152
Lung Fluid Exchange after Uneven
Pulmonary Artery Obstruction in Sheep
KAZUHIRO OHKUDA, KAZUYA NAKAHARA, W. JEFFREY WEIDNER, ANDREW BINDER,
AND NORMAN C. STAUB
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
SUMMARY We studied steady state transvascular fluid and protein exchange after uneven obstruction of the pulmonary arteries. In anesthetized sheep, ventilated by positive pressure, we measured
pulmonary artery and left atrial pressures, cardiac output, lung lymph flow, and lymph/plasma protein
ratios. We calculated pulmonary vascular resistance. In 16 sheep we obstructed the pulmonary arteries
with various microemboli or balloons in lobar arteries until pulmonary vascular resistance was 2-3
times baseline. In every experiment, lung lymph flow increased as pulmonary vascular resistance
increased. The lymph/plasma protein ratio did not change. In four sheep, we increased left atrial
pressure by balloon obstruction of the mitral orifice. Pulmonary artery pressure increased as much as
in the embolization experiments and lymph flow increased but the lymph/plasma protein ratio decreased, meaning that the increased fluid and protein flux after embolization cannot be due to high
pulmonary artery or pulmonary venous outflow pressures alone. In four sheep we compared obstruction of the lower lobe pulmonary arteries by balloons with that of upper lobe pulmonary arteries.
Lower lobe arterial obstruction caused the lymph flow which drains predominantly from the lower
lobes to decrease whereas upper lobe artery obstruction increased lymph flow. This means that the
increased fluid and protein flux occurred mainly in the open, perfused portion of the microvascular
bed. The mechanism of the increased fluid filtration and protein permeability may be related to high
vascular pressure and high linear blood flow velocity through a markedly restricted microvascular
bed, although release of substances that affect endothelial permeability is not ruled out.
IN THE ADULT respiratory distress syndrome,1 a central feature is the occurrence of interstitial and alveolar
edema. Since in pathological sections the edema fluid
stains similarly to blood plasma (proteinaceous) and since
there are often hyaline membranes in the air spaces, the
edema is considered to be of the increased microvascular
permeability variety.2 Hemodynamically, features commonly noted are an increase in pulmonary arterial pressure in the face of a normal pulmonary artery wedge
pressure or a directly measured left atrial pressure. The
calculated pulmonary vascular resistance is elevated. Numerous investigators3"5 have frequently found plugging of
pulmonary small arteries and microvessels by various
types of microemboli.
How is it that obstruction of pulmonary arteries causes
pulmonary edema due to increased microvascular permeability?
From the Cardiovascular Research Institute and Department of Physiology, University of California, San Francisco, San Francisco, California.
This work was supported in part by U. S. Public Health Service
Program Project Grant HL06285, National Heart, Blood and Lung Institute Pulmonary SCOR HL14201, and National Heart, Blood and Lung
Institute Pulmonary Vascular Disease SCOR HL19155. Dr. Weidner is
supported by Training Grant HL05251. Dr. Binder is supported by Training Grant HL05251.
Dr. Ohkuda's present address is: Department of Thoracic Surgery,
Research Institute for Chest Diseases and Cancer, Tohoku University,
Sendai, Japan. Dr. Nakahara's present address is: First Department of
Surgery, Osaka University Medical School, Fukushima-ku, Osaka, Japan.
Dr. Weidner's present address is: Department of Animal Physiology,
University of California, Davis, Davis, California.
Address for reprints: Dr. Norman C. Staub, Cardiovascular Research
Institute, University of California, San Francisco, San Francisco, California 94143.
Received August 1, 1977; accepted for publication March 6, 1978.
Gibbon and Gibbon6 made one of the earliest observations on this syndrome. They found that resection of 70%
of the lung mass in cats was followed by the development
of pulmonary edema, especially if the resection was accompanied by infusions of whole blood or plasma. Although they believed the edema was due to the high
pressure in the microvessels, they did suggest that the
edema fluid was only slightly diluted plasma (that is,
typical of what we call increased permeability edema).2
Visscher7 briefly reported on the effects of microemboli
in the isolated dog lung perfused at constant flow. When
he injected approximately one billion glass microspheres
(25 fim diameter), pulmonary artery pressure rose markedly and edema developed rapidly. He believed that the
increased pulmonary artery pressure was transmitted into
the microvessels and, together with the increased velocity
of blood flow in the patent portions of the microcirculation, caused the edema.
Even pulmonary macroemboli may be associated with
acute pulmonary edema if the initial obstruction is not so
severe as to cause acute cor pulmonale. Parmley and
associates8 reported pulmonary edema as the immediate
cause of death in 8 of 30 dogs given autologous blood clot
emboli.
Hultgren and Grover9 developed the concept of overperfusion as the mechanism of embolic and high altitude
pulmonary edema. In open-thorax dogs, they removed
80% of the lung mass and regularly produced acute pulmonary edema. The total protein concentration of the
edema fluid averaged 4.8 g/dl (Hultgren, personal communication). Although he did not measure the plasma
protein concentration, based on the usual values in labo-
LUNG FLUID EXCHANGE AFTER VASCULAR OBSTRUCTION/OMuda et al.
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
ratory dogs, the edema fluid protein must have been
about 80% of the plasma concentration.
Viswanathan et al.10 hypothesized that uneven pulmonary arterial constriction in acute alveolar hypoxia with
resultant overperfusion of some areas of the lung may be
the underlying mechanism accounting for high altitude
pulmonary edema.
Arfors and co-workers" have measured pulmonary vascular resistance following fibrin microemboli in dogs.
Although they did not relate the degree of edema directly
to the amount of obstruction, their data show a correlation between these two variables. Saldeen12 attributed
the edema following microemboli principally to the biochemical action of fibrin metabolic products acting to
increase microvascular permeability.
In the experiments reported here, we have studied the
relationship between pulmonary embolism and lung fluid
and protein exchange using a sheep model13 in which we
can obtain nearly pure lung lymph. We have tried to
answer the following three questions: (1) can acute pulmonary edema reliably be produced by uneven obstruction of the pulmonary arteries? (2) is the edema of the
increased permeability type? (3) do the edema fluid and
protein escape from the occluded portion or the perfused
portion of the lung?
Methods
We prepared yearling female sheep (weight, 25-58 kg;
average, 38 kg) in order to collect pure lung lymph and to
measure pulmonary hemodynamics as previously described.13' 14 We did our experiments in anesthetized sheep
rather than in chronically instrumented, unanesthetized
ones, for convenience and also because we were uncertain
as to the animals' response.
At preliminary surgery, we resected the distal portion
of the caudal mediastinal lymph node to remove essentially all of its systemic lymph input. We placed a catheter
in the left atrium to record pressure and, in some sheep,
a balloon catheter in the left atrium for later use in
obstructing the mitral valve orifice in a controlled fashion.14
On the day of study, we anesthetized the sheep with
intravenous sodium pentobarbital (30 mg/kg), tracheally
intubated them, and ventilated them with 0.5-1.5% fluothane in 50% oxygen, using a positive pressure ventilator
(Ventimeter, Air Shields Co.) set at 15 cycles/min to
deliver a tidal volume of approximately 600 ml at a peak
airway pressure of 30 cm H2O. Through a right thoracotomy (5th interspace) we cannulated the efferent duct of
the caudal mediastinal lymph node with a small silastic
catheter.13 We did not close the thorax but continued
with the experiment. We placed the sheep in the supine
position.
Under fluoroscopic control, we passed a double lumen
catheter with a thermister bead near the tip via the right
external jugular vein into the main pulmonary artery (7F,
Kimray Medical Assoc). We also passed a large polyvinyl
catheter via a systemic artery into the aorta. We attached
the aortic, left atrial, and distal lumen of the pulmonary
arterial catheters of strain gauges (Statham P23G) to
record pressures on a direct-writing polygraph (model 7,
153
Grass Instruments). Pressures were measured relative to
atmospheric pressure and to the base of the lung.
Through the proximal lumen of the thermister catheter, which was in the right atrium, we injected 5 ml of
iced saline and determined cardiac output by the thermal
dilution method, using a cardiac output computer (model
2800, Kimray).
Experimental Procedure
After the surgical preparation was completed, we studied each sheep for a baseline period of at least 2 hours,
after which we made a variety of interventions. There
were 28 sheep in the entire study; four in each of seven
groups. The first group consisted of our experimental
controls. In these, we followed lung lymph flow and
hemodynamics for 6-8 hours. We did not obstruct the
pulmonary circulation.
In 16 sheep, we obstructed the pulmonary arterial
circulation in one of four different ways. In four sheep we
made mineral oil emboli (200 ± 10 /an in diameter)
prepared by the method of Sackner and associates.15 In
four sheep we used silicone-coated glass beads (215 ± 15
lim in diameter) (glass type IV A, Unispheres, Ferro
Corp.). In four we gave an intravenous infusion of tranexamic acid to inhibit fibrinolysis (150 mg/kg iv initial bolus
followed by 15 mg/kg per min continuous iv infusion) and
thrombin (5000 U.S. N.I.H. units intravenously as a bolus
followed by approximately 3000 U iv in boluses every 15
minutes until the maximum increase in pulmonary vascular resistance was achieved) (Park Davis) in order to
form fibrin microemboli in the pulmonary circulation."
In four sheep, we passed small balloon catheters via
the external jugular vein to obstruct lobar pulmonary
arteries. We obstructed the entire left main pulmonary
artery as well as the right upper and right middle lobe
pulmonary arteries. Right ventricular output then was
forced through the right lower and cardiac lobes which
together account for approximately 30% of lung mass in
the sheep. We will call this group, "upper lobe obstruction," even though the left lower lobe was included. We
want to differentiate this group from another one which
we will call "lower lobe obstruction." This is important
because the lymph in the caudal mediastinal efferent
duct, which we cannulated, comes predominantly from
the lower lobes of the lung, as indicated in Figure 1. It
was necessary in this macroballoon obstruction to leave
as much as possible of the lung from which lymph flow
came, open for blood flow.
The five groups of four sheep each described above
form the primary experimental study. The experiments
were done in random fashion. The following two groups
of experiments were done subsequently.
In four sheep, we inflated a balloon in the left atrium
to raise pulmonary artery pressure passively to approximately the same value as was achieved during the arterial
obstruction experiments.
In the seventh and last group of four sheep, we passed
balloon catheters via the external jugular veins into the
lower lobe pulmonary arteries and selectively occluded
these. This is the "lower lobe obstruction" group, as
previously mentioned (Fig. 1). The purpose of this group
CIRCULATION RESEARCH
154
VOL. 43, No.
2, AUGUST
1978
TrocheoBronchiol
Lymph Nodes
Upper
Lobe
Pulmonary
Artery
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Lower
Lobe
Caudal Mediastinal
Lymph Node
FIGURE 1 Schema of lung lymph drainage in the sheep. The upper (anterior) lobe drains predominantly via the lung hilum and
tracheobronchial lymph nodes into both the right lymph duct and thoracic duct. The lower (posterior) lobes drain predominantly
over the pleural surface, through the pulmonary ligament to the large caudal mediastinal lymph node. The lymph cannula, through
which we measured lung fluid flux, was in the main efferent duct of the caudal mediastinal lymph node.
was to determine whether the increased fluid exchange
occurred in the open or occluded portions of the pulmonary vascular bed.
In all of these experiments, after the baseline period,
there was a transition period of up to 2 hours during
which new steady state conditions were established. For
example, in the pulmonary arterial obstruction experiments, we embolized the lungs at intervals until we had
achieved increases in calculated pulmonary vascular resistance of 2-3 times baseline. We then followed the
parameters of lung fluid exchange over a period of approximately 2 hours of relatively stable lymph flow, which
we defined as a lymph flow whose variance does not
exceed that of the baseline period.
Terminally, all sheep were killed as follows. We opened
the thorax widely, extending the original thoracotomy
incision across the sternum. We isolated and clamped the
right and left lung hila separately. We removed the lungs,
took samples for histology, and homogenized the remainder in order to determine extravascular lung water.16
Statistics
We compared the experimental data between the baseline and steady state experimental periods in each sheep
by the paired <-test. The summary data in the text and
tables give the average ± 1 standard deviation of the
sample of each group of four animals. We made comparisons between experimental groups by an unpaired <-test
and comparison of the slopes of regression lines by the Ftest.17 We accepted P < 0.05 as indicating statistical
significance.
Results
General
Within each group, the responses of the four sheep
were qualitatively similar. All animals remained fully
anesthetized and in relatively stable condition throughout
the experiments. Arterial oxygen tension was usually
above 200 torr during the baseline period. Following
embolization, arterial oxygen tension sometimes decreased to less than 100 torr. Although we did not measure lung mechanics, total pulmonary resistance tended
to increase after embolization as judged by an increase in
the inflation pressure at constant tidal volume. Published
data concerning microemboli indicate that their main
effect is to induce alveolar duct constriction within the
terminal respiratory unit, thereby decreasing lung compliance.1819
Lung lymph was generally free of red blood cells during
the baseline period. Sometimes it became slightly sanguinous after embolization. The lymph hematocrit, however, never exceeded 2%.
At postmortem examination, we did not find any significant accumulation of pleural fluid and the lungs were
not grossly edematous. Histologically, we saw only small
LUNG FLUID EXCHANGE AFTER VASCULAR OBSTRUCTION/OMwria et al.
perivascular or peribronchial fluid cuffs. There was no
alveolar edema.
155
the lymph-to-plasma protein ratio, did not change significantly. Not listed in the table are the data for the
lymph/plasma ratios of the subtractions of albumin and
globulin which averaged 0.77 ± 0.05 and 0.53 ± 0.05,
respectively, in the baseline periods. Neither of these
changed significantly after embolization. The lung water
contents in these animals were increased approximately
10% above the control value.
Control Experiments
The time course of one control experiment is shown in
Figure 2A. The summary data for pulmonary hemodynamics and lung fluid balance for the baseline period and
for the last 2 hours of the "experimental" period are
shown in Table 1A. There were no significant differences
between the baseline and later observation periods. The
lung water content was in the normal range reported by
our laboratory.14'20
Upper Lobe Pulmonary Artery Balloon
Obstruction
Figure 3 shows the time course of one experiment in
which we obstructed the pulmonary arteries to the entire
left lung and the right upper and middle lobes with
balloon catheters, leaving only the right lower and cardiac
lobe perfused. The summary data are in Table 1C. In
every sheep, the pulmonary artery pressure and vascular
resistance increased. Lung lymph flow increased significantly but the lymph/plasma protein ratios did not
change. In these sheep, we measured postmortem water
content of the perfused right lower and cardiac lobes
separately from those that had been obstructed. In every
sheep, the perfused lobes averaged 10% more water than
the obstructed portions of lung; the same difference as
found between the control and microembolized animals.
Microemboli
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Examples of the time course of the microemboli experiments are shown in Figure 2B (mineral oil), 2C (glass
beads), and 2D (fibrin). The summary data for the baseline and 2-hour steady state periods after obstruction are
in Table IB. Pulmonary artery pressure and calculated
pulmonary vascular resistance were increased in all embolization experiments. The rise following the fibrin microemboli was less than with the other two types. Lung
lymph flow increased in every experiment after emboli
whereas the lymph protein concentration, expressed as
Emboli
i.o
J
Protein
Concentration 0.5J
(lymph/plasma) Q|
Albumin
—l_»-
Protein
Concentration
10
Albumin
|lymph/plasmo|
Globulin
6
Pulmonary
Vascular
Resistance
Pulmonary
Vascular
Resistance
l/min
Lung Lymph
Flow
4
("vmTr) 0
20
5
S39-75
I
4
Hours
Lung Lymph
Flow
10
B
Hours
Emboli
Protein
Protein
Concentration
(lymph /plasma]
Concentration
(lymph/plasma)
Lung Lymph
Flow 20
Lung Lymph
Hours
FIGURE 2 Time course of lung lymph flow, calculated pulmonary vascular resistance, and lymph protein concentration expressed
as lymph/plasma ratio for albumin and globulin, in one control sheep (A) and three sheep with different types ofemboli; mineral oil
(B), silicone-coated glass microspheres (C), and fibrin (D).
CIRCULATION RESEARCH
156
VOL. 43, No.
2, AUGUST
1978
TABLE 1 Effects of Uneven Lung Vascular Obstruction on Pulmonary Hemodynamics and Lung Fluid Balance in
Anesthetized Sheep
Pulmonary hemodynamios
N
Condition
0.
A. Controls
Baseline*
Experimental
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
B. Microemboli
1. Mineral oil
Baseline
Experimental
2. Glass beads
Baseline
Experimental
3. Fibrin
Baseline
Experimental
C. Balloon occlusion
Baseline
Experimental
Cardiac
output
(liters/min)
Lung fluid balance
Calculated
resistance
[cm H2O/
(liters/min)]
Lymph flow
(ml/hr)
Lymph protein concentration
(lymph/
plasma)
Lung water
content
(g/g dry
lung)
5.1 ±0.9
6.3±0.6
0.73±0.03
0.68±0.02
4.2±0.1
Pulmonary artery
pressure
(cm H20)
Left atrial
pressure
(cm H20)
25.3±2.1f
27.7±2.2
14.9+1.0
15.0±0.9
3.7+0.6
4.1 ±0.4
3.7±0.3
3.5±0.4
26.5±0.7
41.1±2.1(s)t
16.8±0.6
16.9±0.6
3.8±0.4
3.6+0.3
2.6±0.3
7.6±0.5(s)
5.0±0.5
15.2±1.0(s)
0.64+0.02
0.62±0.02
4.6±0.2
25.6+2.3
44.9+2. l(s)
13.6±1.4
13.7±2.0
4.2±0.6
3.5±0.3
2.8±0.6
9.6±0.8(s)
7.6±1.4
20.5±2.1(s)
0.64±0.03
0.69±0.02
4.6±0.1
24.7±1.2
33.8±1.3(s)
15.0±0.6
14.4±0.5
3.8±0.5
3.1±0.2
2.6±0.3
5.7+0.5(s)
4.8±0.8
7.9±1.2(s)
0.61 ±0.05
0.63±0.03
4.8±0.2
24.9±1.0
33.0±1.3(s)
15.9±1.1
13.7±0.6
3.4±0.4
3.0+0.3
2.8+0.4
6.9±0.4(s)
5.7±0.4
9.3±0.7(s)
0.72±0.02
0.70±0.03
4.0±0.2
4.4±0.3§
4
4
4
4
4
* 2-hr baseline period and 2-hr steady state period after obstruction.
f Average ± standard deviation. Pressures measured relative to atmospheric pressure and the base of the lung.
| (s) = statistically significant at P < 0.05 by paired /-test between baseline and experimental periods.
§ Lobes with arteries occluded are listed above the perfused lobes.
Increased Left Atrial Pressure
Figure 4A shows the time course of one of these experiments. Qualitatively, it is not different from our previous
experiments in unanesthetized animals.14 Nevertheless, it
seemed necessary to test the effects of pulmonary venous
Stort
Occlusion
Albumin
Protein
Concentration
i
Globulin
Lower Lobe Pulmonary Artery Balloon
Obstruction
(lymph/plosmo)
Pulmonary
Vascular
Resistance
/cm
H;0\
\ l/min
hypertension under the conditions of our experiments.
We raised left atrial pressure until the pulmonary artery
pressure was in the same range as in the microemboli
and balloon obstruction experiments. The data are summarized in Table 2A. In every experiment, lung lymph
flow increased but, what is more important, the
lymph/protein ratio decreased. This is different from all
of the pulmonary artery obstruction experiments (Table
1).
/
Lung Lymph
Flow
(ml/h)
S62-76
Hours
FIGURE 3 Time course of uneven pulmonary uascular obstruction obtained by inflating small balloons in lobar pulmonary
arteries. This is referred to as upper lobe artery obstruction
because that is the predominant occlusion, although the left
lower lobe was also blocked. The grey overlay labeled "start
occlusion" is the period during which the balloon catheters
were placed and inflated. Once placed and inflated, the balloons remained so for the duration of the experiment which
explains why pulmonary vascular resistance remained elevated.
We selectively occluded the pulmonary arteries to the
right and left lower lobes and the cardiac lobe with
balloon catheters in four sheep. Under these conditions,
the portion of lung from which most of the caudal mediastinal lymph drainage comes was obstructed.
The time course of one of these experiments is shown
in Figure 4B and the data are summarized in Table 2B.
The protocol of these experiments was slightly different
in that there is a recovery period after the period of
occlusion. In each animal, lung lymph flow decreased
during occlusion in spite of the rise in pulmonary artery
pressure and pulmonary vascular resistance. The pressure
in the pulmonary artery in the occluded lobes was approximately left atrial pressure.
Lung water content is for the entire lung 2 hours after
the occlusion had been released. That is not very informative and may be a protocol error on our part. Nevertheless, we felt that it was more important to have good
before-and-after lung lymph flows in this group than to
have the extravascular lung water content at the end of
the obstruction period.
LUNG FLUID EXCHANGE AFTER VASCULAR OBSTRUCTION/OMutfa et al.
Increose Left
Atrial Pressure
v
Protein
Concentration
Albumin
response is not due to increased pressure alone (Table
2A).
We conclude that in terms of lung fluid exchange there
are no differences among the four types of uneven pulmonary artery obstruction.
A remarkable but perhaps not obvious fact is that lung
lymph flow increased, although much of the lung's microvasculature had been occluded; that is, the surface area
for fluid exchange must have decreased markedly. In the
first stages of embolization, recruitment of additional
microvessels occurs and pulmonary vascular resistance
does not rise very much. When microvascular recruitment
is nearly complete, however, resistance increases rapidly
with further obstruction. The pressure-flow relation is
very nearly linear so that changes in resistance are a
reasonable index of the changing fraction of the microvascular cross section that is obstructed. Thus, a 2-fold
rise in pulmonary vascular resistance means that 50% of
the microvascular cross-section has been obstructed and
50% of fluid exchange area has been sequestered. Likewise, a 3-fold increase in calculated resistance means a
67% decrease in the available exchange area. If there is
too much reduction in surface area, lymph flow may
decrease, as is clearly demonstrated by our experiments
with the selective lower lobe occlusions (Table 2B).
t
(lymph/plosmo)
Pulmonary
Vascular
Resistance
/cm H 2 O \
\l/min
/
Lung Lymph
Flow
(ml/h)
Hours
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Occlusion
Protein
Concentration
(lymph /plasma)
Albumin
1.0
0.5 ~ Globulin
/ c m H20 \
\l/min /
;
•
- -
010 -
Pulmonary
Vascular
Resistance
i
—n_n_.
5-
0
15r-
Lung Lymph
Flow
(ml/h)
10
5
0
B
S30-77
i
i
I
|
157
I
i
Hours
FIGUHE 4 A: Time course of pulmonary venous congestion.
Rising pulmonary vascular pressures increased lung lymph
flow but decreased lymph protein concentration. B: Time course
of lower lobe pulmonary artery balloon obstruction. The regions
of lung from which the caudal mediastinal lymph emanates
were selectively occluded.
Discussion
General
The control sheep (Fig. 2 and Table 1A) establish the
stability of our experimental model over a time course in
excess of 8 hours.
The hemodynamic and fluid balance data for all four
types of uneven pulmonary artery obstruction were similar (Table IB). Figure 5 is a graphic demonstration of
the relationship between pulmonary vascular resistance
and lung lymph flow. There are no significant differences
among the slopes of lines. We correlated the change in
lymph flow to resistance rather than to the pulmonary
artery pressure to emphasize that microvascular flow rate
(cardiac output) may be an important factor, as Gibbon,6
Visscher,7 and Hultgren9 have suggested, and that the
Site of Increased Fluid and Protein Exchange
Although the obvious site of the altered fluid exchange
would be in the open, perfused microvessels, three alternate possibilities ought to be considered, namely, fluid
leakage (1) directly through the thin walls of pulmonary
arteries proximal to the embolization sites; (2) at the site
of embolic impaction due to local injury and (3) in the
microvessels beyond the sites of occlusion due to accumulation of chemically active substances.
1. The pressure increased in all pulmonary arteries
proximal to the occlusion sites. Whayne and Severinghaus,21 Iliff,22 and Butler and associates23 have evidence
that fluid leakage may occur in the arteriolar portion of
the lung's microvessels, that is, proximal to the main
alveolar wall capillary network. In their various experiments, vessels only slightly larger than 15-20 jum in diameter (alveolar wall corner vessels) were suspected. Iliff,
using Evans blue dye, has the most direct evidence for
the size of vessels leaking. On the arterial side, she detected possible leakage only from vessels <75 jum in
diameter.
Since our mineral oil and glass bead microemboli were
200 (im in diameter, such small pulmonary arteries could
not have been involved except for those in the perfused
portions of lung.
Vessels larger than 200 ftm in diameter have not been
seriously considered in terms of transarterial leakage and,
if they were, the internal evidence from our lobar artery
balloon occlusion experiments would tend to rule that out
because the response is not different from the microemboli occlusions, although the surface area available for
transarterial leakage is 3-fold different. Even firmer evidence that transarterial leakage is not occurring in vessels
remote from alveolar walls (<200 ;um) is our recent investigation of acute and chronic alveolar hypoxia in sheep.24
CIRCULATION RESEARCH
158
VOL. 43, No.
2, AUGUST
1978
TABLE 2 Effects of Increased Left Atrial Pressure and of Bilateral Lower Lung Lobe Artery Obstruction on
Pulmonary Hemodynamics and Lung Fluid Balance in Anesthetized Sheep
Pulmonary hemodynamics
No
Condition
A. Increased left
atrial pressure
Baseline*
Experimental
4
B. Lower lobe occlusion
Baseline
Experimental
Recovery
4
Pulmonary artery pressure
(cm H2O)
Left atrial
pressure
(cm H2O)
Cardiac output
(liters/min)
Lung fluid balance
Calculated
resistance
[cm H 2 0/
liters/min)]
24.8+0.7+,
38.7±2.1(s)t
15.6+0.8
33.8±2.2(s)
2.6+0.2
1.8±0.1(s)
3.7±0.3
3.3±0.6
25.2±1.6
34.2±2.1(s)
23.9±1.1
15.7±0.9
15.4±0.4
15.9±0.7
2.6±0.4
2.1±0.2
2.5±0.2
3.6±0.4
9.5±0.9(s)
3.1+0.3
Lymph flow
(ml/hr)
Lymph protein
concentration
(lymph/plasma)
Lung water content
(g/g dry
lung)
11.4±0.5
22.8±1.8(s)
0.70+0.02
0.58±0.03(s)
6.2±1.6
0.69±0.02
0.72+0.01
0.72±0.01
4.4±0.4
9.5+0.8
6.2±0.4(s)
8.9±0.6
* 2-hr baseline period, 2-hr steady state period after intervention, and 1- to 2-hr recovery.
f Average ± standard deviation. Pressures measured relative to atmospheric pressure and the base of the lung.
% (s) = statistically significant at P < 0.05 by paired (-test between baseline and experimental periods.
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
In those experiments, pulmonary artery pressures were
elevated to levels similar to those in the microemboli
experiments, and yet there was no change in lung lymph
flow or its protein composition. We interpreted the hypoxia-induced pulmonary artery hypertension as due to
active vasoconstriction proximal to the sites of fluid exchange.25 Thus, although pulmonary vascular resistance
was increased, neither microvascular flow rate nor microvascular distending pressure was affected.
2. Injury at the site of embolic impaction is possible.
Following the initial impaction, as pulmonary artery pressure rises, the emboli may be pushed downstream,
thereby exposing the original damaged area. Although we
cannot rule out this possibility, it certainly cannot be the
explanation in those animals with the main lobar pulmonary arteries occluded by balloons.
3. Although the arterial end of the microvessels is
occluded, the leakage could be from the occluded microGlass Beads
2Or
Lung
Lymph
|5
Flow
(ml / h )
10
Mineral Oil
j p Balloon Occlusion
Fibrin
vessels. They are open to pressure from the left atrium or
the bronchial circulation. As noted above, pressure distal
to balloon occlusions was approximately left atrial pressure which, as Table 1 shows, is not insignificant at the
base of the lungs in these supine animals.
Clearly, in the experiments of Gibbon6 and Hultgren,9
the increase in fluid flux had to come from vessels with
blood flow, since the remaining lung had been removed.
Because of the pattern of lymph drainage from the sheep
lung, we have been able to make a direct determination
of the principal fluid exchange sites. As Figure 1 showed,
lymph from the caudal mediastinal efferent duct comes
principally from the lower lobes of the lung. Occluding
the lower lobe arteries with balloons decreased the measured lymph flow. Thus, the leakage could not have come
from the occluded vessels.
It is possible that the primary injury was in the occluded portions of the microvessels but that there was
diffusion of chemically active substances through the
lungs interstitium to other perfused vessels which were
then made permeable. Since the microemboli are distributed throughout the lung, occluded vessels were intermixed with those open to perfusion. The microemboli
were distributed initially in proportion to blood flow
distribution; but, since the lungs were embolized repeatedly, the emboli eventually were distributed in all lobes.
On the other hand, the interstitial diffusion of vasoactive
substances does not fit with our lobar artery balloon
occlusion experiments, nor wit^ previously published
studies involving lung resection.
Nature of the Increased Fluid Exchange
«
5
10
Pulmonary Vascular Resistance f c m H 2 ° )
Vl/min /
FIGURE 5 Relationship between calculated pulmonary vascu-
lar resistance and lung lymph flow for each of the four groups
of sheep with uneven pulmonary artery obstruction. The slopes
of the lines of best fit are not significantly different.
In Figure 6, we have compared the relative protein
clearance for the four uneven arterial obstruction groups
(Table 1) with that of the sheep with increased left atrial
pressure (Table 2A). All of the uneven obstruction experiments show the same increases in protein clearance with
respect to lymphflow,whereas the protein clearance with
increased left atrial pressure is significantly less even
though the pulmonary artery and microvascular pressures generated were as high as the emboli experiments.
The protein clearance is a key differential point be-
LUNG FLUID EXCHANGE AFTER VASCULAR OBSTRUCTION/OMuda et al.
3OOr
Relative
Protein
Flow
(%)
200
Uneven
Pulm. A.
Obstruction
I00 1 100
200
Relotive Lymph Flow
300
(%)
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
FIGURE 6 Comparison of lymph protein flow (clearance), to
lung lymph flow for the entire group of 16 sheep with uneven
pulmonary artery obstruction and the four sheep with increased
left atrial pressure. Although pulmonary artery pressure increased following balloon obstruction of the mitral valve as
much as it did with the uneven pulmonary artery obstruction,
the protein flow from the lung increased only half as much. The
significantly different results indicate that increased microvascular pressures alone cannot account for the phenomena in
uneven pulmonary artery obstruction.
tween edema due to increased hydrostatic pressure and
that due to increased endothelial permeability.2'14'20
We conclude that the increased fluid and protein exchange associated with uneven pulmonary artery obstruction involves elevated microvascular pressures which are
transmitted into the open, perfused regions from the
pulmonary artery. We also conclude that the edema
following pulmonary microembolism is chiefly one of
increased microvascular fluid and protein permeability
and that the changes in permeability cannot be accounted
for solely by the elevated vascular pressure.
What is the Mechanism for the Increased
Permeability?
Is there any mechanism that is common to all of the
experiments which will account for the observed results?
To indicate the direction in which our research is headed,
we offer the following speculative review of two different
but not mutually exclusive possibilities.
Biochemical Factors
Lindquist and associates3'26 have identified a permeability-increasing factor that is released from the lungs of
dogs into right duct lymph in the microembolism syndrome.They have tentatively identified it as a fibrin degradation product (FDP), slowly released from the persistent fibrin microemboli in the pulmonary vessels. The
material increases vascular permeability in the hamster
cheek pouch and in the abdominal skin of the rat. The
fact that FDP is released in the lung is not surprising
since their emboli were predominantly fibrin. However,
altered permeability in the hamster cheek pouch is not
definitive evidence that this material functions as a
permeability increasing factor in the lung. The pulmonary
vasculature is notorious for behaving differently from
systemic blood vessels.
159
On the other hand, our mineral oil and Silastic-coated
glass microsphere emboli are probably not as inert as we
had originally thought. For example, Piper and Vane27
reported the release of prostaglandins from lungs after a
variety of microemboli.
Recently, Malik and Van der Zee28 presented evidence
that heparin can partially block the increase in pulmonary
vascular resistance and completely block the increase in
lung water content and tracer fibrinogen deposition after
glass microemboli in dog lungs. Their results support
Saldeen's view that the fibrinogen system is involved in
the altered permeability.3 The difficulty remains, however, in trying to explain the results of our pulmonary
artery balloon occlusion experiments. If an FDP is released downstream from the occlusion site, how does it
reach the open microvessels in another lobe? One would
have to invoke recirculation to explain it.
Lung mechanics have been shown to be affected by
active substances formed or released after microembolization.18'19 In our sheep, we did not measure lung mechanics specifically but noted that peak inflation pressure
increased after emboli, indicating an increase in total
pulmonary resistance. In addition, arterial oxygen tension
usually decreased, indicating an increased right-to-left
shunt-like effect.
Craddock and associates29 have evidence that complement-activated plasma acts to trap circulating neutrophils in the lung's microcirculation with subsequent increase in microvascular permeability, presumably due to
release of lysosomal enzymes. B0 et al.30 find that platelets
are necessary for at least part of the microembolism
response. Although Saldeen3 tends to discount the role of
white cells and platelets in the increased permeability of
the microembolism syndrome, further evaluations by specific depletion experiments are necessary before their role
can be completely defined.
It is possible that local or circulating vasoactive substances released or generated by the microvascular obstruction (for example, histamine31) contribute to the
changes in vascular permeability.
Physical Factors
An increased driving pressure is not the only physical
factor that has been altered in the uneven obstruction
experiments. The linear velocity of blood is also increased,
since cardiac output remained nearly normal in spite of
the decrease in the vascular cross-section.7 This is different from the increase in pulmonary blood flow seen in
exercise. There the increase in cardiac output (linear
velocity of blood flow) is largely offset by the expansion
of the pulmonary microvascular bed.32 In addition, pulmonary vascular resistance decreases so that the distending pressure in the microvessels is not increased; at least,
that is the interpretation we place on Marshall's33 finding
of no increase in lung water content in exercising dogs.
Unfortunately, measurements of steady state lung lymph
flow during exercise have not been made.
How does an increased blood flow rate alter microvascular permeability? Fry34 and Patel and associates35 have
evidence that local stress on endothelial surfaces, at critical points in the circulation, can lead to rapid and marked
CIRCULATION RESEARCH
160
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
changes in endothelial permeability to proteins and other
molecules. Figure 7 is a hypothetical scheme denoting
some of the "stress" points that may occur in the pulmonary microvascular bed. These are analogous to
stresses described by Fry.34 The critical areas appear to
be at branches, sharp bends, or where there is direct
impaction of blood flow jets.35
Substantial evidence for injury in large blood vessels,
such as the aorta and coronary arteries,36 exists, but there
is no direct evidence for injury at the microvascular level.
In the Appendix we have calculated both circumferential
and sheer stresses within the alveolar wall microvessels
based upon estimates of the necessary parameters. Both
stresses increase with uneven pulmonary artery obstruction. The tangential stress increased 7-fold between the
two conditions but was always less than one-tenth the
values for large arteries.36 This is consistent with our data
that distending pressure alone does not increase the lung's
microvascular permeability.14
The sheer stress increased 4-fold but, more important,
it exceeds values that cause increased endothelial fluid
and protein permeability in large arteries.36
Admittedly speculative, the concept of physical injury
by high velocity and high pressure flow in the pulmonary
microcirculation is able to account for all forms of pulmonary edema following uneven arterial obstruction.
Appendix
Estimates of endothelial stresses in the lung's microcirculation provide a basis for comparison between different
conditions. We cannot place too much reliance upon the
VOL. 43, No.
2, AUGUST
1978
TABLE 3 Estimated Parameters and Calculated
Stresses in the Lung's Microcirculation
Baseline
After glass
bead microcirculation
A. Parameter values*
r (Xl(T4 cm)
h (xl(T4 cm)
L (XlO"2cm)
Ptm (xlO3 dynes/cm2)
Pd (XlO:i dynes/cm2)
0.5
6
18
12
44
31
B. Calculated stress
Tangential (XlOr> dynes/cm2)
Shear (dynes/cm2)
1.4
40
10
180
4
7
0.3
6
' r = 4.3 + 0.22 Ptm (ref. 39). Ptm: baseline = 0.4 (Ppa-Pla) + Pla (ref.
14); after emboli = Ppa. Pd = Ppa-Pla.
absolute numbers because of the complex geometry of
the microvascular bed (Fig. 7) and the many simplifying
assumptions involved.
The two major stresses are (1) the tangential stress
tending to rupture the vessel wall and (2) the shear stress
due to viscous drag along the luminal surface of endothehum.3637
Tangential stress (at) in a cylindrical, thin-walled vessel
is directly proportional to the distending pressure (Ptm),
the radius (r), and inversely proportional to the wall
thickness (h).
ot = r(Ptm)/h
Shear stress (T) varies directly with driving pressure (Pd)
along the vessel, the radius (r), and inversely with vessel
length (L).
T = r(Pd)/2L
Arter iole
To obtain numbers, we have assumed the lung's microcirculation to consist of uniform cylindrical tubes with
characteristic length based on measurements in dogs,
cats, and rabbits,'18 and whose radius varies with distending pressure according to the equation of Sobin and
coworkers.39 In Table 3 are listed the necessary parameters and the calculated values of tangential and shear
stress for two conditions; baseline and after glass bead
microemboli (Tables IB and 2).
References
Venule
FIGURE 7 Schema of the pulmonary microvascular bed to
show where physical injury may affect microvascular permeability. The thin arrows represent blood flow stream lines; the
thick blunt arrows indicate points of increased endothelial
stress. [After Fry, D. L., Circulation 39/40 (suppl. IV): 38-59,
1969.]
1. Blaisdell FW, Schlobohm RM: The respiratory distress syndrome: A
review. Surgery 74: 251-262, 1973
2. Staub NC: Pulmonary edema due to increased microvascular permeability to fluid and protein. Circ Res 43: 143-151, 1978
3. Saldeen T: The occurrence of fibrin in the lungs in an (sic) autopsy
material. Virchows Arch 343: 20-25, 1967
4. Moore FD, Lyons JH Jr, Pierce EC Jr, Morgan AP Jr, Drinker PA,
MacArthur JD, Danmin GJ: Post-Traumatic Pulmonary Insufficiency. Philadelphia, W. B. Saunders Co., 1969
5. Reul GJ, Greenberg SD, Lefrak EA, McCollum WB, Beall AC Jr,
Jordan GL Jr: Prevention of post-traumatic pulmonary insufficiency.
Arch Surg 106: 386-394, 1973
6. Gibbon JH Jr, Gibbon MH: Experimental pulmonary edema following
lobectomy and plasma infusion. Surgery 12: 694-704, 1942
7. Visscher MB: The pathophysiology of lung edema; a physical and
physicochemical problem. Lancet 82: 43-47, 1962
8. Parmley LF Jr, North RL, Ott BS: Hemodynamic alterations of acute
pulmonary thromboembolism. Circ Res 11: 450-465, 1962
9. Hultgren NH, Grover RF: Circulatory adaptation to high altitude.
Annu Rev Med 19: 119-152, 1968
LUNG FLUID EXCHANGE AFTER VASCULAR OBSTRUCTION/OMuda et al.
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
10. Viswanathan R, Jain SK, Subramanian S: Pulmonary edema of high
altitude III. Pathogenesis. Am Rev Resp Dls 100: 342-349, 1969
11. Arfors KE, Busch C, Jakobson S, Lindquist O, Malmberg P, Rammer
L, Saldeen T: Pulmonary insufficiency following intravenous infusion
of thrombin and AMCA (tranexamic acid) in the dog. Acta Chir
Scand 138: 445-452, 1972
12. Saldeen T: The microembolism syndrome. Microvasc Res 11:
227-259, 1976
13. Staub NC, Bland RD, Brigham KL, Demling R, Erdmann AJ III,
Woolverton WC: Preparation of chronic lung lymph fistulas in sheep.
J Surg Res 19: 315-320, 1975
14. Erdmann AJ III, Vaughan TR Jr, Brigham KL, Woolverton WC,
Staub NC: Effect of increased vascular pressure on lung fluid balance
in unanesthetized sheep. Circ Res 37: 271-284, 1975
15. Sackner MA, Feisal KA, Karsch DN: Size of gas exchange vessels in
the lung. J Clin Invest 43: 1847-1855, 1964
16. Selinger SL, Bland RD, Demling RG, Staub NC: Distribution volumes
of '"I-albumin, HC-sucrose and MC1 in sheep lung. J Appl Physiol 39:
773-779, 1975
17. Snedecor GW: Statistical Methods, ed. 5. Ames, Iowa, Iowa State
University Press, 1956
18. Stein M, Hirose T, Yasutake T, Tarabeich A: Airway responses to
pulmonary embolism—pharmacologic aspects. In Pulmonary Thromboembolism, edited by KM Moser, M Stein. Chicago, Yearbook, 1973,
pp 166-177
19. Halmagyi DPJ, Colebatch JHJ: Cardiorespiratory effects of experimental lung embolism. J Clin Invest 40: 1785-1796, 19961
20. Brigham, KL, Woolverton WC, Blake LH, Staub NC: Increased sheep
lung vascular permeability caused by pseudomonas bacteremia. J
Clin Invest 54: 792-804, 1974
21. Whayne TF Jr, Severinghaus JW: Experimental hypoxic pulmonary
edema in the rat. J Appl Physiol 25: 729-732, 1968
22. Iliff LD: Extra-alveolar vessels and edema development in excised dog
lungs. Circ Res 28: 524-532, 1971
23. Butler J, Culver BH, Huseby J, Albert R: The hemodynamics of
pulmonary edema. Am Rev Resp Dis 115 (part 2): 173-180, 1977
24. Kato M, Staub NC: Response of small pulmonary arteries to unilobar
hypoxia and hypercapnia. Circ Res 21: 426-439, 1966
25. Bland RD, Demling RH, Selinger SL, Staub NC: Effects of alveolar
hypoxia on lung fluid and protein transport in unanesthetized sheep.
161
Circ Res 40: 269-274, 1976
26. Lindquist 0, Gerdin B, Saldeen T, Svensjo E: Demonstration of a
permeability increasing factor in interstitial fluid from the lung following thrombin induced pulmonary microembolism in the dog and
the rat (abstr). Microvasc Res 10: 238, 1975
27. Piper PJ, Vane JR: The release of prostaglandins from lung and other
tissues. Ann NY Acad Sci 180: 363-383, 1971
28. Malik AB, van der Zee H: Time course of the pulmonary vascular
response to microembolization (abstr). Fed Proc 36: 535, 1977
29. Craddock PR, Fehr J, Brigham KL, Kronenberg RS, Jacob HS:
Complement and leukocyte-mediated pulmonary dysfunction in hemodialysis. N Engl J Med 296: 769-774, 1977
30. B0 G, Hognestad J, Vaage J: The role of blood platelets in pulmonary
responses to microembolization with barium sulphate. Acta Physiol
Scand 90: 244-251, 1974
31. Brigham KL, Owen PJ: Increased sheep lung vascular permeability
caused by histamine. Circ Res 37: 647-657, 1975
32. Vaughan TR Jr, DeMarino EM, Staub NC: Indicator dilution lung
water and capillary blood volume in prolonged heavy exercise in
normal men. Am Rev Resp Dis 113: 757-762, 1976
33. Marshall BE, Soma LR, Neufeld GR: Lung water volume at rest and
exercise in dogs. J Appl Physiol 39: 7-8, 1975
34. Fry DL: Certain histological and chemical responses of the vascular
interface to acutely induced mechanical stress in the aorta of the dog.
Circ Res 24: 93-108, 1969
35. Patel DJ, Vaishnav RN, Atabek HB, Plowman F: Measurement of
endothelial strength using saline jets. In Microcirculation, vol 1,
edited by J Grayson, W Zingg. New York, Plenum Press, 1976, pp
79-80
36. Carew TE, Patel DJ: Effect of tensile and shear stress on intimal
permeability of the left coronary artery in dogs. Atherosclerosis 18:
179-183, 1973
37. Merrill EW, Cox RH: Quantitative relations pertaining to steady flow.
In Respiration and Circulation, edited by PL Altman, DS Dittmer.
Bethesda, FASEB, 1971, pp 4-7
38. Staub NC, Schultz EL: Pulmonary capillary length in dog, cat and
rabbit. Resp Physiol 5: 371-378, 1968
39. Sobin SS, Fung YC, Tremer HM, Rosenquist TH: Elasticity of the
pulmonary alveolar microvascular sheet in the cat. Circ Res 30:
440-445, 1972
Lung fluid exchange after uneven pulmonary artery obstruction in sheep.
K Ohkuda, K Nakahara, W J Weidner, A Binder and N C Staub
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Circ Res. 1978;43:152-161
doi: 10.1161/01.RES.43.2.152
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1978 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/43/2/152.citation
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/