Departments of Anesthesia, Physiology/Biophysics, Pediatrics, and
Surgery, Indiana University School of Medicine, Indianapolis, Indiana
46202; Department of Physiology, Medical College of Wisconsin,
Milwaukee 53226; Research Service, Zablocki Veterans Affairs Medical
Center, Milwaukee 53295; and Department of Biomedical Engineering,
Marquette University, Milwaukee, Wisconsin 53233
pulmonary microcirculation; arterial occlusion; venous occlusion; double occlusion; video fluorescence microscopy; digital image
analysis; fluorescently labeled red blood cells; vascular resistance; mathematical model; isolated dog lung
 |
INTRODUCTION |
PULMONARY ARTERIAL and venous occlusions have been used
to characterize the longitudinal distribution of pulmonary vascular resistance with respect to compliance between the arterial and venous
sites of occlusion (1-3, 6, 9-12, 18). The data obtained have
been the time-varying arterial and venous pressures after occlusion.
Various models (generally compartmental) have been used to interpret
these data in terms of the distribution of compliance and resistance
(2, 9, 11, 12). The process of determining the model elements needed to
explain the data has provided insights into the mechanics of the
pulmonary vascular bed. However, the identification of the arterial or
venous sites of changes in vascular resistance and compliance and a
determination of the pulmonary capillary pressure have been primary
objectives of many of the studies. Such structure-function correlations
have been based primarily on the observation that vasoconstrictor
stimuli affect the model parameters in ways consistent with their known
arterial or venous sites of action (2, 6, 12, 16). These correlations then provide the basis for interpreting the effects of stimuli having
unknown sites of action. There are also other kinds of indirect
evidence linking model parameters and anatomy, e.g., comparison of the
compartmental nodal pressures with pressures estimated by other methods
(6, 11, 18). The interpretations of these results have varied (2).
We have added the measurement of flow in subpleural arterioles and
venules after vascular occlusion as an additional source of information
linking structure and function. The flows were measured in isolated dog
lung lobes under zone 3 conditions by using video microscopy to track
fluorescently labeled red blood cells in vessels with diameters of
~40 µm. Under the assumption that fractional changes in flow in
these vessels are representative of the fractional changes in flow in
similarly sized vessels throughout the lungs, these arteriolar and
venular flows allow for a direct, i.e., model-independent, estimation
of the compliances of the arteries, capillaries, and veins from the
flow response to venous occlusion. The rationale is that when the
venous outflow is occluded while arterial inflow remains constant, the
axial flow observed at any site between the inlet and the site of
occlusion will be the inflow rate minus the rate of volume storage in
the distensible vessels upstream from the site of observation. Because
the rate of change in pressure at all locations within the system will be equal once the quasi-steady state has been established and because
compliance is the change in volume for a given change in pressure, the
rate of upstream volume storage over total flow is equal to the ratio
of upstream compliance to total compliance. If the total compliance is
estimated by measuring the change in pressure resulting from the total
volume stored in the system up to a given time, the actual compliance
of the segments upstream and downstream from the sites of observation,
in this case ~40-µm subpleural arterioles and venules, can be
calculated.
Using the compliances calculated from the flow data after venous
occlusion and the pressure measured after simultaneous occlusion of
arterial inflow and venous outflow (double-occlusion pressure), we used
a previously described three-capacitor-two-resistor (3C2R) compartmental model representation of the lobar vascular bed (2) (Fig.
1) to calculate the londitudinal
distribution of vascular resistance and determine its relationship to
the distribution of compliance.
View larger version (8K):
[in this window]
[in a new window]
|
Fig. 1.
A 3-capacitor-2-resistor (3C2R) compartmental model of pulmonary
circulation. FT, total flow into
lung; Pa, pulmonary arterial pressure; Pc, pulmonary capillary
pressure; Pv, pulmonary venous pressure;
C1,
C2, and
C3, compliances of arteries,
capillaries, and veins, respectively;
R1 and
R2, resistance of arteries
and veins, respectively, with any
resistance in capillaries contributing equally to
R1 and
R2 (3).
|
|
| |
METHODS |
Experimental preparation.
Adult, male, mongrel dogs (n = 6, 19-25 kg) were sedated with acepromazine (10 mg im), and ~100 ml
of blood were withdrawn from each animal by venipuncture. By use of
techniques described in detail elsewhere (19), the red blood cells were
fluorescently labeled with
DiI16(3) and then reinfused. A
1-day delay was allowed to permit removal of any cells damaged during
the labeling process. On the day after labeling, the animals were
anesthetized by intravenous injection of pentobarbital sodium (30 mg/kg) freshly dissolved in 0.9% saline that was supplemented by 5 mg/kg doses as needed to maintain surgical anesthesia, intubated, and
ventilated with air via a constant-volume respirator (model 607D,
Harvard Apparatus). At that time, flow cytometry showed that 2-3%
of the circulating red blood cells were fluorescently labeled. After
heparinization (1,000 U/kg) the animals were rapidly exsanguinated
through a cannula (3 mm ID) placed in the left common carotid artery.
During the exsanguination, 150 ml of 10% dextran (70 kDa) in saline
were infused (6). With the lungs inflated to a constant airway pressure of ~4 mmHg, the left chest wall and the left upper lobe were excised to provide access to the left lower lobe. The left lower lobar artery
was cannulated with a Teflon FEP cannula (6 mm ID), and the left lower
lobe bronchus was clamped to maintain constant inflation. The left
lower lobe was then excised along with a cuff of left atrium and placed
on a microscope stand. The surface of the lobe was kept moist with
saline throughout the procedure. The left atrial cuff was secured
around another Teflon FEP cannula (10 mm ID), and the lobe was perfused
with autologous, heparinized whole blood (hematocrit 31 ± 3%).
Care was taken to exclude all air bubbles from the circuit before
initiation of perfusion. The time interval from complete exsanguination
to reperfusion of the lobe was ~25 min. Blood was pumped (model
7522-10 pump drive and model 7024-20 pump head, Masterflex)
through a pulse dampener to reduce pump vibrations and trap bubbles, a
filter (20-µm pore size, model 4C2423, Fenwal) to remove
microaggregates, and a heat exchanger (model HE-100, Bentley) to
maintain the blood at 37-38°C (Fig.
2). Venous blood drained from the lobe into
a reservoir. Pump flow rate was set between 200 and 400 ml/min. The
height of the reservoir was adjusted so that the venous pressure (Pv) was ~9 mmHg to maintain the lobes in zone 3 conditions when the airway pressure was 5 mmHg.
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 2.
Schematic of experimental setup. ICCD, intensified charge-coupled
device; A to D, analog-to-digital.
|
|
The lobe was ventilated with 6%
CO2-17%
O2-77%
N2. Blood gases were sampled from
the pulmonary venous line periodically and measured by using an
Instrumentation Laboratories model 1304 analyzer. Sodium bicarbonate
solution (1 meq/ml) was added to the venous reservoir as needed to
neutralize metabolic acid. Average values were arterial
PO2 of 112 ± 2 Torr, arterial
PCO2 of 36 ± 1 Torr, and pH of
7.39 ± 0.01. There were no significant differences between blood
gas measurements sampled at the beginning of the experiment and those
sampled at the end (paired 2-tailed t-test,
P > 0.10). A 100-ml tidal volume
kept peak inspiratory pressure <10 mmHg. Expiratory pressure was set
at 5 mmHg by a water overflow on the expiratory limb of the ventilator.
Arterial pressure (Pa), Pv, and airway pressure were measured from side ports in the respective cannulas (model P23 XL, Statham) zeroed at the
level of the subpleural vessels under observation. The output from the
pressure transducers was amplified (model 13-G4615-52, Gould),
processed by a 50-Hz low-pass analog filter, then sampled at 100 Hz by
an analog-to-digital board (model C10-DA508-PGA, Computer Boards) in a
microcomputer (Dell 50-MHz 486).
Video microscopy.
The lobe was suspended by two small spring-backed paper clips attached
to opposite edges of the lobe (21) and raised until the uppermost
pleural surface (the diaphragmatic surface in this orientation) came
into contact with a transparent window. A
1.3-cm2 area on the surface of the
lobe was observed through the window. The windowpane was surrounded by
a vacuum ring to prevent lateral movement of the observed area (20).
Suspending the lobe against the window in this manner allowed free
downward expansion of the lobe during ventilation and prevented
compression of the subpleural alveoli by the window. The remainder of
the lobar surface was covered with a thin sheet of plastic to prevent
drying and to slow the transpleural diffusion of gas.
The subpleural microcirculation under the window was observed with a
modified Olympus BH2 reflectance microscope coupled to a Leitz Ultropak
illuminator with a ×11 objective. Bright-field illumination was
obtained with a 200-W mercury arc lamp that was filtered with a
combination of dichroic infrared-reflecting filters and broad band-pass
ultraviolet-absorbing filters to prevent tissue damage and a narrow
band-pass interference filter to select the mercury green line (546 nm). Illumination for fluorescence microscopy was provided by a 100-W
mercury arc mounted on the sidearm of the BH2 microscope, which was
also filtered by dichroic infrared-reflecting filters and
ultraviolet-absorbing filters. The light from this arc passed through a
green band-pass exciter filter (500-560 nm) and a high-pass
dichroic mirror (cutoff wavelength = 560 nm), which reflected the
exciting light (mercury green line 546 nm) down through the objective
onto the subpleural microcirculation beneath the window. Emitted light
passed back through the objective, the dichroic mirror, and a red
high-pass barrier filter (cutoff wavelength = 590 nm). Video recordings
of the subpleural microcirculation were made with a Panasonic AG7300
SVHS video recorder and a Cohu intensified charge-coupled device camera
(model 5510), which was mounted on the microscope with a Nikon zoom
CCTV adapter (model 79444).
Vascular occlusions.
Arterial, venous, and double-occlusion maneuvers were performed in
random order in each preparation with the lobe held in end expiration.
Flow was occluded by stopcocks in the arterial inflow and venous
outflow limbs of the perfusion circuit (Fig. 2). The stopcocks were
controlled by actuators (cylinders containing a piston and rod), which
closed the stopcocks when they were pressurized. Depending on the type
of occlusion (arterial, venous, or double), one or both of the
stopcocks were closed. Rotation of the stopcock valves interrupted flow
without displacing volume from the site of occlusion. To perform an
occlusion, a solenoid was energized, which pressurized the actuator(s)
with compressed gas at 60 psi, closing the stopcock(s). The computer
simultaneously marked time 0 in the
pressure data file and activated a Panasonic WJ-810 time-date generator, which recorded elapsed time in milliseconds on the videotape. Arterial inflow to the lobe was constant during venous occlusion. Occlusions were performed in triplicate while a subpleural arteriole and venule observed with fluorescence microscopy were videotaped. The duration of each occlusion was ~3 s. Preocclusion arteriolar and venular diameters were measured from the videotaped images as described previously (13).
For the off-line velocity measurements, the videotapes were replayed,
and red blood cell velocities in subpleural precapillary arterioles and
postcapillary venules were calculated by measuring the distance
individual fluorescently labeled red blood cells moved from one video
frame to the next. For each occlusion and in each vessel, one to three
cells were counted in every video frame from 1 s before occlusion to 2 s after occlusion (90 video frames, 90-270 cells/occlusion). To
reduce noise, a single velocity curve for each vessel for each type of
occlusion was obtained by averaging the velocity curves from each of
the triplicate occlusions for that set of conditions. The curves were
averaged by computing the mean of the triplicate measurements at each
point in time after occlusion. Similarly, a single average pressure
decay curve was obtained for each type of occlusion on each lung lobe
by averaging the pressure curves from the triplicate occlusions. The
measured response of the pressure-measuring system (transducer,
amplifier, filter, and analog-to-digital board) to a square-wave
pressure change from 20 mmHg to atmospheric pressure was >90%
complete within 0.033 s or one video frame.
Model-independent analysis.
The venous occlusion data were used to divide the total lobar vascular
compliance into three compliances: one comprised of the compliance of
arteries >40 µm (Ca), another of vessels <40 µm (Cc), and a
third of veins >40 µm in diameter (Cv). If the vascular bed is
visualized as a longitudinally distributed distensible system, when the
outflow is occluded while the inflow is maintained at a constant rate,
the rate of change in pressure at all locations within the system will
be equal and the axial flow observed at any site between the inlet and
the site of occlusion will be the inflow rate minus the rate of volume
storage in the distensible vessels upstream from the site of
observation once the quasi-steady state has been established. The rate
of upstream volume storage over the total flow will be the ratio of the
upstream compliance to the total compliance. If the total compliance is
estimated by measuring the change in pressure resulting from the total
volume stored in the system up to a given time, the actual compliance of the segments upstream and downstream from the site of observation can be estimated. In the present study, two sites along the
longitudinal distribution were available for observation on the
subpleural surface: the ~40-µm arterioles and venules just upstream
and downstream, respectively, from the capillary bed.
It is actually the volume flows, rather than the measured velocities,
that are needed to estimate the compliance distribution. Because the
vessels are distensible and the pressures are continuously increasing
after venous occlusion, we converted the velocities to flows by using
the following relationships
|
(1)
|
where
v, A, D, and F are the local velocity,
vessel cross-sectional area (arterial or venous), diameter, and flow,
respectively. Dividing the postocclusion
(t > 0) velocity,
v(t),
by the preocclusion velocity, v(0),
gives
|
(2)
|
Over
the pressure range encompassed by the venous occlusion, 8.8-20
mmHg,
D(t)
can be related to D(0) by
|
(3)
|
where
Pav(t) = [Pa(t) + Pv(t)]/2 and 
is the vessel
distensibility, which was previously found for the lung preparation used in the present studies to be ~0.018/mmHg for the subpleural venules and 0.031/mmHg for the arterioles (14).
Substituting Eq. 3 in
Eq. 2 results in the following
relationship between flow, velocity, and pressure
|
(4)
|
This
relationship was used to convert the velocity measurements to flows,
correcting for the pressure changes that occurred after venous,
arterial, or double occlusion.
Because we could only measure the diameter of and the velocity in some
individual vessels of the entire parallel set of ~40-µm vessels, it
was necessary to calculate fractional flows and, from them, fractional
compliances. For each experiment, the flows between 0.5 and 2 s after occlusion were averaged, 
, and
divided by the preocclusion flow rates averaged from 
0.5 to 0.0 s, F(0), to obtain the normalized steady-state postocclusion
(t > 0) flows through the arteriole
and venule [a/Fa(0) and
v/Fv(0), respectively]. The total
compliance (CT) was estimated
as
|
(5)
|
where
S is the slope of
Pv(t) obtained by linear regression
over the time interval 0.5-1.0 s after venous occlusion and
FT is the lobar inflow rate.
Before venous occlusion, the flow through 40-µm arterioles (Fa) and
venules (Fv) is equal to the total lobar flow rate
|
(6)
|
After
venous occlusion, after the quasi-steady state is reached, the flow
rate through the arterioles is equal to the total flow rate minus the
volume stored in upstream vessels
|
(7)
|
Solving
Eq. 5 for
FT and substituting for
FT in Eq. 7 gives
|
(8)
|
Similarly,
the flow rate through the venules after venous occlusion is equal to
the total flow rate minus the volume stored in upstream vessels
|
(9)
|
During
the quasi-steady state, Fa(t) = a and
Fv(t) = v. Therefore, dividing Eq. 8 by Eq. 6 gives the
fractional compliance downstream from the arteriolar measurement site
|
(10)
|
and
dividing Eq. 9 by Eq. 6 gives the fractional compliance downstream from the
venular measurement site
|
(11)
|
The occlusion time intercepts, Pai
and Pvi, of the simultaneously
measured Pa and Pv after venous occlusion were obtained by linear
regression of Pa(t) and
Pv(t) over the time interval 0.5-1.0 s. These intercepts were used along with the compliances to put bounds on the resistances upstream from, between, and downstream from the 40-µm arterioles and venules, as previously described (3, 4,
7).
Model-dependent analysis.
The above calculations are model independent, in that no specific
compartmental arrangement of resistances and compliances is assumed. In
addition, they utilize only the quasi-steady-state flow and pressure
data from the venous-occlusion maneuver. Compartmental model
representations can also be used to interpret the transient portions of
the postocclusion data. Audi et al. (1, 2) provided the rationale for
using a 3C2R model representation of the lobar vasculature for
interpreting the occlusion pressure data (Fig. 1). Conceptually, the
upstream resistance (R1) was
visualized as representing primarily the resistance of the small
arteries, and the downstream resistance
(R2) was visualized as primarily the resistance of the small veins, with any capillary resistance contributing approximately equally to both
R1 and
R2 (Fig. 1). By use of the
anatomically defined compliances from the arteriolar and venular flow
measurements (Ca, Cc, and Cv) as model inputs for the upstream
(C1), middle
(C2), and downstream
(C3) compliances, respectively,
the 3C2R model resistances were calculated (16) by using the
double-occlusion pressure (Pd)
![R<SUB>2</SUB> = <FR><NU>[Pd − Pv(0)]C<SC>t</SC> − [Pa(0) − Pv(0)]C<SUB>1</SUB></NU><DE>F<SC>t</SC>C<SUB>2</SUB></DE></FR>](/content/vol84/issue1/fulltext/303/img013.gif)
|
(12)
|
and
|
(13)
|
where
Pa(0) and Pv(0) are the preocclusion steady-state Pa and Pv,
respectively. The 3 C's used as model inputs were obtained by first
estimating CT from the average
pressure data by using Eq. 5. The
values of C1,
C2, and
C3 were then calculated as the products of the fractional compliances and
CT (Table
1).
Once the model resistances were calculated, we investigated to what
extent the arteriolar and venular flows measured during venous,
arterial, and double occlusions were consistent with the compliances
and model resistances derived from the venous occlusion flows and
double-occlusion pressure. This was accomplished by numerically solving
the following governing differential equations for the 3C2R model (16)
describing the time variations in
Pa(t), Pc(t), and
Pv(t)
|
(14)
|
|
(15)
|
|
(16)
|
with
the initial (t = 0) conditions Pa(0) = Pv(0) + FT(R1 + R2) and Pc(0) = Pv(0) + FTR2,
where Fres is the flow into the reservoir.
Equations 14-16 were solved
numerically for Pa(t),
Pc(t), and
Pv(t) after
Fres was set to 0 for venous
occlusion and FT = Fres = 0 for double occlusion. For
arterial occlusion, FT was set
to 0, and only Eqs. 14 and 15 were solved for
Pa(t) and
Pc(t), with Pv(t) set equal to the measured
postocclusion venous pressure. These pressures were then used to
calculate the postocclusion arteriolar and venular flows
|
(17)
|
and
|
(18)
|
| |
RESULTS |
Direct measurements.
The normalized arteriolar and venular velocities from the individual
experiments measured during the occlusion maneuvers are presented in
Figs.
3-5.
The averages of the pressure curves obtained during the occlusion
maneuvers are shown in Fig. 6. The mean
values of the various measurements obtained during the preocclusion
period or from the quasi-steady states during venous or double
occlusion are summarized in Table 1.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 3.
Red blood cell velocities,
v(t),
before and after venous occlusion as a fraction of their mean velocity
during preocclusion measurement period,
v(0). Lines, measurements from a
single ~40-µm arteriole (A) and
venule (B) in each of 6 lobes.
Time 0 (vertical dashed line) is time
of occlusion. Filled bar on x-axis
represents quasi-steady-state data used to calculate compliance.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 4.
Red blood cell velocities,
v(t),
before and after arterial occlusion as a fraction of their mean
velocity during preocclusion measurement period,
v(0). Lines, measurements from a
single ~40-µm arteriole (A) and
venule (B) in each of 6 lobes.
Time 0 (vertical dashed line) is time
of occlusion.
|
|
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 5.
Red blood cell velocities,
v(t),
before and after double occlusion as a fraction of their mean velocity
during preocclusion measurement period,
v(0). Lines, measurements from a
single ~40-µm arteriole (A) and
venule (B) in each of 6 lobes.
Time 0 (vertical dashed line) is time
of occlusion.
|
|
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 6.
Arterial pressure, Pa, and venous pressure, Pv, before and after
arterial, venous, and double occlusion. Curves are averages of mean
pressure curves from each of 6 lobes. Data were digitally filtered
using a Butterworth low-pass filter with a cutoff frequency (3
dB) of 10 Hz and zero-phase shift (2). Filled bar on
x-axis represents quasi-steady-state
data used to calculate compliance.
|
|
Model-independent analysis.
To carry out the analysis of these measurements, the velocities were
converted to normalized flows (Fig. 7)
using Eq. 4. The fractional
compliances estimated from these flows (Eqs.
10 and 11) are shown
in Table 2. The Cc fraction was
significantly larger than either Ca or Cv
(P < 0.01, analysis of variance
followed by Newman-Keuls test). The upper and lower bounds placed on
the resistances partitioned by the 40-µm arterioles and venules are also shown in Table 2.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 7.
Calculated flow rates, F(t), before
and after venous occlusion as a fraction of mean flow rate during the
preocclusion measurement period, F(0). Flows were calculated from red
blood cell velocities through ~40-µm subpleural arterioles (thin
solid lines, n = 6) and ~40-µm
subpleural venules (thin dotted lines,
n = 6). A single arteriole and a
single venule were measured in each of 6 lobes. Thick lines, 3C2R
predictions for flows through resistances upstream and downstream from
C2, with R and C values estimated
from venous and double occlusions. Time
0 (vertical dashed line) is time of occlusion. Filled
bar on x-axis represents
quasi-steady-state data used to calculate compliance.
|
|
Model-dependent analysis.
When the 3C2R model and Pd were used to interpret the data,
R1/RT
was nearly equal to
R2/RT
(Table 3). The simulated venous and
double-occlusion pressure curves, generated by solving
Eqs. 14-16 for
Pa(t) and
Pv(t), and the simulated
arterial-occlusion pressure curves, generated by solving
Eqs. 14 and 15 for
Pa(t), are shown in Fig.
8. The arteriolar and venular flows
predicted by Eqs. 17 and 18 after arterial and double occlusion
followed the trend in the measured flows to the extent shown in Figs.
9 and 10.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 8.
Arterial pressure, Pa, and venous pressure, Pv, before and after
arterial, venous, and double occlusion. Thick lines, pressures predicted by 3C2R model using model parameters estimated as indicated in text; thin lines, average of actual mean pressure curves from each
of 6 lobes as shown in Fig. 5. Filled bar on
x-axis represents quasi-steady-state
data used to calculate compliance.
|
|
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 9.
Calculated flow rates, F(t), before
and after arterial occlusion as a fraction of mean flow rate during
preocclusion measurement period, F(0). Flows were calculated from red
blood cell velocities through ~40-µm subpleural arterioles (thin
solid lines) and ~40-µm subpleural venules (thin dotted lines). A
single arteriole and a single venule were measured in each of 6 lobes.
Thick lines, 3C2R predictions for flows through resistances upstream
and downstream from C2, with R and
C values estimated from venous and double occlusions.
Time 0 (vertical dashed line) is time
of occlusion.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 10.
Calculated flow rates, F(t), before
and after double occlusion as a fraction of mean flow rate during
preocclusion measurement period, F(0). Flows were calculated from red
blood cell velocities through ~40-µm subpleural arterioles (thin
solid lines) and ~40-µm subpleural venules (thin dotted lines). A
single arteriole and a single venule were measured in each of 6 lobes.
Thick lines, 3C2R predictions for flows through resistances upstream
and downstream from C2, with R and
C values estimated from venous and double occlusions.
Time 0 (vertical dashed line) is time
of occlusion.
|
|
| |
DISCUSSION |
The key observation in this study was a large decrease in venular flow,
but only a small fall in arteriolar flow, after venous occlusion. Thus,
after occlusion, a relatively small fraction of the flow continued into
veins >40 µm and a relatively small fraction was diverted into
distending arteries >40 µm. The largest fraction, ~65% of the
flow, increased the volume of vessels <40 µm, presumably
mostly capillaries. Therefore, the capillaries contained the largest
fraction of lobar vascular compliance. The relatively rapid fall in
arteriolar flow in comparison to venular flow after arterial occlusion
indicates that the time constant for emptying vessels upstream from the
40-µm arterioles was small in comparison to that upstream from the
40-µm venules, again consistent with a relatively small fraction of
the compliance in the arteries compared with the capillaries. Unlike
previous occlusion studies, our video-microscopic observations were
made at specific anatomic locations (40-µm-diameter arterioles and
venules), which allowed us to partition compliance into anatomically
defined segments.
The relative arterial-capillary-venous distribution of compliance
implied by these results of 8:65:27 can be compared with compliance
distributions obtained for the intrapulmonary vessels of the lungs by
other methods, which include 30:49:21 for the dog lung lobe vascular
bed (7) and 7:79:14 for the whole dog lung vascular bed (5). Consistent
with previous studies, the results of the present study lead to the
conclusion that the arteries and veins account for smaller fractions of
the total intrapulmonary vascular compliance than the capillaries. Thus
the results further support the concept that the intrapulmonary
vascular compliance is concentrated in the capillary bed.
The compliances estimated from the venous occlusion flows are based on
the concept that the fractional changes in flows in the subpleural
vessels are close to those in similarly sized vessels in the interior
of the lungs. This is a difficult assumption to test directly. It is
conceivable that at least one reason for any differences between these
estimates for arterial, capillary, and venous compliances based on the
subpleural vessel flows and those estimated by other means, for what
might be assumed to be the same vascular segments (5, 7), might be the
result of deviations from this assumption. However, the basic agreement between the results of this study and those obtained by very different methods (5, 7) supports the concept that the surface arterioles and
venules reflect the behavior of the internal vessels. Also supporting
this concept, Short et al. (17) found that recruitment of subpleural
capillaries in the rat lung paralleled recruitment of interior
capillaries over a range of pressures that spanned low zone 3 to high
zone 1. Finally, Hillier et al. (15), using fluorescent microspheres,
found no redistribution of flow between subpleural and interior vessels
in the isolated canine lobe during airway hypoxia. Although these
studies do not directly compare subpleural with interior flow changes
after vessel occlusion, they demonstrate that subpleural perfusion
parallels interior perfusion over a wide range of hemodynamic
conditions.
We divided total pulmonary vascular compliance into three segments
bounded by 40-µm arterioles and venules, because these were the
vessels that were observable on the surface of the lung. Because one
cannot assume that compliance is evenly distributed over the entire
diameter range, these bounds do not tell us anything about the
distribution of compliance within these three segments. Instead, the
sites where the measurements were made put lower bounds on the diameter
of arteries and veins contributing to the compliance of the upstream
and downstream segments, respectively. Likewise, the measurement sites
put upper bounds on the size of vessels contributing to the compliance
of the middle segment. Nevertheless, the available anatomic evidence
(8) suggests that there is little volume and, therefore, little
compliance in noncapillary vessels <40 µm. Most of the compliance
in the middle segment, therefore, is probably in the capillaries.
If the pressure data are included in the analysis, the occlusion
maneuvers also provide information about how resistance is distributed
with respect to compliance. The bounds placed on resistance associated
with each of the three compliances using the quasi-steady-state data,
i.e., the model-independent interpretation, are rather wide. The most
important conclusion from these bounds is that no more than ~50% of
the total resistance is associated with vessels <40 µm. If,
instead, the 3C2R model is used to interpret the data, the two
resistances, R1 and
R2, are nearly equal under the
conditions of this study. Like the finding that the major locus of
compliance is in the capillaries, this distribution of resistance is
consistent with previous interpretations based on this model (2). More importantly, this model also provides a means for simulating the time
course of flows immediately after arterial, venous, and double occlusion that appear to be reasonably consistent with the observed flow transients (Figs. 9 and 10). Thus, even with all the resistance in
the model lumped downstream from
C1 and upstream from
C3, the rate of fall in simulated
arterial flow after arterial occlusion and the rate of fall in
simulated arterial and venous flows after double occlusion (i.e., the
flow transients) are at least as rapid as the measured rates.
Alternatively, if a significant fraction of the total resistance had
been upstream from C1 or
downstream from C3, the rates of
decrease in flows would have been expected to be even faster. Thus
these transient flow data are consistent with most of
C1 and
C3 being located upstream from any
resistance in arteries >40 µm and downstream from any resistance in
veins >40 µm.
In summary, we have obtained anatomically referenced measurements of
flow after occlusion of the venous outflow which indicate that the bulk
of pulmonary vascular compliance is in the capillary bed. These
compliance measurements, when combined with the 3C2R model (2) and flow
measurements after occlusion of arterial inflow and simultaneous
occlusion of arterial inflow and venous outflow, suggest that most of
the compliance in arteries of >40 µm is upstream from any
resistance in these vessels and the compliance in veins of >40 µm
is downstream from any resistance in these vessels.
The authors thank Dr. Tawfik S. Hakim for assistance with pilot
studies.
This work was supported by National Heart, Lung, and Blood Institute
Grants HL-36033 and HL-19298 and the Department of Veterans Affairs.
Address for reprint requests: R. G. Presson, Jr., Riley Children's
Hospital, 702 Barnhill Dr., Rm. 2001, Indianapolis, IN 46202-5200.
Top
Abstract
Introduction
Previously, the
pressure changes after arterial and venous occlusion have been used to
characterize the longitudinal distribution of pulmonary vascular
resistance with respect to vascular compliance using compartmental
models. However, the compartments have not been defined anatomically.
Using video microscopy of the subpleural microcirculation, we have
measured the flow changes in ~40-µm arterioles and venules after
venous, arterial, and double occlusion maneuvers. The quasi-steady
flows through these vessels after venous occlusion permitted an
estimation of the compliance in three anatomic segments: arteries >40
µm, veins >40 µm, and vessels <40 µm in diameter. We found
that ~65% of the total pulmonary vascular compliance was in vessels
<40 µm, presumably mostly capillaries. The transient portions of
the pressure and flow data after venous, arterial, and double occlusion
were consistent with most of the arterial compliance being upstream
from most of the arterial resistance and most of the venous compliance
being downstream from most of the venous resistance.
