
the
findings
for
the
coronary
circulation
The
present
results
illustrate
that
the
time-varying
elastance
hypothesis
as
an
explanation
for
phasic
changes
in
coronary
arterial
flow
and
lymphatic
pressure
is
also
consistent
with
the
generation
of
coronary
venous
pressure.
Indeed
time-varying
elastance
as
the
cause
of
intramyocardial
venous
pressure
generation
implies
an
effective
mechanism
to
empty
the
coronary
vessels.
Thus,
a
local
increase
of
outflow
resistance
would
cause
local
venous
pressure
to
rise
thereby
increasing
the
driving
pressure
for
venous
outflow.
Moreover,
unperfused
'sleeping'
anastomoses
might
be
opened
and
the
pressure
would
tend
to
normalize
via
increased
flow
through
all
alternative
outflow
channels.
In
other
words,
if
local
intra-
myocardial
venous
pressure
is
determined
by
variations
in
elastance
this
would
form
an
intrinsic
control
mechanism
that
would
tend
to
optimize
coronary
venous
outflow.
REFERENCES
BELLAMY,
R.
F.
(1978).
Diastolic
coronary
artery
pressure-flow
relations
in
the
dog.
Circulation
Research
43,
92-101.
CHILIAN,
W.
M.,
EASTHAM,
C.
L.
&
MARCUS,
M.
L.
(1986).
Microvascular
distribution
of
coronary
vascular
resistance
in
beating
left
ventricle.
American
Journal
of
Physiology
251,
H779-788.
COHEN,
M.
V.,
MATSUKI,
T.
&
DOWNEY,
J.
M.
(1988).
Pressure-flow
characteristics
and
nutritional
capacity
of
coronary
veins
in
dogs.
American
Journal
of
Physiology
255,
H834-846.
DOWNEY,
J.
M.
&
KIRK,
E.
S.
(1975).
Inhibition
of
coronary
blood
flow
by
a
vascular
waterfall
mechanism.
Circulation
Research
36,
753-763.
Fox,
J.
R.
&
WIEDERHIELM,
C.
A.
(1973).
Characteristics
of
the
servo-controlled
micropipet
pressure
system.
Microvascular
Research
5,
753-760.
GRAYSON,
J.,
DAVIDSON,
J.
W.,
FITZGERALD-FINCH,
A.
&;
SCOTT,
C.
(1974).
The
functional
morphology
of
the
coronary
microcirculation
in
the
dog.
Microvascular
Research
8,
20-43.
GREGG,
D.
E.
&
GREEN,
H.
D.
(1940).
Registration
and
interpretation
of
normal
phasic
inflow
into
a
left
coronary
artery
by
an
improved
differential
manometric
method.
American
Journal
of
Physiology
130,114-125.
HAN,
Y.,VERGROESEN,
I.,
GOTO,
M.,
DANKELMAN,
J.,
VAN
DER
PLOEG,
C.
P.
B.
&
SPAAN,
J.
A.
E.
(1993a).
Left
ventricular
pressure
transmission
to
myocardial
lymph
vessels
to
myocardial
lymph
vessels
is
different
during
systole
and
diastole.
Pfluigers
Archiv
423,
448-454.
HAN,
Y.,
VERGROESEN,
I.
&
SPAAN,
J.
A.
E.
(1993b).
Stopped-flow
epicardial
lymph
pressure
is
affected
by
left
ventricular
pressure
in
anesthetized
goats
American
Journal
of
Physiology
264,
H1624-1628.
HOFFMAN,
J.
I.
E.
&
SPAAN,
J.
A.
E.
(1990).
Pressure-flow
relations
in
coronary
circulation.
Physiological
Reviews
70,
331-390.
KOUWENHOVEN,
E.,
VERGROESEN,
I.,
HAN,
Y.
&
SPAAN,
J.
A.
E.
(1992).
Retrograde
coronary
flow
is
limited
by
time
varying
elastance.
American
Journal
of
Physiology
63,
H484-490.
KRAMS,
R.,
SIPKEMA,
P.
&
WESTERHOF,
N.
(1989).
The
varying
elastance
concept
may
explain
coronary
systolic
flow
impediment.
American
Journal
of
Physiology
257,
H1471-1479.
RABBANY,
S.
Y.,
KRESH,
J.
Y.
&
NOORDERGRAAF,
A.
(1989).
Intramyocardial
pressure:
Interaction
of
myocardial
fluid
pressure
and
fibre
stress.
American
Journal
of
Physiology
257,
H357-364.
SPAAN,
J.
A.
E.
(1985).
Coronary
diastolic
pressure
flow
relation
and
zero
flow
pressure
explained
on
the
basis
of
intra-
myocardial
compliance.
Circulation
Research
56,
293-309.
SPAAN,
J.
A.
E.,
BRUINSMA,
P.
&
LAIRD,
J.
D.
(1983).
Coronary
flow
mechanics
of
the
hypertrophied
heart.
In
Cardiac
Left
Ventricular
Hypertrophy,
ed.
TER
KEURS,
H.
E.
D.
J.
&
SCHIPPERHEIN,
J.
J.,
pp.
170-193.
Martinus
Nijhoff
Publishers,
Dordrecht,
The
Netherlands.
STEINER,
C.
&
KOVALIK,
A.
T.
W.
(1968).
A
simple
technique
for
production
of
chronic
complete
heart
block
in
dogs.
Journal
of
Applied
Physiology
25,631-632.
SUGA,
H.,
SAGAWA,
K.
&
SHOUKAS,
A.
A.
(1973).
Load
independence
of
the
instantaneous
pressure-volume
ratio
of
the
canine
left
ventricle
and
effects
of
epinephrine
and
heart
rate
on
the
ratio.
Circulation
Research
32,
314-322.
TILLMANS,
H.,
STEINHAUSEN,
M.,
LEINBERGER,
H.,
THEDERAN,
H.
&
KtBLER,
W.
(1981).
Pressure
measurements
in
the
terminal
vascular
bed
of
the
epimyocardium
of
rats
and
cats.
Cardiovascular
Research
49,
1202-1211.
VAN
DER
PLOEG,
C.
P.
B.,
DANKELMAN,
J.
&
SPAAN,
J.
A.
E.
(1993).
Functional
distribution
of
coronary
vascular
volume
in
beating
goat
hearts.
American
Journal
of
Physiology
264,
H770-776.
VERGROESEN,
I.,
NOBLE,
M.
I.
M.
&
SPAAN,
J.
A.
E.
(1987).
Intramyocardial
blood
volume
changes
in
first
moment
of
cardiac
arrest
in
anesthetized
goats.
American
Journal
of
Physiology
253,
H307-316.
WESTERHOF,
N.
(1990).
Physiological
hypotheses
-
intramyocardial
pressure.
A
new
concept,
suggestions
for
measurements.
Basic
Research
in
Cardiology
85,105-119.
Acknowledgements
The
authors
thank
Mr
A.
Boekee
for
his
assistance
with
the
experiments
and
Dr
H.
B.
van
Wezel
for
his
advice
on
the
anaesthesia
of
the
dogs.
We
thank
Professor
N.
Westerhof
for
the
use
of
the
ultrasonic
crystals
and
equipment
developed
in
his
laboratory.
Received
22
October
1992;
accepted
1
March
1994.
