Vol. 84, Issue 4, 1350-1358, April 1998
Improved oxygenation with prostaglandin
F2
with and without inhaled
nitric oxide in dogs
Bryan E.
Marshall,
Linda
Chen,
H. Fred
Frasch,
C. William
Hanson, and
Carol
Marshall
Center for Anesthesia Research, Department of Anesthesiology,
University of Pennsylvania, Philadelphia, Pennsylvania 19104
 |
ABSTRACT |
Dogs of mixed
breed (n = 7) were anesthetized, right
lung atelectasis was established, and the cyclooxygenase pathway was blocked with ibuprofen. Measurements of pulmonary gas exchange were
performed (fractional concentration of inspired
O2 = 0.95) after infusions of
prostaglandin F2
(PGF2; 2 µg · kg
1 · min1),
ventilation with nitric oxide (NO; 40 ppm), or both
(PGF2 + NO) in random order.
The arterial PO2
(PaO2) under control conditions was 117 ± 16 Torr (shunt = 33 ± 2.5%), was unchanged with NO alone
(PaO2 = 114 ± 17 Torr; shunt = 35.7 ± 3.1%), but was significantly
improved with PGF2 alone
(PaO2 = 180 ± 28 Torr; shunt = 23.2 ± 2.8%) and with the combination of
PGF2 + NO
(PaO2 = 202 ± 30 Torr; shunt = 20.9 ± 2.5%). The addition of NO did
not significantly enhance the effectiveness of the
PGF2 on
PaO2.
Simulation of these data in a computer model, combining pulmonary gas
exchange and pulmonary blood flow, reproduced the results on the basis
that vasoconstriction with PGF2
was maximal under hypoxia in the atelectatic lung and reduced by
hyperoxia in the ventilated lung, consistent with the hypothesis of
O2 dependence of
PGF2 vasoconstriction.
almitrine; nitric oxide synthase inhibition; pulmonary
vasoconstrictors; hypoxic pulmonary vasoconstriction; computer
modeling
| |
INTRODUCTION |
WHEN ATELECTASIS is induced, whether by pathology or
deliberately, as during one-lung anesthesia for thoracic surgery, the arterial PO2
(PaO2) is
determined primarily by the fraction of the cardiac output (CO) that
continues to perfuse the atelectatic lung. The
PO2 of the atelectatic lung
approximates that of the mixed venous blood so that hypoxic pulmonary
vasoconstriction (HPV) is induced and blood flow is diverted to the
opposite, ventilated lung. The more effective the HPV, the greater is
the diversion of blood flow and the higher the
PaO2.
In a theoretical analysis (15), it was proposed that infusion of a
small artery pulmonary vasoconstrictor should enhance the effectiveness
of HPV in atelectasis. Furthermore, because vasoconstriction occurs in
both lungs, administration of nitric oxide (NO) by inhalation to the
ventilated lung (20) should further enhance the beneficial effects of
the vasoconstrictor, whereas administration of NO in the absence of
vasoconstriction was unlikely to be effective (21). A variety of
studies has substantiated these predictions in animal and human
investigations using phenylephrine (3), NO synthase inhibitors (4), or almitrine (13, 30) as the vasoconstrictor agent with inhaled NO.
Our original prediction, from the theoretical analysis, was that any
small-artery pulmonary vasoconstrictor would enhance the effectiveness
of HPV. However, it has been hypothesized by others (12, 28) that when
prostaglandin F2
(PGF2) is the
vasoconstrictor, O2 acts like a
competitive antagonist. The same dose of
PGF2 induces a greater
vasoconstriction in the small pulmonary arteries under hypoxic
conditions (atelectatic lung) than in those same arteries under
normoxic or hyperoxic conditions (ventilated lung). Furthermore,
because vasoconstriction, by HPV or exogenous
PGF2, is partially offset by
local release of endogenous vasodilator prostacyclin, the strength of
the constrictor response is enhanced by cyclooxygenase inhibition (1,
26).
This study has therefore tested, both experimentally and by computer
modeling, the hypotheses that infusion of
PGF2 improves gas exchange in
one-lung atelectasis in dogs with cyclooxygenase block. NO was
administered to the ventilated lung to test the O2 sensitivity of
PGF2 by differentiating the
constriction in the atelectatic and ventilated lung.
| |
METHODS AND MATERIALS |
Anesthesia and surgery.
After approval of the protocol by the Institutional Review Board, seven
female dogs of mixed breed with mean weight of 19.8 ± 0.5 kg were
anesthetized with an initial bolus of 30 mg/kg pentobarbital sodium
intravenously, followed by an infusion of 0.8-2.5 mg/min via a
Harvard infusion pump (model 902). The trachea was intubated with a
double-lumen Kottmeier endobronchial tube (Rüsch), and ventilation of both lungs was initiated. Muscle paralysis was established with 0.05 mg/kg pancuronium intravenously, supplemented with 0.2-0.5 mg every 30 min. Lung isolation was tested by adding He to the inspired gas to the right lung and analyzing the left lung
expired gas for absence of He by mass spectrometry (Perkin-Elmer model
1100 medical gas analyzer). If cross contamination was detected, the
Kottmeier tube was repositioned through a subcricoid tracheostomy until
isolation was complete. The lungs were ventilated synchronously but
independently with 100% O2 using
a Harvard dual-piston respirator with 7 cmH2O positive end-expiratory
pressure (PEEP) applied by a water seal.
Tidal volumes to both lungs were selected to achieve equal peak airway
pressures (~20 cmH2O), and the
respiratory rate was adjusted to maintain end-tidal
PCO2 at 35-40 Torr. Each piston of the
Harvard ventilator was part of a separate gas circuit with gas
composition determined by separate banks of flowmeters. Right and left
lung inspired, end-tidal, and mixed expired
PO2 and
PCO2 were measured by mass
spectrometer (Perkin-Elmer 1100 medical gas analyzer), which was
calibrated with gases of known composition and corrected for barometric
pressure, temperature, and water vapor.
Peripheral veins were cannulated for intravenous fluid administration
(~8.1 ± 0.3 ml/min Normosol) and administration of drugs. Body
temperature was maintained close to 37°C by heat lamps and a
thermostatic pad. Sodium bicarbonate was administered to correct acidosis after each phase, and urine was collected via a Foley catheter. To block the cyclooxygenase pathway for prostaglandin synthesis, ibuprofen (12.5 mg/kg) was infused intravenously over 10 min
before the beginning of surgery.
A femoral artery was cannulated percutaneously (Becton-Dickinson,
Angiocath, 18 gauge) for measurement of pressure and sampling of
arterial blood. A femoral vein was cannulated percutaneously (Arrow
percutaneous introducer cannula 8.5-Fr gauge), and a balloon-tipped catheter (American Edwards 7-Fr) was advanced into a pulmonary artery
for measurement of pulmonary arterial (PAP) and pulmonary arterial
occlusion pressures (PAOP), sampling of mixed venous blood, and
measurement of CO by thermal dilution (Edwards cardiac output computer
model 9510-A) using quadruplicate injections of ice-cold 5% dextrose
in water. Intravascular and ventilation pressures were measured with
transducers (Statham P 23BB and Gould-Statham P 23Db) zeroed at the
midcardiac level, and the transduced pressures were recorded on an
eight-channel Grass polygraph and on computer disk after processing
through Viewdac software. Respired volumes and respiratory rate were
measured by turbine spirometers (Boehringer Laboratories model 8830) in
the expired limb of each ventilator circuit.
A midline thoracotomy was performed, and the position of the Kottmeier
tube was confirmed. In four of the dogs, the origin of the bronchus to
the right upper lobe was occluded when the endobronchial tube was
positioned for complete separation of the two lungs, and the right
upper lobe was already atelectatic. The presence of an HPV response was
checked by ventilating the right lung with an hypoxic gas mixture (3%
CO2-4%
O2-balance
N2) for ~10 min, followed by a
return to 100% O2. This sequence
was repeated until a stable HPV response of at least
3-cmH2O increase of mean pulmonary
artery pressure was observed.
After both lungs were ventilated with 100%
O2 for 15 min, the ventilation of
the right lung was discontinued, and the connection of the Kottmeier
tube to the right lung was opened briefly to atmospheric pressure to
allow the lung to deflate before the orifice was occluded with a rubber
stopper. The thoracotomy permitted direct confirmation that complete
atelectasis of the right lung was induced in ~4 min and was
maintained for the entire subsequent study. After anticoagulation
(heparin, 50 U/kg iv) and a 30-min stabilization period, the study
protocol commenced.
Study design.
The experiment was performed in two parts. The first part examined the
effects of 40 ppm of inhaled NO and infused
PGF2 (2 µg · kg1 · min1),
both separately and combined, in random order. The second part examined
the influence of infused sodium nitroprusside (SNP) and was used to
characterize the pulmonary vascular bed for the computer modeling. The
two parts were each bracketed by control phases during which no
vasoactive drugs were administered; these control phases are identified
as control 1,
2, or
3. Control
2 separated the two parts of the experiment, and the
same phase is used in presenting the results of both parts. The total
experiment therefore consisted of seven phases.
NO (2,200 ppm) was mixed with N2
using a series 2000 computerized multicomponent mass-flow controller
(Environics) to obtain an inspired concentration of 40 ppm. The
concentration of NO and the absence of higher oxides of
N2 were analyzed continuously by
chemiluminescence (CLD 700 AL analyzer, EcoPhysics). This addition of
~300 ml of NO-N2 reduced the
inspired O2 concentration to 95%, and therefore 5% N2 was added to
all the other phases so that the alveolar
PO2
(PAO2)
was maintained constant.
The PGF2 was infused at a
constant rate of 2 µg · kg1 · min1 into a central venous
cannula. The SNP was made up in 5% dextrose at a concentration of 400 µg/ml and administered at a rate (10-50 µg · kg1 · min1)
to achieve and maintain a 50% reduction of the mean systemic arterial
pressure.
Measurements.
Each phase occupied ~45 min, and the following measurements were
obtained: peak, mean, and end-expiratory (PEEP) airway pressures, PAP,
systemic arterial pressures, central venous pressure, PAOP, CO, body
temperature, and inspired, end-tidal, and mixed expired O2 and
CO2 of the left lung ventilation
circuit. Arterial and mixed venous blood-gas samples were collected and
analyzed immediately to determine pH,
PO2,
PCO2 (Corning analyzer model 168),
and hemoglobin quantity and saturation (Instrumentation Laboratories
CO-oximeter model 482). The left lung tidal volume, respiratory rates,
and minute ventilation were recorded.
Calculations.
From the measured data, the pulmonary vascular resistance (PVR),
pulmonary shunt, and dead space were derived. Total PVR was calculated
as pulmonary perfusion pressure [(PAP 
PAOP)/CO]. PAO2
was calculated from the following form of the mixing equation
where
PIO2
is inspired PO2,
PACO2
is alveolar PCO2, and
O2
and
CO2 are mixed expired PO2 and
PCO2. End-capillary PO2 was assumed equal to
PAO2.
Saturation (Sat), corrected for pH and temperature, was calculated from
a nomogram for canine hemoglobin (22). The
O2 contents of end-capillary, arterial, and mixed venous blood was calculated from
Pulmonary
shunt was calculated using the traditional shunt equation.
Statistics.
The experimental variables were analyzed by within-subject ANOVA with
Newman-Keuls test for specific differences between means. A two-tailed
value of P < 0.05 was considered
significant. Results are expressed as means ± SE.
Modeling.
The theoretical basis for the action of vasoactive drugs was examined
using a previously described computer model
[ventilation/perfusion-pressure/perfusion (
A/
-P/)].
The model is based on a biodynamic representation of the pulmonary
circulation as 24 generations of arteries and veins and a capillary
sheet (6). This model of pressure-flow relations was generalized,
and active regulation by HPV was incorporated (P/
model) (16). Finally, the
A/-P/
model was developed by combining this model of the pulmonary
circulation with the familiar model for representing gas exchange as a
multicompartment distribution of ventilation/perfusion ratios
(A/ model) (29). The nature of the
A/-P/
model, the assumptions incorporated, and its ability to predict
experimental outcomes have been described elsewhere (10, 14).
The following assumptions are relevant for the present application. It
is assumed that alveolar gases influence only vessels with diameters of
<500 µm, and for simplicity, these are referred to here as small
arteries, small veins, and the capillary sheet, the remaining vessels
being grouped as large arteries or veins. HPV is assumed to act only at
small arteries as a function of the hypoxic stimulus
(PsO2), the
strength of the HPV responsivity, and the level of preexisting
constriction. The hypoxic stimulus is a function of both
PAO2
and mixed venous PO2 (
O2),
i.e.,
PsO2 = PAO20.6 × O20.4
(14), and in an atelectatic lung,
PsO2 = O2.
NO acts only on small vessels and inhibits HPV and other
causes of constriction, according to the dose-response relationship
defined by others (7, 23). The NO concentration is determined by the
A/ ratio
for each lung compartment and the assumption that the concentration within the blood phase is immediately reduced to zero by combination with hemoglobin.
From the present experimental phases, the results of the study with SNP
provided a basis for defining the metabolic state and characterizing
the gas exchange and pulmonary circulation when atelectasis is present
but all pulmonary vascular tone is abolished. The changes observed
during the control measurements therefore permit adjustment of the
spontaneous vascular tone and HPV parameters in the
A/-P/
model to match the results observed in the experiments.
PGF2 is reported to act only on
the arterial side of the pulmonary circulation in dogs of mixed breed
(8). Therefore, only the levels of constriction of the large and small arteries need to be differentially adjusted in the model to approximate the observed results. NO at 40 ppm is assumed to exert a maximal inhibition of the constriction of small vessels, and changes observed when NO is administered, in the absence and presence of
PGF2, permit the model to
identify the interactions at the small arteries. A
A/ ratio
maldistribution amounting to the upper limit of normal, log
SD(A/) = 0.6 [where log
SD(A/) is logarithm of standard deviation of
A/ ratio
distribution for ventilation () and perfusion
()], was assumed, but the inspired
O2 concentration was 95%, and
therefore the effect of this maldistribution is not directly on the
efficiency of O2 exchange or HPV.
However, the alveolar NO concentration varies directly with the
A/ ratio of the
alveolar compartment and therefore influences blood flow distribution
to ventilated alveoli when small arteries are constricted by hypoxic or
nonhypoxic causes.
| |
RESULTS |
Experimental animals.
General information for the seven dogs included in this study is shown
in Table 1. The values shown for
PaO2, arterial
PCO2, arterial pH, and body
temperature were obtained after completion of the surgical preparations
but before induction of complete right lung atelectasis, and they
confirm normal initial pulmonary function (lowest
PaO2 values
correspond to four dogs with right upper lobe atelectasis as a result
of endobronchial tube position).
NO and PGF2.
All seven dogs completed the five phases that constituted this part of
the study. The values for the control phases, controls 1 and 2, were not
statistically different by paired
t-test. Therefore, for each animal,
the mean value was calculated (
), and
this value is shown in the tables and for the subsequent analysis. The
mean data are shown for the ventilation and gas exchange in Table
2 and for pressure and flow data in Table
3. Analysis of these results reveals that
NO was accompanied by a small increase of CO, but no other significant
effects were observed. When
PGF2 was infused, there were
significant increases of PAP, PVR, and PaO2, with a
30% decrease in pulmonary shunt. A small but significant increase of
peak inspiratory pressures occurred, but inhalation of NO had no
significant effect on airways (11). When NO and PGF2 were combined, small but
significant decreases in CO, PAP, and PVR occurred, but the small
increase of
PaO2 was not significant compared with PGF2
alone. PaO2,
PAP, pulmonary shunt, and PVR values at each phase for each dog are
shown in Fig. 1.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 1.
Individual data and means ± SE from 7 dogs receiving prostaglandin
F2
(PGF2) and/or nitric
oxide (NO) showing changes for arterial
PO2
(PaO2;
A), pulmonary shunt
(B), pulmonary artery pressure (PAP;
C), and pulmonary vascular
resistance (PVR; D). Pattern is
evident, with NO alone having little effect,
PGF2 alone improving
O2 exchange (increased
PaO2 and
decreased shunt) and increasing PAP and PVR, and combination of NO and
PGF2 demonstrating only small
further improvements of oxygenation while reducing PAP and PVR
increases.
|
|
SNP.
Five dogs completed the three phases (control
2, SNP infusion, and control
3) that constituted this part of the study. Analysis of the data for the two series of control observations demonstrates a
significant decrease in the pulmonary shunt between
controls 2 and
3, and therefore they were not
combined. Compared with control 2, the
effect of SNP was to reduce systolic blood pressure from 127 ± 4 to
69 ± 2 mmHg. Systemic vascular resistance, PAP, and PVR also
decreased while PO2 decreased from
134 ± 24 to 64 ± 5 Torr, and pulmonary shunt increased from 31 ± 3 to 53 ± 3%.
PaO2, PAP,
pulmonary shunt, and PVR values at each phase for each dog are shown in
Fig. 2. These data demonstrate that the
total right lung blood flow was ~53% of the total CO in the absence
of any vasoconstriction and that atelectasis was associated with a 48%
diversion of blood flow from the right lung in the presence of normal
HPV.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 2.
Individual data and means ± SE from 5 dogs receiving sodium
nitroprusside (SNP) showing changes for
PaO2
(A), pulmonary shunt (B), PAP
(C), and PVR
(D). Control-2, control 2 phase. Infusion of SNP impaired oxygenation (decreased
PaO2 and
increased shunt) and decreased PAP and PVR.
|
|
Modeling.
The dogs were represented in the
A/-P/
model by the following general properties. The CO during the
anesthetized open-chest control phase ()
was assumed to be 75% of the normal awake value. The
P50 for hemoglobin was estimated
at 29.3 Torr from a plot of the measured
PO2, and saturation of mixed venous blood from all phases and was assumed to be constant. The other values
used in the model, including barometric pressure, body temperature,
alveolar ventilation, hemoglobin, mixed venous
PO2, PCO2, and base excess, airway
pressures, inspired PO2, and NO
concentrations (ppm), were the means of the values observed in the
experimental preparations. The amount of the lung that was atelectatic
was assumed to remain constant throughout the study and to be the same
as the mean shunt flow, 53% of the CO, calculated during the SNP
phase. This value includes atelectatic areas of the left lung as well
as the deliberately induced atelectasis of the right lung. For the
model, a shunt flow of 53% is therefore entered at the start of the
simulation of all phases. It is the model's calculation of how HPV
and vasoconstriction
(PGF2) or dilation (NO)
influences blood flow distribution and gas exchange that
determines the
PaO2 and PAP,
which are the essential outputs generated. The model parameters for the
sites and extent of vascular constriction are the only values adjusted
until the calculated outputs match those observed. Although there are
reports of constriction of veins by
PGF2 in some species (9, 18,
19), constriction is confined to pulmonary arteries in dogs of mixed
breed (8), and therefore only actions on pulmonary arteries were
considered in the simulations.
With so many potential variables and the approximations inherent in any
computer model, other solutions are undoubtedly possible, but the model
parameters displayed in Table 4 are the
simplest solutions that fit the observations. The derivation of these
parameters and the arguments on which they are based are as follows. A
value of 1.0 for the large or small arteries means that at the start of
the simulation the vessels have diameters corresponding to no intrinsic
tone. A value of 1.0 for HPV means that the small arteries have
normally reactive responses to hypoxia, and therefore a maximal HPV
response is reached when the vascular diameter is reduced by 40% of
the relaxed value. During administration of SNP, when all tone in the
lung is abolished, the arteries have initial diameters corresponding to
the relaxed state while HPV is reduced to zero. With the control
measurements, HPV is enhanced to 1.25 because the prostacyclin release
that normally modulates the HPV response is eliminated by the
cyclooxygenase inhibition. Initial tone is necessary in the large
arteries (0.85) to account for the changes in PAP in addition to that
resulting from HPV. However, the tone of the small vessels (1.0) is not
increased; if it were, the effects of inhaled NO would be to
reduce PAP and increase
PaO2, results
not observed experimentally. The influence of
PGF2 is arrived at by adjusting
the initial constriction of the small and large arteries until the
observed values are reproduced. These same initial values are then used
for the simulation when NO is added. The selections converge for this
experimental design because increased tone in the large arteries
(initial diameter = 0.85) primarily changes total vascular resistance
and slightly impairs oxygenation, HPV is only active in the atelectatic
lung, NO acts on the ventilated lung, and
PGF2 acts initially on both
lungs, but its final efficacy is modified as a function of the local
PO2. It is the reduction of
small-artery diameters that determines the extent to which NO improves
oxygenation under these conditions, and it is evident that the relative
ineffectiveness of NO in the presence of
PGF2 vasoconstriction is due to the reduced vasoconstriction in the hyperoxic ventilated lung.
| |
DISCUSSION |
The experimental results demonstrate that, with atelectasis of one
lung, infusion of PGF2 improves
oxygenation, whereas inhaled NO alone does not, and the combination of
NO and PGF2 is not
significantly more effective than
PGF2 alone. A reasonable explanation for these results is that the difference in
PO2 on the two sides influences the
action of the vasoconstrictor so that there is a reduced constrictor
action in the hyperoxic lung compared with the atelectatic lung; the
HPV response is therefore apparently enhanced. This is in striking
contrast to studies using other small-artery vasoconstrictors in the
presence of one-lung atelectasis or hypoxia, in which their
administration increased PVR markedly but only modestly increased
oxygenation, whereas in combination with NO the vascular resistance
increase was lessened and oxygenation markedly improved (3, 5, 16, 34).
O2 sensitivity of
PGF2.
Modeling shows that apparent enhancement of HPV of small pulmonary
arteries subjected to nonhypoxic constriction is simply because the
same hypoxic constriction has a proportionately greater effect on PVR
when applied to already narrowed arteries (15). If administration of a
vasoconstrictor does not show this effect, then it is evidence that the
endogenous vasoconstrictor is not acting at the same site as HPV (e.g.,
histamine and serotonin; Ref. 12). However, when
PGF2 is the vasoconstrictor,
the interaction with HPV is exaggerated (12, 28) beyond that predicated on the above concept. Comparison of the results (Fig.
3, A and B) observed by Tucker et al. (28)
with those from the
A/-P/ model allows two interpretations as follows (see
APPENDIX for model methodology). If
the small-artery constriction due to
PGF2 is adjusted
(K = 0.769 in Fig.
3B) so as to reproduce the observed constriction at the highest
PsO2 (~200
Torr), then the model underestimates the constriction
observed at the lowest
PsO2.
Therefore, under this first interpretation, the model can reproduce the
observed behavior only if HPV is enhanced by
PGF2 at
PsO2 <100
Torr. Conversely, if the small-artery constriction due to
PGF2 is applied
(K = 0.666 in Fig.
3B) so as to reproduce the
observed results at the lowest
PsO2 (~60
Torr), then the calculated PVR is too high at greater
PsO2, and the
model small arteries are too constricted. Accordingly, for the second
interpretation, the model can simulate the observed result if the
efficacy and therefore the initial constriction due to the
PGF2 are reduced as
PsO2 increases.
Both these interpretations were suggested by Lonigro and Dawson (12)
and Tucker et al. (28), although with the data available at that time
they could not distinguish between them. These authors did not suggest
a basis for the enhancement of HPV, but they postulated that a
mechanism for the second interpretation was
O2-dependent metabolism of
PGF2. It should be noted that this discrepancy between observed values and the
A/-P/
model has so far only been noted with
PGF2 and is not observed when
phenylephrine, almitrine, or an NO synthase-blocking drug is the
vasoconstricting agent (see Fig. 4).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 3.
Quantitation of O2-dependent
PGF2 effect. In
A, data of Tucker et al. (28) have
been replotted to show mean PVR as small-artery
O2 tension
(PsO2; see text
for definition) was reduced in absence and in presence of
PGF2 (2 mg · kg1 · min1).
Vertical difference between 2 lines at each of 4 points is increase of
PVR with PGF2, and this
increase is plotted in B ( ).
Response lines generated by
ventilation/perfusion-pressure/perfusion (A/-P/)
model as hypoxia is applied are calculated with constriction factor
adjusted to match maximal constriction at PsO2 = 60 Torr ( ; K = 0.666) and minimal
constriction at
PsO2 = 200 Torr ( ; K = 0.769). Dot-dashed line
shows continuous
A/-P/ model simulation of observed data after application of
O2-dependent change in efficacy of
PGF2 (see text for
derivation).
|
|
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 4.
Influence of vasopressor characteristics on
PaO2 and PAP.
Results of changing vasopressor characteristics for
A/-P/ model in absence (closed symbols) and presence (open symbols) of NO (40 ppm) are illustrated. General conditions were maintained as described
for modeling experimental data. Only vasoconstriction of large arteries
(CLA) or small arteries
(CSA) or activity of HPV used
for
A/-P/
model have been altered as follows.
1) Baseline condition where
CLA = 0.85 and
CSA and HPV = 1.0. 2) Further vasoconstriction of large
pulmonary arteries only (CLA = 0.7, CSA and HPV = 1.0) results in
a small deterioration of PaO2 because
distribution of blood flow between 2 parallel pathways depends on ratio
of their resistances. Vasoconstriction of large arteries increases
resistance relatively more to hyperoxic lung. 3) Even though vasoconstriction of
large arteries is more than that of small arteries
(CLA = 0.7, CSA = 0.9, HPV = 1.0),
vasoconstriction of small arteries underlies marked improvement in
PaO2 and PAP when NO is inhaled. 4) When
vasoconstriction is as for 3 but HPV is enhanced (CLA = 0.7, CSA = 0.9, HPV = 1.4), there is
further improvement in
PaO2 and marked
sensitivity to NO. 5) With addition of O2-dependent small-artery
vasoconstriction as described for PGF2
(CLA = 0.8, CSA = 0.7, HPV = 1.25),
improvement in
PaO2 is equivalent to combined vasopressor-NO condition illustrated by
4 and NO has little additional effect.
6) When vasoconstriction of small
arteries is greater than that of large arteries
(CLA = 0.8, CSA = 0.7, HPV = 1.0), profound
importance of small-artery vasoconstriction to dramatic response to NO
is evident. 7) Enhancement of HPV
added to conditions otherwise as for 6 (CLA = 0.8, CSA = 0.7, HPV = 1.4) acts to
further improve
PaO2 and
retains maximal responsiveness to NO.
|
|
The following evidence now provides strong support for the second
interpretation and specifically for the concept that the constriction
stimulated by a particular dose of
PGF2 is reduced as a direct
function of
PsO2. In the
present studies, the unexpectedly small dilator response to NO
administered to the ventilated lung in the presence of
PGF2 (Fig. 1) suggests that the
increase in the overall PVR induced by the vasoconstrictor was due
primarily to constriction of the atelectatic, or hypoxic, lung despite
administration of the same concentration of drug to both lungs. Also
consistent with this conclusion is a study of the influence of
PGF2 in dogs with one-lung
atelectasis (24) that was very similar to the present study except that
the PGF2 was deliberately administered only to the atelectatic lung. Thus constriction primarily of the hypoxic (atelectatic) lung is observed whether the
PGF2 is administered to the
hypoxic lung alone or to both the hypoxic and hyperoxic lungs.
Increased metabolism of the
PGF2 as a direct function of
PO2 could account for the reduced effect in the hyperoxic lung, and biochemical studies in hepatocytes support this conclusion (27).
When the differential effect on hypoxic-hyperoxic lungs is considered,
the dose-response relationship for
PGF2 is important. The action
is not apparent at infusion rates of 0.01-0.1
µg · kg1 · min1
(25), becomes just detectable at ~0.4
µg · kg1 · min1
(28), is maximal at 2 µg · kg1 · min1
(23, 27), and becomes less effective again at high doses (24).
Vasoconstriction induced by
PGF2 in vivo is partially opposed by the release of endogenous vasodilator prostaglandins, and
the sensitivity of this effect is enhanced by blockade of the
cyclooxygenase synthetic pathway (1, 26). Therefore, optimal conditions
were obtained in the present study by combining ibuprofen-mediated
cyclooxygenase block and a PGF2
infusion dose of 2 µg · kg1 · min1.
Comparison of vasoactive agents.
Others have reported different combinations of vasoactive drug
infusions with inhalation of NO. Freden et al. (4) reported that in
pigs with left lower lobe hypoxia
(PaO2 = 300 Torr), NO alone slightly worsened oxygenation
(PaO2 = 266 Torr), infusion of the NO synthase inhibitor
NG-nitro-L-arginine methyl ester
(L-NAME) improved oxygenation
(PaO2 = 320 Torr), and the combination of NO with
L-NAME greatly augmented oxygenation
(PaO2 = 435 Torr). In a sheep model in which lung damage was induced by bilateral
lung lavage
(PaO2 = 68 Torr), Rovira et al. (23) reported that NO alone improved gas exchange (PaO2 = 115 Torr) and, in a second series, that infusion of
L-NAME had no effect (baseline
PaO2 = 109 Torr), whereas the combination of NO with
L-NAME augmented oxygenation
(PaO2 = 195 Torr). Infusion of L-NAME alone
was accompanied by a 30% decline in CO and a corresponding decrease of
O2.
That PaO2 did
not change indicates that the efficiency of
O2 exchange was improved even by
L-NAME alone. Finally, in humans
with severe acute respiratory distress syndrome
(PaO2 = 88 Torr), Wysocki et al. (30) showed that NO alone improved oxygenation
(PaO2 = 98 Torr) and infusion of the HPV-enhancing agent almitrine improved
oxygenation
(PaO2 = 106 Torr) and was further enhanced
(PaO2 = 130 Torr) with the combination. In all these investigations, the inspired
O2 concentration was 
80%, and
therefore the impairment of gas exchange is primarily due to pulmonary
shunting.
It is evident that the effectiveness of the vasopressor alone and of
the combination with NO varies with the vasopressor and the lung
abnormality. Normally, a low level of NO release from the vascular
endothelium occurs continuously and is enhanced when pressure
and/or flow increases the wall stress. The characteristics of
L-NAME are therefore that
abolition of NO results in both increased pulmonary (and systemic)
vascular tone in normoxic conditions and enhancement of HPV. Almitrine,
in contrast, stimulates vasoconstriction only in the pulmonary
circulation and appears to do so by acting, as does hypoxia,
principally on small pulmonary arteries. Almitrine and hypoxia are
therefore additive, but almitrine does not enhance HPV in the sense
that inhibition of NO release does. Several studies in animal and human
subjects have previously reported improved oxygenation with pulmonary
shunting when NO is inhaled in the presence of an infused
vasoconstrictor (3-5, 13, 30). These observations are
qualitatively consistent with the concept that inhaled NO will improve
oxygenation when there is significant atelectasis and small-artery
constriction at sites accessible to the NO. However, modeling permits a
more detailed analysis of the quantitative basis for the observations
and therefore a way to decide whether one vasoconstrictor appears
theoretically preferable. The purpose of the modeling is to identify
underlying factors that may serve as guides for research and clinical
applications. In this analysis, we consider only the effects of
vasopressors and/or NO on oxygenation with atelectasis and not
their effects on other abnormalities of gas exchange or other
properties (e.g., changing hemodynamics) that may profoundly influence
their value in practice.
Although experimental examples for conditions
2 and 4 in Fig. 4 do
not appear to have been reported, conditions
3 and 5-7 correspond respectively to phenylephrine (3),
PGF2 (present study), almitrine
(30), and NO synthase inhibition (4). From the results of Fig. 4 the
following general guidelines are derived for predicting the outcome of
vasopressors on pulmonary O2
exchange with or without NO. In the absence of NO, the efficiency of
O2 exchange will be increased if
the vasopressor enhances HPV, acts primarily at small pulmonary
arteries, or demonstrates
O2-dependent efficacy. NO will
only be effective to the extent to which the vasopressor acts
differentially on small pulmonary arteries in ventilated lung regions.
For general vasopressors and particularly NO synthase inhibition, there
are changes elsewhere in the systemic circulation that may be
undesirable for an unstable patient or may result in changes (e.g.,
decreased CO with increasing afterload) that more than offset an
improved efficiency of O2
exchange. Of the drugs presently considered,
PGF2 has the advantage of simplicity, safety, and brevity of action, whereas almitrine appears closest to ideal because its actions are confined to the pulmonary vasculature. For both agents, further clinical trials are warranted.
| |
APPENDIX |
The method of deriving and applying the specific correction factors to
simulate the O2-dependent activity
of PGF2 is described further
below to illustrate the functional basis of the
A/-P/
model. Each generation of artery and vein in the A/-P/
model is defined by resting diameter, length, compliance, and number of
branches. The action of a vasoconstrictor is incorporated by reducing
the resting diameter of the vessels affected, by a constriction factor.
If there is no constriction, the factor is 1.0, but if the constricted
diameter is one-half the preconstricted state, the factor is 0.5; in
Fig. 3B the constriction factors for
the two model curves that bracket the data of Tucker et al. (28) are
indicated as K = 0.666 and 0.769. Analysis of the change in constriction factor
(KI) that could
account for these data yields a simple exponential form, where
KI = 0.155 × {1 exp[0.0557(PsO2 60)]}. Note that Fig.
3B shows that the increase of PVR
reported by Tucker et al. (28) reaches a maximum at ~60 mmHg. This
result is the expected one when the whole animal is hypoxic, because we (17) and others (2) have shown that HPV is impaired
when the systemic arterial
PsO2 is <60
Torr. We have therefore chosen not to further adjust the constriction
factor for
PsO2
at <60 Torr. The
A/-P/
model interpolation shown in Fig.
3B, for the
O2-dependent
PGF2 effect at all intermediate
PO2 for the data of Tucker et al.
(28), provides a good fit to their data and also predicts that the
primary region of O2 sensitivity is in the 60- to 100-Torr range corresponding to sea-level inspired O2 concentrations of <35%. The
O2-dependent
PGF2 effect has been
incorporated in the general
A/-P/
model by correcting the resting diameters of the small arteries in each
of the 50 compartments according to their individual
PsO2.
The correction is applied iteratively because the change of
constriction changes the compartment blood flow and hence the
PsO2, which in
turn changes the constriction and so on until it converges at the
required tolerance (usually 0.1%). With this advance, the
A/-P/
model is used to simulate the present experimental results, and outputs for PaO2 and
PAP were within 5% of the observed values when the parameters shown in
Table 4 were selected.
| |
ACKNOWLEDGEMENTS |
The authors acknowledge the technical assistance provided by
P. Lilagen, R. Y. Katz PhD, and R. Andrews in the conduct of these studies. Thanks are also due to Dr. L. R. Soma, Dr. G. Acland, and A. Nickle of the New Bolton Campus of the University of
Pennsylvania Veterinary School, whose cooperation made this work
possible.
| |
FOOTNOTES |
This work was supported in part by National Institutes of Health Grants
GM-29628 and HL-09040.
Address for reprint requests: B. E. Marshall, Center for Research in
Anesthesia, 781 Dulles, Hospital of the University of Pennsylvania,
Philadelphia, PA 19096.
Received 31 March 1997; accepted in final form 2 December 1997.
| |
REFERENCES |
1.
Bishop, M. J.,
L. D. Artman,
and
F. W. Cheney.
Ibuprofen potentiates diversion of blood flow from hypoxic to normoxic lung.
J. Crit. Care
3:
112-115,
1988.
2.
Brimioulle, S.,
P. Lejeune,
J. L. Vachiiry,
M. Delcroix,
R. Hellemans,
M. Leeman,
and
R. Naeije.
Stimulus-response curve of hypoxic pulmonary vasoconstriction in intact dogs: effects of ASA.
J. Appl. Physiol.
77:
476-480,
1994[Abstract/Free Full Text].
3.
Doering, E. B.,
C. W. Hanson III,
D. J. Reilly,
C. Marshall,
and
B. E. Marshall.
Improvement in oxygenation by both phenylephrine and nitric oxide in adult respiratory distress syndrome.
Anesthesiology
87:
18-25,
1997[Medline].
4.
Freden, F.,
S. Z. Wei,
J. E. Berglund,
C. Frostell,
and
G. Hedenstierna.
Nitric oxide modulation of pulmonary blood flow distribution in lobar hypoxia.
Anesthesiology
82:
1216-1225,
1995[Medline].
5.
Frostell, C. G.,
H. Blomqvist,
G. Hedenstierna,
J. Lundberg,
and
W. M. Zapol.
Inhaled nitric oxide selectively reverses human hypoxic vasoconstriction without causing systemic vasodilation.
Anesthesiology
78:
413-416,
1993[Medline].
6.
Fung, W. C.
Biodynamics, Circulation. New York: Springer-Verlag, 1984, p. 290-364.
7.
Gerlach, H.,
R. Rossaint,
D. Pappert,
and
K. J. Falke.
Time-course and dose-response of nitric oxide inhalation for systemic oxygenation and pulmonary hypertension in patients with adult respiratory distress syndrome.
Eur. J. Clin. Invest.
23:
499-502,
1993[Medline].
8.
Hakim, T. S.,
C. A. Dawson,
and
J. H. Linehan.
Hemodynamic responses of dog lung lobe lobar venous occlusion.
J. Appl. Physiol.
47:
145-152,
1979[Abstract/Free Full Text].
9.
Hakim, T. S.,
M. King,
C. G. Wang,
and
M. Cosio.
Effect of chronic cigarette smoke exposure on pulmonary vasomotion in beagle dogs.
J. Appl. Physiol.
59:
1815-1822,
1985[Abstract/Free Full Text].
10.
Hanson, C. W.,
B. E. Marshall,
H. F. Frasch,
and
C. Marshall.
Causes of hypercarbia with oxygen therapy in patients with chronic obstructive lung disease.
Crit. Care Med.
24:
23-28,
1996[Medline].
11.
Hogman, M.,
C. G. Frostell,
H. Hendenstrom,
and
G. Hedenstierna.
Inhalation of nitric oxide modulates adult human bronchial tone.
Am. Rev. Respir. Dis.
148:
1474-1478,
1993[Medline].
12.
Lonigro, A. J.,
and
C. A. Dawson.
Vascular responses to prostaglandin F2 in isolated cat lungs.
Circ. Res.
36:
706-712,
1975[Abstract/Free Full Text].
13.
Lu, Q.,
E. Mourgeon,
J. D. Law-Koune,
S. Roche,
C. Vezinet,
L. Abdennour,
E. Vicaut,
L. Puybasset,
M. Diaby,
P. Coriat,
and
J. J. Rouby.
Dose-response curves of inhaled nitric oxide with and without intravenous almitrine in nitric oxide-responding patients with acute respiratory distress syndrome.
Anesthesiology
83:
929-943,
1995[Medline].
14.
Marshall, B. E.,
C. W. Hanson,
H. F. Frasch,
and
C. Marshall.
Role of HPV in pulmonary gas exchange and blood flow distribution. I. Physiologic concepts.
Intensive Care Med.
20:
291-297,
1994[Medline].
15.
Marshall, B. E.,
C. W. Hanson,
H. F. Frasch,
and
C. Marshall.
Role of HPV in pulmonary gas exchange and blood flow distribution. II pathophysiology.
Intensive Care Med.
20:
379-389,
1994[Medline].
16.
Marshall, B. E.,
and
C. Marshall.
A model for hypoxic constriction of the pulmonary circulation.
J. Appl. Physiol.
64:
68-77,
1988[Abstract/Free Full Text].
17.
Marshall, B. E.,
C. Marshall,
M. Magno,
P. Lilagan,
and
G. G. Pietra.
The influence of bronchial arterial oxygen tension in pulmonary vascular resistance.
J. Appl. Physiol.
70:
405-415,
1991[Abstract/Free Full Text].
18.
Matsuki, T.,
and
T. Ohnashi.
Endothelium and mechanical responses of isolated monkey pulmonary veins to histamine.
Am. J. Physiol.
259 (Heart Circ. Physiol. 28):
H1032-H1037,
1990[Abstract/Free Full Text].
19.
McCormack, D. G.,
J. C. Mak,
M. O. Coupe,
and
P. J. Barnes.
Calcitonin gene-related peptide vasodilation of human pulmonary vessels.
J. Appl. Physiol.
67:
1265-1270,
1989[Abstract/Free Full Text].
20.
Puybasset, L.,
J. J. Rouby,
E. Mourgeon,
T. E. Stewart,
P. Cluzel,
M. Arthau,
P. Poete,
L. Bodin,
A. M. Korinek,
and
P. Viars.
Inhaled nitric oxide in acute respiratory failure: dose-response curves.
Intensive Care Med.
20:
319-327,
1994[Medline].
21.
Rich, G. F.,
S. M. Lowson,
R. A. Johns,
M. O. Daugherty,
and
D. R. Uncles.
Inhaled nitric oxide selectively decreases pulmonary vascular resistance without impairing oxygenation during one-lung ventilation in patients undergoing cardiac surgery.
Anesthesiology
80:
57-62,
1994[Medline].
22.
Rossing, R. G.,
and
S. M. Cain.
A nomogram relating PO2, pH, temperatures, and hemoglobin saturation in the dog.
J. Appl. Physiol.
21:
195-201,
1966[Free Full Text].
23.
Rovira, I.,
C. Tong-Yan,
M. Winkler,
N. Kawai,
K. D. Bloch,
and
W. M. Zapol.
Effects of inhaled nitric oxide on pulmonary hemodynamics and gas exchange in an ovine model of ARDS.
J. Appl. Physiol.
76:
345-355,
1994[Abstract/Free Full Text].
24.
Scherer, R. W.,
G. Vigfusson,
E. Hultsch,
H. Van Aken,
and
P. Lawin.
Prostaglandin improves oxygen tension and reduces venous admixture during one-lung ventilation in anesthetized paralyzed dogs.
Anesthesiology
62:
23-28,
1985[Medline].
25.
Sprague, R. S.,
A. H. Stephenson,
L. J. Heitman,
and
A. J. Lonigro.
Differential response of the pulmonary circulation to prostaglandins E2 and F2 in the presence of unilateral alveolar hypoxia.
J. Pharmacol. Exp. Ther.
229:
38-43,
1984[Abstract/Free Full Text].
26.
Sprague, R. S.,
A. H. Stephenson,
and
A. J. Lonigro.
Prostaglandin I2 supports blood flow to hypoxic alveoli in anesthetized dogs.
J. Appl. Physiol.
56:
1246-1251,
1984[Abstract/Free Full Text].
27.
Tran-Thi, T.-A.,
A. Holstege,
and
K. Decker.
Effects of hypoxia on the oxygen-dependent metabolism of prostaglandins and adenosine in liver cells.
J. Hepatol.
20:
570-579,
1994[Medline].
28.
Tucker, A.,
E. K. Weir,
R. F. Grover,
and
J. T. Reeves.
Oxygen-tension-dependent pulmonary vascular responses to vasoactive agents.
Can. J. Physiol. Pharmacol.
55:
251-257,
1977[Medline].
29.
Wagner, P. D.,
H. A. Saltzman,
and
J. B. West.
Measurement of continuous distribution of ventilation-perfusion ratios: theory.
J. Appl. Physiol.
36:
588-599,
1974[Free Full Text].
30.
Wysocki, M.,
C. Delclaux,
E. Roupie,
O. Langeron,
N. Liu,
B. Herman,
F. Lemaire,
and
L. Brochard.
Additive effect on gas exchange of inhaled nitric oxide and intravenous almitrine bismesylate in the adult respiratory distress syndrome.
Intensive Care Med.
20:
254-269,
1994[Medline].
J APPL PHYSIOL 84(4):1350-1358
8570-7587/98 $5.00
Copyright © 1998 the American Physiological Society
Top
Abstract
Introduction
