Download Effects of increased pressure inside or outside ventricles - AJP

Am J Physiol Heart Circ Physiol
279: H2927–H2938, 2000.
Effects of increased pressure inside or outside
ventricles on total and regional myocardial blood flow
G. S. ALDEA, H. MORI, W. K. HUSSEINI, R. E. AUSTIN†, AND J. I. E. HOFFMAN
Cardiovascular Research Institute and the Department of Pediatrics,
University of California, San Francisco, California 94143-0544
Received 24 January 2000; accepted in final form 19 July 2000
subendocardial flows; subepicardial flows; zero flow pressure,
coronary venous pressure; pressure-flow curves
or outside the heart have
been associated with subendocardial ischemia and
hemorrhage (10, 20). Investigators from Sendai, Japan, (28, 35) showed that increasing pressures outside
the heart or in either ventricle in an isolated supported
dog heart increased the zero flow pressure (Pzf), the
coronary arterial perfusing pressure at which flow
ceases. The experiments were done with maximally
dilated vessels during episodes of postpacing arrest.
The greatest effect was shown by increasing pericardial, or surrounding heart pressure (SHP); for increases of pericardial pressure from 0 to 15 and then to
30 mmHg, Pzf increased from 6 to 22 mmHg and then
to 36 mmHg. Similar increases in right ventricular
pressure (RVP) raised Pzf levels from 0 to 15 mmHg
and then 30 mmHg, whereas similar increases in left
ventricular pressure (LVP) increased Pzf from 6 to 12
mmHg and then to 15 mmHg. Such increases in Pzf
PRESSURE ELEVATIONS INSIDE
† Deceased 22 August 1998.
Address for reprint requests and other correspondence: J. I. E.
Hoffman, Box 0544, Univ. of California, San Francisco, CA 941430544 (E-mail: [email protected]).
http://www.ajpheart.org
would be expected to decrease driving pressures across
the coronary circulation and so render the heart liable
to global or regional ischemia.
We wondered whether these different pressure
changes would have different effects on flows in different layers of the left ventricular free wall, hypothesizing that the effects would be to decrease flows predominantly in the subendocardium for increases in LVP,
predominantly in the subepicardium for increased
pericardial pressures, and perhaps more evenly across
the wall for increases in RVP. We also wondered
whether the mechanism for the changes in Pzf involved
the extramural venous waterfall that had been described previously by Scharf et al. (29) and Uhlig et al.
(31) or whether it involved selective drop out of layers
or smaller regions of the left ventricle.
To examine these problems, we developed an isolated blood-perfused dog heart and repeated the experiments described above during cardioplegia, but with
the addition of radioactive microspheres to measure
regional flows. Because ischemia does not usually occur with adequate perfusion pressures, we concentrated on low perfusing and driving pressures. Furthermore, because we were becoming aware of the
heterogeneity of myocardial blood flow, we examined
not only average flow values in any myocardial layer,
but in some experiments we examined the flows in
several pieces of heart muscle within the myocardial
layers.
METHODS
Preparation of the Heart
The experimental protocol described was approved by the
Committee on Animal Research of the University of California at San Francisco and was in compliance with the “Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the National
Institutes of Health Guide for the Care and Use of Laboratory
Animals [DHHS Publication No. (NIH) 80–23, Revised 1985,
Office of Science and Health Reports, Bethesda, MD 20892].
We induced anesthesia in 40 healthy adult mongrel dogs
weighing between 12.6 and 32.9 kg with 2 ml im Innovar
followed by pentobarbital sodium (10–20 mg/kg iv) and mainThe costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
0363-6135/00 $5.00 Copyright © 2000 the American Physiological Society
H2927
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
Aldea, G. S., H. Mori, W. K. Husseini, R. E. Austin, and
J. I. E. Hoffman. Effects of increased pressure inside or
outside ventricles on total and regional myocardial blood
flow. Am J Physiol Heart Circ Physiol 279: H2927–H2938,
2000.—Increasing pressures to 30 mmHg in right (RV) and
left (LV) ventricles and surrounding heart (SH) in isolated,
arrested, maximally vasodilated, blood-perfused dog hearts
shifted pressure-flow (PF) curves rightward and increased
zero flow pressure (Pzf) by an amount equal to the RV applied
pressure, SH applied pressure, or two-thirds of the LV applied pressure. There were comparable increases in coronary
venous pressures. Increasing LV or SH pressures decreased
coronary blood flows, especially in the subendocardium. Decreases in driving pressure decreased flows in all layers, but
even with driving pressure of 5 mmHg, a few subepicardial
pieces had flow. We conclude with the following: 1) raising
pressures inside or outside the heart shifts PF curves and
raises Pzf by increasing coronary venous pressure; 2) the
effects are most prominent in the subendocardial muscle
layer; and 3) as driving pressures are decreased, there is a
range of Pzf in the heart with the final Pzf recorded due to the
last little piece of muscle to be perfused.
H2928
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
Fig. 1. Diagram of the preparation. Aorta (Ao); great
cardiac vein (GCV); inferior vena cava (IVC); left
atrium (LA); left main coronary artery (LMCA); left
ventricle (LV); right atrium (RA); right ventricle (RV);
superior vena cava (SVC); H2O heating for oxygenator; O2 to and from oxygenator. The dashed lines
indicate intracardiac or intravascular tubing; all
these tubes have only one end hole. Tubing not shown
to scale. Drainage from RA to oxygenator is by gravity.
was filtered through a 20-␮m PDF filter (Fenwal Electronics/
APD, Milford, MA) and infused with a roller pump (Masterflex R-7523, Cole-Parmer, Chicago, IL) through the Gregg
cannula. For the rest of the experiment, blood was kept at
37°C. The ventricles were vented, and blood from the right
atrium was returned to the oxygenator. In each dog, coronary
perfusion was reestablished within 15 min after aortic cross
clamping, and cardiac arrest was maintained for the rest of
the experiment.
The heart was then excised from the body by ligating and
cutting the appropriate vessels. The right coronary artery
was ligated. Large bore tubes were placed in each ventricle
through the atria and connected to reservoirs that could be
set at any desired pressures (Fig. 1). These tubes had bullet
tips with a single draining hole. The positions of the tips of
these tubes were verified by the response to fluid infusion. If
the end of the tube was in the ventricle, then that ventricle
was distended by the fluid. If in error, the end of the tube was
in the atrium, then the atrium and not the ventricle was
distended; if that occurred the tube was repositioned. The
reservoirs were also raised and then returned to control level
to make sure that all blood returned to the reservoir and was
not lost from the heart by leakage. The stump of the superior
vena cava was tied around the cannula that entered the right
ventricle. The stump of the inferior vena cava was tied
around a catheter that drained blood from the right atrium
and the coronary sinus by gravity to the oxygenator.
Fluid-filled catheters for pressure measurement were
placed in the left ventricle and the coronary sinus or great
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
tained anesthesia with 1% halothane in oxygen. We ventilated the dogs’ lungs through a cuffed endotracheal tube with
a Harvard pump (Harvard Apparatus, S. Natick, MA). We
kept arterial blood gases and pH within the normal range by
adjusting inspired oxygen concentration, minute ventilation,
or giving boluses of sodium bicarbonate. A heating pad was
used to keep the dogs’ body temperature close to 37°C.
Bilateral thoracotomies were performed and the sternum
was divided. A modified Gregg cannula (4 mm, ID 3 mm) was
introduced into the aorta through a left subclavian arteriotomy and secured in place with its tip just above the aortic
valve. The dogs received 10,000 units iv of heparin sodium.
They also received 5,000 units iv every 30 min throughout
the experiment. A bolus of dipyridamole (2 mg iv) was also
given.
The heart was isolated by clamping the azygos vein, inferior and superior vena cava, pulmonary hilar vessels, brachiocephalic trunk, and descending aorta. The femoral arteries were cannulated with 24-Fr cannulas. As soon as the
aorta was clamped, oxygenated St. Thomas’ Hospital cardioplegic solution (24) at 4°C was infused into the proximal
aortic root at 60 mmHg pressure through the Gregg cannula.
Asystole always occurred within 10 s of cross clamping.
Each dog was exsanguinated through the femoral vessels,
and the blood was collected for coronary perfusion. After
reoxygenation with a Bio-2 membrane oxygenator (Bentley
Laboratories, Irvine, CA) and the addition of potassium chloride (10 meq/l) and carbochromene hydrochloride (40 mg/kg,
Hoechst-Roussel Pharmaceuticals, Somerville, NJ), blood
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
Regional Flow Determinations
Regional myocardial flow was measured with radioactive
microspheres 15 ⫾ 3 ␮m in diameter (means ⫾ SD) labeled
with one of nine radionuclides (157Gd, 57Co, 114In, 51Cr,
113
Sn, 85Sr, 95Nb, 54Mn, and 65Zn). Microsphere aggregation
was prevented by suspending the microspheres in 1% Tween
80 in saline and then shaking the suspension vigorously for 5
min just before injection. One million microspheres were
injected ⬃40 cm proximal to the Gregg cannula to allow
adequate mixing. At the end of the experiment, surface
vessels, fat, and the right ventricular free wall were removed
from the heart, and the left ventricle was opened along the
left anterior descending artery. The left ventricular free wall
and septum were then unrolled, weighed, flattened under a
weight, and placed in 4% Formalin for 5–7 days. Then it was
reweighed and sectioned into four longitudinal segments:
anterior, lateral, posterior, and septal; four vertical layers:
basal, upper middle, lower middle, and apical; and three
layers across the wall: subendocardial, midwall, and subepicardial. Each of the 48 pieces was weighed and then the
radioactivity was counted in a 3-in. NaI (Tl) well counter, the
counts were recorded on a 1,024 channel pulse-height analyzer (TM Analytic, Elk Grove, IL), and the contributions of
each radionuclide were separated by a least-squares method
(7). Flow (ml 䡠 min⫺1 䡠 100 g⫺1) was calculated as
Qt ⫽
Q LMCA ⫻ 共C t /C total兲
Wt
where Qt is flow in the tissue sample (ml 䡠 min⫺1 䡠 100 g⫺1);
QLMCA is the flowmeter determined flow in the LMCA
(ml/min); Ct is the count for a particular radionuclide in the
tissue sample (counts/min); Ctotal is the total radioactivity of
that radionuclide in the heart (counts/min); and Wt is the
weight of the tissue sample (g).
We weighed pieces of the left ventricles of two dogs after
cardioplegia and four dogs that did not receive cardioplegia.
The ventricle pieces were then dehydrated in a desiccator for
48 h until the dry weight was constant.
Experimental Protocol
Pressure-flow relations. All pressure-flow (PF) relations
were determined during a steady-state period. With Pzf inside and around the heart, PF relations were determined at
several coronary arterial pressures in descending order by
lowering the reservoir pressure 1–2 mmHg every 2–4 min
until steady flows were reached. Pzf levels were determined
by lowering coronary arterial pressures until no flow was
observed, and then the coronary arterial pressures were
raised and lowered again to obtain a second Pzf measurement.
Regional flows. In 31 hearts, regional flows were determined by nine radioactive microspheres at selected coronary
arterial perfusion pressures during one of four preload conditions: zero in and around the heart, LVP 30, RVP 30, or
SHP 30 mmHg. Coronary arterial perfusion pressures were
changed in 5-mmHg steps, and recordings were made after
pressures and flows had been stable for 5–10 min; these
included pressures lowered until Pzf ceased. The reservoirs
that controlled pressures in the ventricles were raised or
lowered slowly to the desired pressures. Pressure around the
heart was changed by admitting air into the box and recorded
from a strain gauge in the box. At selected coronary arterial
pressures, microspheres were injected to determine regional
flows. Because we had only nine types of microsphere labels,
various combinations of coronary arterial pressures with
pressures in or around the heart were used in different
hearts. In four hearts, comparisons were made in the same
heart of different preloads (0, LVP ⫽ 30, and SHP ⫽ 30
mmHg) at three different low coronary arterial pressures. In
other hearts, pressures were raised in either ventricle or
around the heart in increments of 5 or 10 mmHg up to a
maximum of 30 mmHg, and regional PF relations were
determined at different perfusion pressures.
Constant coronary arterial perfusing pressures and flows.
Because the heart was quiescent and the vessels were maximally dilated, increases in preloads decreased flows, often
markedly. For fixed resistances, the upstream and downstream pressures will vary with flows, so we determined
what would happen to arterial and venous pressures with
changes in preloads at constant pressures or flows.
Constant coronary arterial perfusing pressures. In one
heart we raised LVP in 5-mmHg increments from 0 to 30
mmHg while we kept coronary arterial perfusion pressure
constant at each of the five pressures. In the same heart we
repeated the measurements at four different coronary arterial perfusion pressures while raising surrounding heart
pressure by 5-mmHg increments from 0 to 30 mmHg. In
another heart we changed LVP and then RVP at four different coronary arterial perfusion pressures. In a third heart we
changed RVP at a fixed coronary arterial perfusion pressure.
Constant flows. In five hearts we changed the preparation
to keep flows constant at three or four rates: 25, 50, 75, and
100 ml/min. In three other hearts we kept flows constant at
one low rate (28.5–33 ml/min) and at one high rate (80–87
ml/min). In two hearts we varied RVP, in four hearts we
varied the surrounding heart pressure, and in five hearts we
varied the LVP; all of these pressures were changed in
5-mmHg increments from 0 to 30 mmHg. For any intervention, the other two pressures were kept at zero. In three of
these hearts, two different pressures were varied one at a
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
cardiac vein. The Gregg cannula was then advanced just
beyond the orifice of the left main coronary artery (LMCA)
and secured in place with a silk suture. Adenosine infusion at
a rate of 20 ␮g 䡠 kg body wt⫺1 䡠 min⫺1 was begun. After the
first few experiments, to ensure maximal vasodilatation, we
added 40 mg/kg chromonar hydrochloride to the reservoir
and then doubled the rate of adenosine infusion until flow did
not increase further. All subsequent experiments were done
with this combined vasodilator protocol. LMCA pressure was
measured at the tip of the Gregg cannula via an internally
placed stainless steel tube. All pressure lines were connected
to Statham D23 Db pressure transducers (Spectra-Med, Critical Care Division, Oxnard, CA), and zero level was set at the
coronary sinus ostium. LMCA flow was measured by an
in-line electromagnetic flow transducer (Howell Instruments,
Camarillo, CA) attached to a Narcomatic flowmeter. The
transducer was placed just proximal to the Gregg cannula,
and Pzf was checked frequently by transient occlusion of the
circuit. To avoid perfusion disturbances when checking Pzf,
the flow transducer was in one arm of a bypass; during flow
recording, the other arm of the bypass was occluded, and
during Pzf testing, the bypass was opened and the tube
clamped proximal to the flow transducer. The flowmeter was
calibrated by timed collection of blood in a measuring cylinder. Once the perfusion had been established, the isolated
perfused heart was transferred to an airtight Plexiglas box.
It was suspended from the lid of the box with an aortic cross
clamp, and the various tubes were brought out through ports.
We were careful to make sure that the box remained airtight.
Figure 1 shows the preparation.
H2929
H2930
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
time. Two other hearts, in which changes in the left heart
and surrounding pressure were studied, had abnormally
high coronary vascular resistances and were excluded from
further analysis.
Ventricular volumes. In two hearts, we measured left ventricular volume during changes of LVP, SHP, or both simultaneously at the end of the experiment. The left ventricular
volume was measured by adding saline from a reservoir
through the left ventricular tube. Volume changes due to
altering pressures in or around the heart were then measured from the height of fluid in the cannula.
Coronary vascular resistances. These were calculated by
dividing the flow rate (ml/min) into the pressure difference
(mmHg) between coronary arterial pressure and pressure in
the coronary sinus or great cardiac vein. This difference is
also called the driving pressure.
Regression analysis was done with Statview version 4.5
and SuperANOVA version 1.11. Statistical significance was
set at 0.05. We used linear and quadratic regression analyses
and analysis of covariance.
RESULTS
Global PF Relations
Figure 2 shows typical PF relations in the arrested
heart with maximally dilated vessels when pressures
are zero in and around the ventricles and when pressures are raised to 30 mmHg in the right ventricle (Fig.
2, top) and around the heart or in the left ventricle (Fig.
2, bottom). When pressures were raised, the PF curves
showed parallel shifts to the right, the change being
always more marked when pressure was raised outside
the heart. Intermediate increases in pressure had intermediate effects. The effect of raising these pressures
on Pzf is shown in Table 1; duplicate determinations of
Pzf agreed closely. A 30-mmHg increase in pressure
around the heart increased Pzf from a mean of 8.6 to
41.3 mmHg, the same increase in the left ventricle
increased Pzf from a mean of 8.6 to 23.8 mmHg, and the
same increase in the right ventricle increased Pzf from
a mean of 8.6 to 32.5 mmHg. Pzf varied in the control
states, that is, when no external pressures were applied, and the differences appeared to be related to the
different flow rates.
Venous Pressures
Pressures in the great coronary vein or coronary
sinus behaved identically and will be referred to hereafter as coronary venous pressures. The effect of increasing pressures in or around the heart on coronary
venous pressures is complex. On the one hand, venous
pressures increased, but in addition coronary arterial
pressures increased, coronary flows decreased, or both
of these occurred simultaneously, and the final consequences for venous pressures depend on interactions
between all these events. To clarify the effects, we
conducted experiments with both controlled coronary
arterial pressures and controlled flows.
Fig. 2. Top: representative pressure-flow (PF) curves made with 0
pressures in and around the heart (0/0/0) and with a pressure of 30
mmHg in the RV (0/30/0). Bottom: representative PF curves made
with 0 pressures in and around the heart (0/0/0), a pressure of 30
mmHg around the heart (0/0/30), and a pressure of 30 mmHg in the
LV (30/0/0). Increases of pressure above 0 produce an almost parallel
shift to the right. L, LV pressure; R, RV pressure; S, surrounding
heart pressure.
Controlled Coronary Arterial Pressure
We studied controlled coronary pressure in three
hearts (Fig. 3). As expected, the higher coronary arterial pressures were associated with the higher flows.
With Pzf in and around the heart, coronary venous
pressures were 5–10 mmHg higher at the highest than
the lowest coronary arterial pressures (and flows). As
pressures in or around the heart were raised in
5-mmHg increments, flows decreased at all coronary
arterial pressures, the rate of decrease being greatest
for the highest flows. When RVP was increased in
5-mmHg increments, there was initially no or little
increase in coronary sinus pressure in one of two
hearts. At higher RVP, the increase in coronary sinus
pressure became greater until in both hearts it increased
with a slope of 1 to reach the same pressure as that in the
right ventricle. Increments in surrounding heart pressure tended to be accompanied by similar increments in
pressure in the great cardiac vein. When LVP was raised
from 0 to 5 mmHg, coronary venous pressure decreased
slightly; after that it increased steadily up to LVP of 30
mmHg. The rate of increase was similar at different
coronary arterial pressures, but the maximal pressure
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
Analysis of Results
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
Table 1. Pzf under different conditions
Statistic
Control
Plv30
Prv30
Psh30
n
Maximum
75%
Mean
Median
25%
Minimum
18
18.0
9.0
8.6
8.0
7.5
1.0
12
38.5
29.8
24.4
23.8
16.5
14.0
2
35.0
35.0
32.5
32.5
30.0
30.0
9
50.0
49.0
41.3
41.0
37.5
26.0
Values are means ⫾ SE; n ⫽ number of dogs. Plv30, left ventricular pressure of 30 mmHg; Prv30, right ventricular pressure of 30
mmHg; Psh30, surrounding heart pressure of 30 mmHg; and Pzf, zero
flow pressure.
Controlled Flow Rates
Changing RVP. In two hearts, RVP was increased
from 0 to 30 mmHg at each of four different constant
flow rates between 25 and 100 ml/min (Fig. 4, left).
When RVP was 0, the coronary venous pressures were
8 and 13 mmHg at a flow rate of 25 ml/min and
increased to 11 and 32 mmHg, respectively, at a flow
rate of 100 ml/min. When RVP was increased from 5 to
15 mmHg, the pressure in the great cardiac vein rose
⬍4 mmHg, and then at higher RVP the venous pressure of the venous pressure-imposed pressure relation
increased at a slope near 1 in six of eight studies; the
RVP at which venous pressure increased seemed to be
greater at the higher flow rates. By contrast, all of the
coronary arterial pressures increased almost linearly
as imposed pressures increased, with slopes that were
independent of flow rates. All of the coronary arterial
slopes were ⬍1, significantly so in six of eight.
The pressure differences from the coronary artery
to the great cardiac vein were relatively constant,
with a tendency to be highest at higher RVP.
Changes in coronary vascular resistance showed
similar patterns and, as expected, decreased with
increasing flow rates.
Changing surrounding heart pressure. In four
hearts, as surrounding pressures were increased from
0 to 30 mmHg at different flow rates, there were
parallel increases in venous pressures with no or minimal interaction between flow rates and pressures (Fig.
4, center). Six of the slopes of the venous pressureimposed pressure relations or their differences were
linear, and three slopes had a significant but minimal
Fig. 3. Coronary inflow flow rates (top) and pressure in the coronary sinus or great cardiac vein (bottom) with
perfusion pressures fixed at different levels, as denoted by the symbols. Numbers indicate the individual hearts.
In each panel, flows and pressures are recorded at 5-mmHg increments in RV pressure (Prv), surrounding heart
pressure (Psh), or LV pressure (Plv). Increases of pressure in or around the heart decrease coronary flows and
increase coronary venous pressures. When pressures are increased to 30 mmHg around the heart or in the RV, the
venous pressures consistently reach 30 mmHg. When pressure in the LV is increased to 30 mmHg, the venous
pressure consistently reaches only 20 mmHg.
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
achieved in the coronary sinus was never greater than
⬃60% of that in the left ventricle, even at the highest
flows and coronary arterial pressures.
H2931
H2932
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
quadratic component. None of the nine individual
slopes was significantly different from 1.
Coronary arterial pressures rose in parallel at each
flow rate as SHP were increased at constant flow rates,
and there was no interaction between flow and slope.
Quadratic components were small, usually positive,
and significant in only three of nine studies. Three of
the nine slopes of the arterial-imposed pressure relations were slightly but significantly ⬎1.
Pressure differences between the coronary artery
and vein were almost constant at all surrounding heart
pressures for any given flow rate. Coronary vascular
resistances tended to increase slightly but not significantly at the higher surrounding heart pressures.
Changing LVP. In six studies in five hearts, as LVP
was increased from 0 to 30 mmHg, cardiac vein pressures increased roughly linearly, with average venous
pressure-imposed pressure slopes well below 1 (Fig. 4,
right)
Coronary arterial pressures increased roughly linearly with increasing LVP; the slopes of the coronary
arterial-imposed pressure relations were all significantly below 1. The pressure differences from the coronary artery to vein tended to remain constant. In the
control state, the resistance was lowest at the highest
flows and approximately doubled at the lowest flows.
The steady-state vascular resistances at each flow rate
remained constant or increased slightly as pressures
were increased in the left ventricle.
Ventricular volumes. Changing LVP in one heart
from 10 to 30 mmHg increased left ventricular volume
from 20 to 65 ml. The SHP was increased incrementally in that heart and one other to 25 and 30 mmHg
when LVP was high (28 and 30 mmHg, respec-
tively); there was an incremental decrease in ventricular volume to near control values (Fig. 5).
Regional PF Relations
In four hearts we examined the effects of raising LVP
or SHP on flow to the left ventricular free wall (Fig.
6A). Driving pressure (coronary arterial minus venous
pressure) was set at 30, 20, and 10 mmHg. Without any
pressure in or around the heart, total coronary flow
(measured in ml 䡠 min⫺1 䡠 g⫺1) decreased from 1.5 to 0.85
to 0.4 as driving pressure decreased. When LVP was
raised to 30 mmHg, the flows at each driving pressure
were ⬃40–50% of control flows, and almost identical
Fig. 5. LV volume as Plv was increased in 5-mmHg increments
(dotted line) or as SHP was increased in 5-mmHg increments after
the ventricle had been distended by pressures of 30 mmHg (F) or 28
mmHg (䊐). Note the large changes in volume caused by these imposed pressures.
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
Fig. 4. Representative examples of changes in coronary arterial and venous pressures and their differences at
various fixed perfusion rates, as indicated by the symbols in the key, at 5-mmHg increments in pressure in the RV
(left), around the heart (middle), and in the LV (right). Top shows the coronary arterial (inflow) pressures with solid
symbols connected by solid lines, and the coronary venous (outflow) pressures with open symbols connected by the
dashed lines. Bottom shows the arteriovenous pressure differences at each flow rate. Numbers indicate the
individual hearts. Increases in pressure in or around the ventricles increase both coronary arterial and venous
pressures in approximately linear fashion to similar extents, so that the arteriovenous differences remain almost
constant.
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
H2933
Fig. 6. A: coronary flow at driving pressures of
10, 20, or 30 mmHg in control state (solid bars),
when Plv was 30 mmHg (open bars), and when
Psh was 30 mmHg (hatched bars). B: inner-toouter flow ratio per gram at driving pressures of
10, 20, or 30 mmHg in control state (solid bars),
when LVP was 30 mmHg (open bars), and when
SHP was 30 mmHg (hatched bars). n ⫽ 4 for
each measurement. Vertical lines represent
standard deviations. Increasing pressures in the
LV or around the heart decreased coronary flows
and inner-to-outer flow ratios per gram to similar extents.
each layer. The heterogeneity of flow in these hearts
when pressures in and around the heart were zero has
already been reported (1). In another heart, at driving
pressures of 30, 20, and 10 mmHg, we examined local
flows per minute per gram when all pressures in and
around the heart were zero, and again when LVP was
raised to 30 mmHg. As expected, flows were heterogeneous in all layers. At 30 mmHg driving pressure,
mean flows were higher in the subendocardium than
the subepicardium when LVP was zero (Fig. 8) and
decreased to similar low flow rates as driving pressure
was reduced to 10 mmHg. When LVP was raised to 30
mmHg at a driving pressure of 30 mmHg, flows decreased more in the subendocardium than in the subepicardium. The effect of raising LVP was almost
equivalent to decreasing the driving pressure by ⬎10
mmHg. At a driving pressure of 20 mmHg, not only
were the flow rates lower in both layers, but also the
degree of heterogeneity was reduced. All pieces of muscle, however, had detectable flows. Only at a driving
pressure of 10 mmHg was flow indistinguishable from
zero in all the subendocardial pieces and in most but
not all of the subepicardial pieces. The dry weight of
the left ventricle was unaffected by cardioplegia.
DISCUSSION
Fig. 7. Representative examples of flow (ml 䡠 min⫺1 䡠 100 g⫺1) in subepicardium (epi) and subendocardium (endo) for 10-mmHg increments in LV pressure (left) or around the heart (right). For both
imposed pressures, changes were more marked in the subendocardium.
We have confirmed the findings of others (28, 35)
that changes of pressure in or around the arrested
heart with maximally dilated vessels shift the PF
curves to the right and increase Pzf. The relationship of
the changes in Pzf to the changes in applied pressure
was also confirmed. We also confirmed the relative
decrease in subendocardial blood flow when, in hearts
with maximally dilated vessels, left ventricular diastolic pressure was increased (17). The new findings in
our study are as follows. First, increases in pressure in
the ventricles or around the heart are accompanied by
increases in coronary venous pressure. The increases
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
values were found when surrounding pressure was
raised to 30 mmHg. When in the same hearts the flows
were examined in the three layers of the left ventricular free wall, with no pressures loading the heart, the
inner:outer flow ratio per gram was just below 1.5 at
driving pressures of 30 and 20 mmHg and 1.0 at a
driving pressure of 10 mmHg (Fig. 6B). When LVP was
raised to 30 mmHg, the inner-to-outer ratio decreased
to 1, 0.6, and 0.5, respectively. When surrounding
heart pressure was increased to 30 mmHg, the innerto-outer ratios were ⬃0.8, 0.7, and 0.6, respectively.
These changes in regional flows were due mainly to
changes in subendocardial flows. If driving pressure
was held at 20 mmHg and LVP or SHP was raised to
10, 20, and 30 mmHg, subepicardial flow changed little, but subendocardial flow decreased profoundly
(Fig. 7).
We also examined the heterogeneity of flow in the six
hearts in which flows were measured in 16 pieces in
H2934
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
in venous pressure matched the increases in pressure
applied around the heart or in the right ventricle. The
increase in venous pressure when pressure was raised
in the left ventricle was about two-thirds of the applied
pressure. Second, Pzf is not uniform across the heart
wall or in any given layer of the left ventricular wall. It
tends to be higher in the subendocardial than the
subepicardial muscle layer. Third, increased pressures
both inside the left ventricle and outside the heart
cause a decrease in myocardial blood flow that is most
marked in the subendocardium. Fourth, there were
greater differences between the arterial than the venous pressures as flows changed. These differences
probably reflected the greater resistance to flow in the
microvascular bed than in the larger extramural veins.
Critique of Methods
We found this technique difficult. Several experiments had to be discarded because of leaks in the box
or the heart or angulation of a cannula. Occasionally,
the cardioplegia failed, and the heart began to beat. If
this happened, we gave more cardioplegic solution and
repeated the experiment. The data from two hearts
were discarded because the coronary vascular resistances were abnormally high; we found no specific
reasons for these abnormalities. Even in the acceptable
experiments, there were large differences in coronary
vascular resistance that could not be explained entirely
by differences in left ventricular mass. One possibility
is that the length of the muscle fibers was altered by
prior distention (strain softening) (18) so that measurements made without prior distention could differ from
those made after the heart had been distended. Nevertheless, the pattern of response from heart to heart
was consistent despite the changes in basal vascular
resistance.
The arrested heart is not an ideal model for a beating
heart in several respects. We doubt whether the effects
of raising pressures around the heart would be much
affected by beating. On the other hand, ventricular
contraction might well affect coronary venous drainage
and certainly increases coronary venous pressure because coronary venous flow is largely systolic (2, 23).
Nevertheless, in at least one study (28), the effect of a
raised left ventricular diastolic pressure in the beating
heart was quantitatively and qualitatively similar to
that reported in the present study. We note too that
these hearts had maximally dilated vessels, so that
vascular tone and metabolic vascular regulatory effects
did not vary as they would in most states in the beating
heart. On the other hand, loss of coronary flow reserve
in a beating heart will produce loss of vascular tone
and metabolic regulation (12), so that our findings
might be applicable to such regions. Despite its disadvantages, working on a heart arrested by cardioplegia
has the advantage of prolonged diastolic arrest without
ischemia. This allowed us to make careful repeated
measurements and to attempt to separate confounding
mechanisms.
Effects on Venous Pressures
Increasing pressures around the heart above atmospheric compresses the extramural coronary veins that
will then act as vascular waterfalls (21, 29, 31) and
regulate venous pressures at the level of the pressure
outside them. The fact that coronary venous pressures
rose linearly and to the same extent with pressures
around the heart explains the increase in Pzf, the
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
Fig. 8. Column graph shows the numbers of pieces (vertical y-axis) with each flow rate/gram/minute (x-axis) for LV
pressures of 0 or 30 mmHg and driving pressures across the coronary circulation of 30, 20, and 10 mmHg; these
pressures are shown on the z- axis. Thus 30/20 indicates that driving pressure was 20 mmHg and LV pressure was
30 mmHg. Data taken from 1 heart. For discussion, see text. A: subepicardial layer. B: subendocardial layer.
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
compresses them between the myocardium and the
connective tissue that surrounds the vessels. The initial negligible response of venous pressure up to LVP of
10 mmHg is probably related to the normal vascular
waterfall pressure in those veins, but after this pressure they become collapsible vessels with a vascular
waterfall dependent on the surrounding pressures. The
fact that at higher ventricular pressures the venous
pressure is about one- to two-thirds of that in the
ventricle may be due to dissipation of some of the LVP
across the wall of the ventricle, especially at the subepicardium. The effects of raising LVP on venous pressures were similar to those on Pzf. Other investigators
have found similar effects on Pzf when preload was
increased. Ellis and Klocke (17) found that, after abolishing coronary autoregulation, raising left atrial pressure from 6 to 20 mmHg increased Pzf from 12 to 19
mmHg. The same group later used a capacitance-free
method to assess Pzf in long diastoles in a vasodilated
beating heart; increasing preload from 6 to 10 mmHg
to 31–35 mmHg increased Pzf from 14 to 23 mmHg (6).
Duncker et al. (15) observed that the higher Pzf in dogs
with left ventricular hypertrophy (24 vs. 12 mmHg in
controls) was not only associated with a higher left
ventricular end-diastolic pressure (14 vs. 6 mmHg) but
also noted that increasing left ventricular diastolic
pressure from 6 to 16 mmHg in the control dogs by
infusing blood increased Pzf in them from 14 to 23
mmHg. Finally, the same group showed that the
higher Pzf in dogs with chronic left ventricular hypertrophy (26 vs. 13 mmHg in controls) was associated
with a greater left ventricular diastolic pressure (22
mmHg in hypertrophy vs. 9 mmHg in controls) (16).
With exercise, Pzf went from 26 to 41 mmHg in dogs
with hypertrophy, and in controls Pzf went from 13 to
24 mmHg. At the same time, left ventricular enddiastolic pressure rose with exercise from 22 to 39
mmHg with hypertrophy and from 9 to 14 mmHg in
controls. In both the studies with left ventricular hypertrophy, the increments of Pzf equaled or exceeded
the increments in left ventricular diastolic pressure;
whether this is a specific feature of hypertrophy is not
known. Thus the relationship of increasing left ventricular diastolic pressure to Pzf appears to be similar in
beating and nonbeating hearts. Whether the relationship is achieved by an increased coronary venous pressure in beating hearts remains to be determined.
Some caution is needed in interpreting these results
as due entirely to the interaction between left ventricular distension and coronary venous pressures.
Duncker et al. (14) observed in exercising dogs with
maximal coronary vasodilatation that, although left
ventricular end-diastolic pressure increased from 5.5
to 10.6 mmHg, measured Pzf increased from 12.6 to
23.3 mmHg, that is, by more than the increase in
end-diastolic pressure. When the exercise-mediated increments in heart rate, contractility, and ␣-adrenergic
vasoconstriction were prevented, the increases in Pzf
and end-diastolic pressure were then equivalent.
Whether the effects of heart rate, contractility, and
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
parallel shift of the PF curves, the decrease in coronary
flows at fixed perfusing pressures, and the increases in
coronary arterial pressures at fixed coronary flows.
Increasing RVP should have no effect on coronary
venous pressures until the RVP exceeds the normal
vascular waterfall pressures of ⬃10–12 mmHg (31).
After that pressure, the venous pressure and the RVP
increased at the same rate with a slope of 1 in two
studies with fixed flows and one study with fixed perfusion pressures. At the highest RVP of 30 mmHg,
coronary venous pressures were equal to or close to 30
mmHg. Pzf also increased to a similar extent at the
highest RVP.
How does an increased RVP affect coronary venous
pressures? Stretching the right ventricle will stretch
the veins on its surface and even impede direct septal
venous drainage into the right ventricle but is unlikely
to have direct mechanical effects on veins over the left
ventricle. These veins constitute most of the left coronary arterial venous drainage, and they and some right
ventricular venous drainage return blood via the coronary sinus to the right atrium. In the experiments done
by Mafuyama et al. (28, 35), the tube inserted into the
right ventricle had multiple holes in it and was tied
into the right atrium at its entry. Therefore, changes in
right ventricular and atrial pressures were identical,
and the shift of Pzf was probably due to increases in
coronary venous pressures due to the raised right
atrial pressures once venous waterfall pressures had
been exceeded. We did not think that this mechanism
explained our results, because in all our experiments
the right atrium was undistended because of the gravity drainage. Had right atrial pressures increased, distension of the right atrium would have been obvious.
Furthermore, the maintenance of different steady RVP
from the reservoir required the reservoir-right ventricular system to be a closed system. Had there been
significant tricuspid incompetence, blood would have
leaked out into the right atrium and the oxygenator,
thereby depleting the reservoir and preventing the
RVP from remaining steady. These changes were not
seen. Therefore, we considered other possible mechanisms for the coronary venous pressure changes that
followed an increase in RVP. Direct pressure on or
obstruction of the ostium and the first part of the
coronary sinus seems unlikely because there is no
anatomic connection between the coronary sinus and
the right ventricle. There is no doubt that there would
have been massive septal shift, but whether this could
have changed coronary pressures is uncertain. We are
left without a clear explanation of the rise in coronary
venous pressure unless it was due to an unappreciated
and unexplained rise in right atrial pressure.
Increasing LVP to 5 or 10 mmHg had no or little
effect on coronary venous pressures or even decreased
them slightly; at higher LVP the coronary venous pressures rose roughly linearly but achieved only ⬃33–84%
of the pressures in the ventricle. We believe that the
mechanism of the rise in coronary venous pressure is
related to the distention of the left ventricle that
stretches the veins on the surface of the heart and
H2935
H2936
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
Effects on Arterial Pressures
Coronary arterial pressures needed to maintain a
constant flow increased linearly or almost linearly with
all changes of pressures in or around the heart. When
pressures were increased in the right ventricle, coronary arterial pressures rose substantially even when
venous pressures had initially risen very little. The
initial rise in coronary arterial pressure was therefore
not due to the change in venous pressure but probably
reflects an increase in septal vascular resistance as the
right ventricular volume and pressure increased.
When pressures were increased in the left ventricle, a
similar finding was noted in some dogs, and perhaps in
these the increased coronary arterial pressure was due
to compression of the left ventricular subendocardium.
Slopes of Coronary Arterial Pressure
Imposed pressure relations were above 1 and were
greater than the slopes of the venous pressures when
pressures were changed around the heart. Increasing
pressures around the heart decreases left ventricular
volume and probably increases coronary vascular resistance, so that coronary arterial pressures were affected both by the raised venous pressures and the
increased vascular resistance. For increased RVP, the
coronary arterial slopes were below 1, even when at the
higher pressures the venous slopes were 1. This probably reflects the effects of venous distention in distending microvessels and reducing coronary vascular resistance without a counterbalancing increase in resistance in the left ventricular free wall. For increased
LVP, the coronary arterial slopes were well below 1
and usually greater than the venous pressure slopes.
We wonder whether this reflects a balance between
lowering vascular resistance by venous and ventricular
distention (remembering that the empty ventricle is
abnormally small) and raising vascular resistance by
compression of the wall, particularly the subendocardium.
Effects on Regional Flows
As expected in a maximally vasodilated coronary
vascular bed, subendocardial flows were higher than
subepicardial flows until driving pressures were very
low. Increasing LVP at a fixed coronary arterial pressure had little effect on subepicardial flows but reduced
subendocardial flows in proportion to the increase in
LVP. We interpret this to mean that the increased LVP
or the accompanying distention of the left ventricle
compressed or stretched (or both) subendocardial more
than subepicardial vessels. Inasmuch as high LVP
increased the volume of the left ventricle considerably,
the stretching of the subepicardial vessels had little
effect on the flow through them; this speaks for radial
compression of the subendocardial vessels as the main
reason for the decreased flows in the inner layer, or
else for a marked difference in the length of the subepicardial and subendocardial vessels in response to
increase in ventricular volume.
We were surprised to find similar effects after increases of surrounding heart pressure; we had hypothesized that inwardly directed radial compression
would have had its greatest effect on subepicardial
vessels. The greater effect on subendocardial vessels
might have been due to the considerable decrease in
left ventricular volume caused by an increase in pressure around the heart. Such a decrease in volume
would severely distort subendocardial vessels and increase the resistance to flow through them. Utley et al.
(33) reached a similar conclusion. Four studies by others of experimental pericardial tamponade are consistent with our results in showing a decreased inner-toouter flow ratio during tamponade (13, 25, 30, 36). In
one of these studies, tamponade was shown to cause
diffuse, punctate subendocardial hemorrhages (36). An
additional study demonstrated decreased systolic coronary flow during tamponade (22); this is compatible
with decreased subendocardial contractility due to
ischemia. Our data support these studies and are noteworthy in that, in our studies, these changes occurred
despite a maintained driving pressure, which does not
usually occur during pericardial tamponade.
The heterogeneity of flow in pieces within a layer of
the left ventricle agrees with other studies (3, 5, 37),
including those done with an arrested heart (1). The
greater flow in the subendocardium than the subepicardium at low driving pressures has been noted before
in our laboratory (34) and is almost certainly due to the
greater conductance in subendocardial vessels in the
arrested heart (11). Raising LVP decreased flows more
in the subendocardium than in the subepicardium, as
discussed above. At a driving pressure of 10 mmHg and
no pressure in the left ventricle, all pieces of muscle in
all layers had flow, albeit small, so that Pzf levels had
to be below 10 mmHg, as we found. With a high LVP,
however, all the subendocardial pieces had flows close
to zero and not distinguishable from it by the microsphere method; Austin et al. (4) concluded that the
confidence limits for the measurement of flows by the
microsphere method were ⫾0.10 ml/min. However, two
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
␣-adrenergic stimulation on Pzf are mediated through
coronary venous pressure is unknown.
Although the heart was arrested and the vessels
were maximally dilated, at zero pressures in and
around the heart vascular resistances were lowest at
the highest flow rates and about twice as high at the
lowest flow rates. This indicates that the increased
flows and accompanying higher perfusing pressures
dilated the resistance vessels passively, as previously
described (19). Interestingly enough, the arteriovenous
pressure difference and therefore the coronary vascular resistance tended to be constant or even to increase
at any given flow as pressures were raised in or around
the heart, even though the increases in coronary arterial and venous pressures would be expected to cause
resistance to decrease (19). This finding suggests that
the dilating effect of raised vascular pressures was
opposed by increased vascular resistances, probably
mainly subendocardial, related to compression of the
left ventricular wall or, for RVP, the septum.
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
This work was supported in part by National Heart, Lung, and
Blood Institute (NHLBI) Program Project Grant HL-25847. G. S.
Aldea received NHLBI Training Grant HL-07192, H. Mori was
supported by the American Heart Association, California Affiliate,
and R. H. Austin was a Stanley J. Sarnoff Fellow.
REFERENCES
1. Aldea GS, Austin RE, Flynn AE, Coggins DL, Husseini W,
and Hoffman JIE. Heterogeneous delivery of cardioplegic solution in the absence of coronary artery disease. J Thorac Cardiovasc Surg 99: 345–353, 1990.
2. Armour JA and Klassen GA. Pressure and flow in epicardial
coronary veins of the dog heart: response to positive inotropism.
Can J Physiol Pharmacol 62: 38–48, 1984.
3. Austin, RE Jr, Aldea GS, Coggins DL, Flynn AE, and Hoffman JIE. Profound spatial heterogeneity of coronary reserve.
Discordance between patterns of resting and maximal myocardial blood flow. Circ Res 67: 319–331, 1990.
4. Austin, RE Jr, Hauck WW, Aldea GS, Flynn AE, Coggins
DL, and Hoffman JIE. Quantitating error in blood flow measurements with radioactive microsphere. Am J Physiol Heart
Circ Physiol 257: H280–H288, 1989.
5. Austin, RE Jr, Smedira NG, Squiers TM, and Hoffman JIE.
Influence of cardiac contraction and coronary vasomotor tone on
regional myocardial blood flow. Am J Physiol Heart Circ Physiol
266: H2542–H2553, 1994.
6. Aversano T, Klocke FJ, Mates RE, and Canty JM Jr. Preload-induced alterations in capacitance-free diastolic pressureflow relationships. Am J Physiol Heart Circ Physiol 246: H410–
H417, 1984.
7. Baer RW, Payne BD, Verrier ED, Vlahakes GJ, Molodowitch D, Uhlig PN, and Hoffman JIE. Increased number of
myocardial blood flow measurements with radionuclide-labeled
microspheres. Am J Physiol Heart Circ Physiol 246: H418–H434,
1984.
8. Bellamy RF. Comments on interpretation and physiological
significance of diastolic coronary artery pressure-flow relationships in the canine coronary bed. Circ Res 57: 485–486, 1985.
9. Bellamy RF. Diastolic coronary pressure-flow relations in the
dog. Circ Res 43: 92–101, 1978.
10. Buckberg GD, Fixler DE, Archie JP, and Hoffman JIE.
Experimental subendocardial ischemia in dogs with normal coronary arteries. Circ Res 30: 67–81, 1972.
11. Chilian WM. Microvascular pressures and resistances in the
left ventricular subepicardium and subendocardium. Circ Res
69: 561–570, 1991.
12. Coggins DL, Flynn AE, Austin RE Jr, Aldea GS, Muehrcke
D, Goto M, and Hoffman JIE. Nonuniform loss of regional flow
reserve during myocardial ischemia in dogs. Circ Res 67: 253–
264, 1990.
13. Cohen MV. Myocardial blood flow in dogs with coronary occlusion: effects of isoproterenol. Am Heart J 109: 1032–1038, 1985.
14. Duncker DJ, Van Zon NS, Crampton M, Herrlinger S,
Homans DC, and Bache RJ. Coronary pressure-flow relationship and exercise: contributions of heart rate, contractility, and
␣1-adrenergic tone. Am J Physiol Heart Circ Physiol 266: H795–
H810, 1994.
15. Duncker DJ, Zhang J, and Bache RJ. Coronary pressureflow relation in left ventricular hypertrophy. Importance of
changes in back pressure versus changes in minimum resistance. Circ Res 72: 579–587, 1993.
16. Duncker DJ, Zhang J, Pavek TJ, Crampton MJ, and
Bache RJ. Effect of exercise on coronary pressure-flow relationship in hypertrophied left ventricle. Am J Physiol Heart Circ
Physiol 269: H271–H281, 1995.
17. Ellis A and Klocke FJ. Effects of preload on the transmural
distribution of perfusion and pressure-flow relationships in the
canine coronary vascular bed. Circ Res 46: 68–77, 1980.
18. Emery JL, Omens JH, and McCulloch AD. Biaxial mechanics of the passively stretched left ventricle. Am J Physiol Heart
Circ Physiol 272: H2289–H2305, 1997.
19. Hanley FL, Messina LM, Grattan MT, and Hoffman JIE.
The effect of coronary inflow pressure on coronary vascular
resistance in the isolated dog heart. Circ Res 54: 760–772, 1984.
20. Hoffman JIE. Transmural myocardial perfusion. Prog Cardiovasc Dis 29: 429–464, 1987.
21. Hoffman JIE and Spaan JAE. Pressure-flow relations in coronary circulation. Physiol Rev 70: 331–390, 1990.
22. Jarmakani JMM, McHale PA, and Greenfield JC Jr. The
effect of cardiac tamponade on coronary hemodynamics in the
awake dog. Cardiovasc Res 9: 112–117, 1975.
23. Klassen GA and Armour JA. Canine coronary venous pressures: responses to positive inotropism and vasodilation. Can
J Physiol Pharmacol 61: 213–221, 1983.
24. Ledingham SJM, Braimbridge MV, and Hearse DJ. Improved myocardial protection by oxygenation of St. Thomas’
Hospital cardioplegic solutions: studies in the rat. J Thorac
Cardiovasc Surg 95: 103–111, 1988.
25. Martins JB, Manuel WJ, Marcus ML, and Kerber RE.
Comparative effects of catecholamines in cardiac tamponade:
experimental and clinical studies. Am J Cardiol 46: 59–66, 1980.
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
pieces of the subepicardium had flows above this
threshold. Hence, in this heart, zero flow at the main
coronary arterial level would not have been detected at
a driving pressure of 10 mmHg and a LVP of 30 mmHg.
These results suggest that there is probably a distribution of Pzf across the left ventricular wall and that
the concept of Pzf applies to only a small portion of the
heart rather than being a function of global flow (see
Bellamy, Refs. 8 and 9). Rouleau and White (26) raised
coronary sinus pressure to 18–28 mmHg at fixed inflow
pressures and observed that flow was decreased in the
subepicardium but not in the subendocardium. They
concluded that this finding indicated that the high
venous pressures exceeded intramyocardial pressures
only in the subepicardium and that an intramyocardial
vascular waterfall with a high surrounding pressure
regulated subendocardial flow. A more direct demonstration of this was provided by Satoh et al. (27), who
found that as LVP was raised in the arrested heart, Pzf
increased and was similar to subepicardial intramyocardial pressure that was lower than that in the subendocardium. Moreover, there are differences of Pzf
even within a layer. It is likely that the curvature
observed frequently at the lower end of the PF curve
(21) is in part due to sequential drop out of small
regions within a layer as well as between layers. These
differing regional effects suggest that Pzf cannot be a
function only of extracardiac venous waterfalls but is
related also to intramyocardial pressure or different
regional microvascular time constants (21).
Our study has some important clinical implications.
Clinically, a dilated heart with an elevated left ventricular end-diastolic pressure has been associated with a
vicious cycle of further dilatation and failure. Our
studies suggest that elevation of transmural pressure
produces a disproportionate reduction in subendocardial blood flow, even if coronary arterial pressures are
maintained, and could lead to subendocardial ischemia. This may be one of the reasons for the subendocardial ischemia reported in congestive cardiomyopathy (32). Furthermore, if cardioplegia is done in a
patient with a high left ventricular diastolic pressure,
there may be inadequate delivery of cardioplegic solution and thus inadequate myocardial protection in the
subendocardium, even at normal perfusing pressures.
H2937
H2938
VENTRICULAR TRANSMURAL PRESSURE AND CORONARY BLOOD FLOW
32. Unverferth DV, Magorien RD, Lewis RP, and Leier CV.
The role of subendocardial ischemia in perpetuating myocardial
failure in patients with nonischemic congestive cardiomyopathy.
Am Heart J 105: 176–179, 1983.
33. Utley JR, Michalsky GB, Mobin-Uddin K, and Bryant LR.
Subendocardial vascular distortion at small ventricular volumes. J Surg Res 17: 114–24, 1974.
34. Verrier ED, Baer RW, Hickey RF, Vlahakes GJ, and Hoffman JIE. Transmural pressure-flow relations during diastole in
the canine left ventricle (Abstract). Circulation 3: 62, 1980.
35. Watanabe J, Maruyama Y, Satoh S, Keitoku M, and Takishima T. Effects of the pericardium on the diastolic left coronary
pressure-flow relationship in the isolated dog heart. Circulation
75: 670–675, 1987.
36. Wechsler AS, Auerbach BJ, Graham TC, and Sabiston DC.
Distribution of intramyocardial blood flow during pericardial
tamponade. J Thorac Cardiovasc Surg 68: 847–856, 1974.
37. Yipintsoi T, Dobbs WA Jr, Scanlon PD, Knopp TJ, and
Bassingthwaighte JB. Regional distribution of diffusible tracers and carbonized microspheres in the left ventricle of isolated
dog hearts. Circ Res 33: 573–587, 1973.
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.1 on May 15, 2017
26. Rouleau JR and White M. Effects of coronary sinus pressure
elevation on coronary blood flow distribution in dogs with normal
preload. Can J Physiol Pharmacol 63: 787–797, 1985.
27. Satoh S, Maruyama Y, Watanabe J, Keitoku M, Hangai K,
and Takishima T. Comparison of zero-flow pressure (Pf ⫽ 0) in
diastolic coronary pressure-flow relationship with diastolic pressures in outer (IMPo) and inner (IMPi) myocardial layers. Circulation 76: 489, 1987.
28. Satoh S, Watanabe J, Keitoku M, Itoh N, Maruyama Y, and
Takishima T. Influences of pressure surrounding the heart and
intracardiac pressure on the diastolic coronary pressure-flow
relation in excised canine hearts. Circ Res 63: 788–797, 1988.
29. Scharf SM, Bromberger-Barnea B, and Permutt S. Distribution of coronary venous flow. J Appl Physiol 30: 657–662,
1971.
30. Taichman GC, Byrne P, Forester GV, and Keon WJ. Regional myocardial blood flow during pericardial tamponade. Can
J Physiol Pharmacol 62: 539–543, 1983.
31. Uhlig PN, Baer RW, Vlahakes GJ, Hanley FL, Messina LM,
and Hoffman JIE. Arterial and venous coronary pressure-flow
relations in anesthetized dogs. Evidence for a vascular waterfall
in epicardial coronary veins. Circ Res 55: 238–248, 1984.