L I BR A R Y
Michigan Stave
University

This is to certify that the

thesis entitled

Distribution of Ventilation
in Anesthetized Horses

presented by
Paul R. Sorenson

has been accepted towards fulfillment
of the requirements for

M. S. degree in Physiology

 

Major professor

Date May 11, 1978

 

0-7639

 

DISTRIBUTION OF VENTILATION

IN ANESTHETIZED HORSES

BY

Paul R. Sorenson

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Physiology

1978

ABSTRACT

DISTRIBUTION OF VENTILATION
IN ANESTHETIZED HORSES

BY

Paul R. Sorenson

Inhomogeneities of ventilation as they may relate to impaired gas
exchange during general anesthesia in the horse were assessed in four
body positions, 1) left lateral, 2) right lateral, 3) dorsal, and
4) sternal recumbency, using multiple and single breath N2 dilution
tests. Changes in body position did not shift the distribution of tidal
volume between functional lung units. Regional asynchronous ventilation
was significantly increased in the lateral and dorsal postures. Vital
capacity (VC), expiratory reserve (ER) and functional residual capacity
(FRC) were significantly elevated in sternal recumbency, suggesting a
reduction in lung elastic recoil. Closing volume (CV) did not change
with posture, and exceeded resting lung volume in all body positions
except sternal recumbency. Increased lung elastic recoil and reduction
in resting lung volume below CV indicate that small airway closure may

occur during tidal breathing and be a predominant factor limiting gas

exchange in the laterally recumbent anesthetized horse.

ACKNOWLEDGEMENTS

I must extend my appreciation to Dr. N. E. Robinson for his
support and guidance throughout this study, and the professional free-
dom he has extended to me over the past years. I would also like to
thank the members of my committee, Doctors L. F. Wolterink and S. R.
Heisey. A special appreciation is extended to Dr. Wolterink for his
interest and ideas on all phases of my research.

The members of our laboratory are to be acknowledged. I thank
Dr. T. Kozlowski, Roberta Milar, and Lynne Olson for their support and
technical assistance.

A very special acknowledgement is given to all members of the
'Fish lab' past and present, and, in particular, Doctors P. 0. Fromm,
J. R. Hoffert, H. L. Bergman, R. L. Walker, and Ester Brenke for their
continued support over the years. My deepest thanks are expressed to
William Jackson for his interest and assistance.

This investigation was supported in part by Grant 1-ROl-HL-17768

from the National Heart and Lung Institute.

 

ii

TABLE OF CONTENTS

Page

LIST OF TABLES.......0..0....................................... iv

LIST OF FIGURES.................................. ..... .0........ v

 

INTRODUCTION.. ............... .............. ..... ....... ..... .... 1
LITERATURE REVIEW............................................... 4
MATERIALS AND METHODS........................................... 25

Pressure-volume Experiments............................... 25

Closed Circuit N2 Experiments............................. 35
RESULTS......................................................... 41
DISCUSSION...................................................... 64

LIST OF REFERENCESOOOO..........0.........000......000000000O... 79

APPENDIX..O....................................0................ 85

iii

TABLE

1.

LIST OF TABLES

Lung volumes (liters) determined from quasi-static P—V
curves of the lung as a function of body position...........

. Values for parameters of the single breath N washout test

2

in different body positions.................

Estimated functional residual capacity (FRC, liters) given
by closed and open circuit N2 dilution experiments..........
4

for the three basic lung models investigated................

. Variance of residuals (measured FN - predicted FNZ) X 10

Estimated compartment volumes and tidal volumes (liters) as
a function of position given by the 2 compartment models
incorporating deadspace.....................................

Specific tidal volume (5, V'T/V) predicted from the model

parameters using the effective tidal volume (V'T = VT - VD)
as a function of posture....................................

iv

Page

44

47

49

55

57

6O

10.

ll.

12.

LIST OF FIGURES

A hypothetical regional pressure—volume (P—V) curve.......
Four compartmental representations of the 1ung............

Schematic representation of the experimental equipment
used for simultaneous recording of quasi-static pressure—
volume curves and single breath N2 washouts...............
Representative recordings of quasi-static lung (FTP) and
thoracic (PTT) P-V curves and the single breath N2 washout

Results of non-linear least squares fit of equation 8.....

Schematic diagram of equipment used to perform closed
CirCUit N2 equ111brations...............................O.

Representative recording of fractional N concentration

versus time during a closed circuit N equilibration and
2

subsequent open circuit N2 washin.........................

Composite expiratory quasi-static lung P-V curves deter-
mined from 8 horses in four body positions................

Effect of position on the single breath N2 washout for one

horse.............................O.......................

Dimensionless N concentration (F'N ) plotted against
breath number (n) for a horse in right lateral recumbency.

Mean distributions of specific tidal volume (g) given by
the model of Gomez for 7 horses in three body positions...

Estimated subdivisions of lung volume (V) in liters as a
function 0f bOdy POSitionooooooooooo00.0000000000000000000

Page

ll

16

27

31

34

37

4O

43

46

54

59

62

INTRODUCTION

Impairment of gas exchange during general anesthesia has been
demonstrated in both man and horse, and is associated with increases in
both venous admixture and alveolar-arterial oxygen tension difference

(P

A—aOZ) (Campbell et al., 1958; Hall et al., 1968). In man, increases

in PA_aoz have been correlated with a reduction in resting lung volume
(functional residual capacity, FRC) and an increase in trapped thoracic
gas (Don et al., 1972). A reduction in PRC has been demonstrated in
the anesthetized horse, and is primarily related to the change in body
position; standing versus lateral recumbency (McDonell and Hall, 1974).
Further, PA_802 in both anesthetized horses and man is dependent on the
mode of ventilation, with A-a gradients in positive pressure ventilated

patients being lower than in spontaneously ventilated patients

(Gillespie et al., 1969; Nunn, 1969). The dependence of PA-aO on lung

2
volume and mode of ventilation suggests that impaired gas exchange dur-

ing anesthesia may be the result of increases in regional inhomogenei-
ties of ventilation.

The lung is an elastic structure suspended in the thorax and sub-
ject to gravitational forces by virtue of its own mass. Static
gravitational stress distributions within the lung result in a vertical
gradient of pleural pressure and regional lung volumes, with upper

regions, relative to the gravitational field, being more distended than

lower regions (Milic-Emili et al., 1966). Ventilation is preferentially
distributed to upper lung units at low lung volumes while lower regions
are better ventilated at high lung volumes.

For efficient gas exchange to occur, regional ventilation (V)
must match regional blood flow (O). The distribution of perfusion is
also gravity dependent with higher regions, relative to the gravita-
tional field, being relatively underperfused (Kaneko et al., 1966).
Due to these static gravitational influences, the reduction in PRC dur-
ing anesthesia may cause a shift in ventilation toward upper, poorly
perfused lung units, resulting in an increase ventilation-perfusion

(V/Q) inequality and elevated P In addition, there is a critical

A-aOZ'

lung volume at which small airways in lower lung regions may close
(closing volume, CV). Airway closure results in a lung region where
V/O = 0 contributing to venous admixture. The presence of closed air-
ways during tidal breathing has been implicated as a major factor
limiting gas exchange in anesthetized man (Craig et al., 19713).

Regional distributions of ventilation are dependent on the trans-
mission of forces from the thoracic and abdominal surfaces through the
lung parenchyma, and also on local stress distributions between mechanic-
ally interdependent lung units (Mead et al., 1970). The change in body
position and thoracic muscle tone with anesthesia may alter lung-chest
wall interactions and contribute to ventilatory inhomogeneities (Rehder
et al., 1977).

The predominant factor influencing regional inhomogeneities of

ventilation is the distribution of gravitational stress through the

lung. Because of the large thorax, reorientation of the lung in the
gravitational field and the constraints on thoracic and diaphragmatic
movement, impaired gas exchange in the anesthetized horse may be
attributable to changes in the regional distribution of ventilation and
either intermittant or continuous airway closure in dependent lung
regions during tidal breathing. This study was designed to assess

these factors through the use of multiple and single breath indicator
dilution experiments with the application of several mathematical models

to describe lung ventilation as a function of body position.

LITERATURE REVIEW

In conscious man, alveolar—arterial oxygen tension difference
(PA-a0 ) averages 8 mm Hg with inspired oxygen concentrations of 20-30%
2

(F =0.2-O.3) (Raine and Bishop, 1963; Mellemgaard, 1966) and increases

IO2

to approximately 50 mm Hg in anesthetized subjects (Nunn et al., 1965;

Bergman, 1967). When F =1.0,

10 increases from 70 mm Hg in con-

P
2 A-aO2
scious subjects to 200 mm Hg following anesthesia, and venous admixture

(QS/QT) increases from 4% to 14% (Nunn, 1964; Mellemgaard, 1966; Bergman,

PA-aoz increases from 18

mm Hg in the awake standing animal to 37 mm Hg (FIo =0.2) immediately
2

following induction of anesthesia and lateral recumbency (Hall et al.,

1967; Marshall et al., 1969). In the horse,

1968). Alveolar-arterial oxygen tension differences in spontaneously
ventilated and positive pressure ventilated horses during halothane

anesthesia (FIo =1.0) averaged 400 mm Hg and 300 mm Hg respectively
2 . .
with a 14% shunt fraction (QS/QT) (Hall et al., 1968; Gillespie et al.,

1969). The similar magnitudes of P and QS/QT suggest that the

A-aO2

mechanism involved in impaired gas exchange with anesthesia is similar
in both man and horse.

Increased PA-aO may be the result of l) impaired alveolo-
2

capillary diffusion, 2) mismatching of ventilation (V) and perfusion (Q),
or 3) right to left pulmonary vascular shunt. Bergman (1970) demon-

strated a significant increase in PA-aO with no change in carbon
2

L\

monoxide diffusing capacity in anesthetized man, indicating that the
development of an alveolo—capillary diffusion barrier is not a major

factor impairing gas exchange during anesthesia.

In order to determine the relative contribution of V/Q inequalities

and right to left shunts to elevated P P and shunt fraction

- ’ ~
A a02 A a02

(QS/QT) are measured while breathing 100% 02. In this case, only regions

and are the result of anatomic
A-aO2
and physiologic intrapulmonary shunts. Anatomically, venous admixture

with V/Q=0 (shunt) contribute to P

occurs via arterio-venous anastomoses in the pulmonary circulation and
the bronchial and thebesian circulation. Physiologic shunts are due to
small airway closure with the subsequent equilibration of trapped

alveolar gas with mixed venous blood. The large P seen in both

A-aO2

anesthetized man and horse when FIo =l.0 indicates a major contribution
2

of intrapulmonary shunting to impaired oxygenation.

To separate the components of intrapulmonary shunting, a solution
of a relatively insoluble inert gas is administered intravenously.
Because of the low solubility, the inert gas marker evolves rapidly when
the blood contacts an airspace, and the concentration of the marker gas
in the blood leaving an alveolar capillary is essentially zero.
Therefore, any inert gas retained in the arterial blood must have been
contributed by blood flow through intrapulmonary arterio-venous anasto-
moses. Using tritium (3H2), Mellemgaard (1966) determined that less
than one-third of the 3% shunt fraction in normal man was attributable
to arterio-venous anastomoses. Further, there was an age related

increase in PA—aO with no change in blood flow through arterio-venous
2

anastomoses. While an increase in bronchial and thebesian circulation

could account for the increased P Anthonisen et a1. (1969), using

_ 9
A a02

a modified single breath N washout maneuver, reported evidence for

2
increases in small airway closure with age. These data suggest that
small airway closure is the predominant factor influencing the develop-
ment of intrapulmonary shunting in normal man.

A common feature of increasing age and general anesthesia is a
reduction in resting lung volume (FRC). In man, general anesthesia
results in a 44% decrease in PRC. Sixty percent of this reduction in
FRC is accounted for by the change from seated to supine posture, with
40% occurring after induction of anesthesia (Craig et al., 1971A,B;

Don et al., 1972).: At low lung volumes, dependent lung regions may be
subject to airway closure (Milic-Emili et al., 1966). When FRC falls
below a critical lung volume at which small airways are thought to

close (closing volume,(fi0 there is an increase in P and trapped

A-aO2

thoracic gas in both conscious and anesthetized man (Craig et al.,
1971B; Don et al., 1972). Wyche et a1. (1973) observed that elevating
FRC with positive airway pressures can improve arterial oxygenation
during anesthesia, presumably by reducing the amount of airway closure.
Functional residual capacity is reduced by 50% in the anesthetized
horse, primarily as a result of the change in body position (McDonell
and Hall, 1974). The relationship of FRC to CV has not been investi-
gated in the horse, and, as in man, may be involved in the increased

P seen with anesthesia.
A-aO2

As inspired oxygen percentage is reduced from 100%, V/Q inequali-

A—aO , with areas of low V/Q exerting the
2

greatest influence on arterial oxygenation due to the shape of the hemo-

ties start contributing to P

globin dissociation curve. The calculation of shunt fraction when

F < 1.0 assumes that increased P is solely the result of adding
102 A-aO2

mixed venous blood to ideal arterial blood (P =P ).
A02 a02

shunt fraction as FIO is reduced from 1.0 reflect an 'as if' shunt
2 . .

resulting from the addition of a component due to areas of low V/Q to

Increases in

the intrapulmonary shunt component. In normal man, QS/QT increases

from 3% to 5% when FIO is reduced from 1.0 to 0.2 (Mellemgaard, 1966).
2

Shunt fraction is also increased as alveolar P0 is reduced during
2 . O

anesthesia (Nunn et al., 1964). The combination of increases in V/Q
inequalities and intrapulmonary shunting due to small airway closure

are the major determinants of P and venous admixture associated

A-aO2

with general anesthesia. Since small airway closure is an extreme form
of regional inhomogeneity of ventilation, increases in airway closure
and the large shift in resting lung volume during anesthesia suggest
that changes in regional ventilation may be the predominant factor
determining V/O inequalities.

Regional inhomogeneities of ventilation have been investigated in
man by measuring regional concentrations of an inert gas indicator.
Following inspiration of an inert gas of known concentration, regional
inert gas concentration is given by dividing the product of the inspired

indicator concentration and the regional inspired volume by the end

inspiratory regional volume. Milic-Emili et al. (1966) arranged 12

scintillation counters at various levels behind the thorax of seated
men, and recorded regional count rates of 133Xe introduced in the
inspirate. By knowing the inspired and regional 133Xe concentrations,
the topographic distribution of regional lung volumes as a function of
overall lung volume can be determined. At any given lung volume,
apical regions were found to be relatively distended over more basal
areas, and ventilation was preferentially distributed to apical units
at low lung volumes while basal units were better ventilated at high
lung volumes. Topographic distributions of regional volume determined
in prone, supine, and laterally recumbent man by the same technique
demonstrate that the distribution of regional lung volume is dependent
on the orientation of the lung in the gravitational field, with upper
(relative to the gravitational field) lung regions always relatively
more distended than lower regions (Kaneko et al., 1966). Direct histo-
logic measurement of alveolar size in dog lungs frozen in situ also
demonstrates a vertical distribution of volume relative to the gravita-
tional field which can be accentuated by increases in gravitational
acceleration (Glazier et al., 1967).

In any plane taken through the lung normal to the gravitational
field, alveoli within the plane are subjected to a distending force
proportional to the mass of the lung below the plane and the cross
sectional area of the lung at this level. By measuring cross sectional
area of serial transverse sections and the weight of the lung below
each section in a dog lung frozen in situ, Glazier et a1. (1967) pre-

dicted the regional distending force due to gravity. This regional

distending force correlated well with their determinations of the verti-
cal (relative to the gravitational field) gradient of alveolar size.
Since the lung is mechanically interdependent with the thorax, regional

L
vertical gradients of PPL (Krueger et al., 1961) and regional lung

distending forces are reflected by pleural pressure (PP ). Thus,

volume are a direct consequence of static gravitational stress distribu-
tions within the lung.

If the lung is considered as a large number of gas exchange units
each exhibiting a characteristic pressure-volume (P-V) relationship
similar to the whole lung, the influence of pleural pressure gradients
on regional ventilation is illustrated in Figure l. Ventilation will
depend on both the resting P

PL and the change in PPL (A PPL). By virtue

of the gravity dependent gradients of P and regional volume, lung

PL

regions will operate at different points along the P-V curve. If each

unit is subjected to the same A P L’ optimum ventilation occurs when

P
the lung acts as a single homogenous unit operating at the inflection
point of the P—V curve (dV/dP = maximum). Regional inhomogeneities of
ventilation develop when the lung deviates from this homogenous state.
Ventilation is now determined by the distribution of units along the
P—V curve, with ventilatory inhomogeneities dependent on both the width
of the distribution and the mean Operating point of the lung. At low
lung volumes, apical lung regions may operate near the inflection point
(A Vr ) while basal areas are near their regional residual volume (er)

2

(A Vr ). As overall lung volume is increased, each unit moves along the
l
P-V curve with apical regions approaching their regional total lung

10

.uQMHomuw mam Hmofiuuo> oHumuw ecu mo

msuua> he m>u=o >Im ecu wcoam musaoa mamuommav um wnfiuMummo msowwou

m

mesa mounu you vmumuumsaafi mum Hm < sm>fiw m now Au> <v mesao> Hmku

aw mGOfiumflum> Hmaoawmm .ONm Eu aw Amen <v whammoum mumcoaasamawuu
Ca mwcmso msmuo> AHUAHV muflommmo mesa Hmuou Hmaowwou ou Au>mv madao>

Hmsvflmou Hmaowwwu Scum mafia: humuufinum ca wouuoaa ma oadao> Hmcowmom

.w>M:o A>Imv madao>lou=mmmum Hm:o«wmu HmUfluonuonhs <

.H munmflm

11

m
c><

H ounwwm
AV.N—L .hho ambfim AOV
m__ o 2-

:2 fl .2

>6

 

..
I 0.2.

12

capacity (TLCr) (A Vr3) and basal areas approaching their regional
inflection point (A Vrz) resulting in a progressive shift in the tidal
volume toward the basal lung units.

This simple model can account for shifts in regional ventilation
with lung volume (Milic-Emili et al., 1966) on the basis of static

gravitational stresses, but the actual change in P seen by a given

PL
region is dependent on the interaction of the lung and chest wall, and
may not be constant for all regions. Pleural surface pressure has been
measured in rabbits and dogs without the introduction of a cannula into
the pleural space by placing a clear, rigid capsule over an exposed

area of the parietal pleura and adjusting capsule pressure to remove

any visible deformations in the local lung surface (Agostoni et al.,
1970). The distribution of pleural pressure is dependent on body posi-
tion, the gradient of transdiaphragmatic pressure and chest wall con—
formation (Agostoni et al., 1970; Agostoni and D'Angelo, 1971).

In particular, diaphragmatic position and visceral contents exert a
major effect on pleural surface pressures, with evisceration resulting
in a 2-3 fold reduction in the vertical PPL gradient. In laterally
recumbent man, voluntary diaphragmatic contraction reduces the vertical
differences in 133Xe concentrations, reflecting a more uniform distribu-
tion of regional volumes and ventilation (Roussos et al., 1977). This
shift in the regional volume distribution may be the result of a change

in diaphragmatic conformation and the gradient of PP during voluntary

L
diaphragmatic contraction. Rehder et al. (1977) used the same technique

to assess changes in regional ventilatory inhomogeneities following

l3

anesthesia and paralysis in man. In lateral recumbency, anesthetized-
paralyzed mechanically ventilated subjects show greater regional differ-
ences in 133Xe concentration than awake spontaneously breathing man.
Thus, changes in chest wall shape and alterations in the transmission

of forces between the lung and chest wall appear to be major factors
influencing the regional distribution of ventilation.

While lung and chest wall interactions determine intrapleural
pressure, lung expansion is dependent on the transmission of PPL through
the lung parenchyma. Mead et al. (1970) modelled lung expansion as a
sum of all distending and recoil stresses acting on a lung unit. When
the lung is expanded homogeneously, net distending force on inter-
dependent airspaces depends directly on pleural pressure. When inter-
dependent regions do not expand uniformly, the effective distending
pressures acting on the regions reflect the degree of mechanical inter-

dependence and no longer bear a direct relation to P Mechanical

PL'
interdependence acts to reduce local inhomogeneities of ventilation by
changing the stress distributions between interdependent regions. The
distribution of ventilation within the lung is then a function of static
gravitational stresses, the transmission of forces from the thoracic and
diaphragmatic surfaces through the lung parenchyma, and local stress
distributions between mechanically interdependent structures.

The uptake or clearance of an inert gas by ventilation has also
been used to assess distributions of lung volume. If the lung is

assumed to be a single well mixed compartment of volume V and tidal

volume VT (Figure 2A), the expired concentration of an inert gas such as

14

N2 distributed in the lung obeys the following difference equation when

the inspired indicator concentration is zero:

 

V
F (n+1) = F (n) (1)
NL N2 [v+vT]
where:
n = breath number
FN (n) = fractional expired N2 concentration at breath n

2

Equation 1 may be written in terms of a dimensionless N2 concentration

(F'N ), and in general form is given by (Appendix):

 

2
FN2 (n) - FN2 (00) v 11
I = -_- _.._...
FN “0 F (0) - F (00) v+v (2)
2 N2 N2 T
where:
F'N (n) = dimensionless N2 concentration at breath n
2
F (0), F (n), F (00) = fractional N concentration at breath
N2 N2 N2 2

0, n, and 00 respectively.

Deviations of expired N2 concentration from the relationship given by
equation 2 may be, in part, accounted for by the presence of more than
one compartment. The model can be extended to k parallel, independent
compartments (k = 2, Figure 2B) by the addition of a weighting factor

to account for differences in compartmental tidal volumes (VT ), and
1
assuming all k compartments start at the same initial condition (FN (0)),
2

can be shown to follow:

Figure 2.

15

Four compartmental representations of the lung.

A) Single compartment, no deadspace.

B) Two compartment, no deadspace.

C) Single compartment with deadspace.

D) Two compartment, common deadspace.

Compartments are assumed to be instantaneously mixed, with
the volume and tidal volume of the compartment represented
by V and VT respectively. Deadspace volume (VD) is
assumed to have a flat velocity profile and undergo no

diffusive mixing with the compartments.

 

 

 

 

 

 

 

 

 

 

l6

 

 

 

 

~r

 

l l

 

h--

 

 

 

 

 

 

 

 

Figure 2

 

 

 

 

 

 

[ l

l7

1 k Vi n
F' (n) = -' Z V -—-—-—- (3)
N2 VT i=1 Ti V1+V i
where:
k
V = Z V
T i=1 Ti

This model separates the lung into functionally different compartments
based on their clearance rates. The number of unique compartments that
can be resolved is limited by the method of analysis. Numerical solu-
tions have been used to resolve up to four compartments and analog simu-
1ation has been used to define as many as six independent compartments
(Fowler et al., 1952; Hashimoto et al., 1967).

One major criticism of these models is the failure to incorporate
a deadspace volume (VD). The addition of deadspace to the model of
Figure 2A is shown in Figure 2C. Again, the compartment is assumed to
be instantaneously well mixed. In addition, flow within VD is assumed
to have a flat velocity profile (plug flow) with no diffusive mixing

occurring between V and V. Because of the initial expiration of dead-

D
space gas at the inspired N2 concentration, F'N (n) is now based on an
2
'end-tidal' sample and is given by:
v+vD n
I a __
F N (n) v+v (4)

2
This equation was presented by Darling et a1. (1944). The extension to
k parallel compartments with independent deadspaces is similar to the
previous case (Appendix). This multiple compartment model reveals a

significant amount of uneven ventilation in normal man, and an increase

18

in these ventilatory inhomogeneities with obstructive pulmonary disease
(Fowler et al., 1952; Hashimoto et al., 1967; Tsunodo et al., 1972).

An alternative to assuming an independent deadspace associated
with each compartment is the addition of a common deadspace to a two

compartment system (Figure 2D) as developed by Weber and Bouhuys (1959)

 

 

 

 

(Appendix):
v V k m n 1
. T1 T2 0 o
F' (n) = -— —— (5)
N2 VT VT po qo 1
where: V 2 V V
T T T
Vl+ Vl VD V:L V2 VD
k = T m _ T T
0 V + V 0 V + V
1 T1 1 T1
2
V V V
T1> T2 v v2+ T2 v
p - VT VT D q = VT D
0 V + V 0 V + V

In the presence of a common mixing chamber (VD), clearance of the inert
gas from the two compartments are interdependent, and clearance rates
for the individual compartments are no longer exponential; better venti-
lated unit clearances being relatively retarded and poorly ventilated
unit clearances enhanced. As a result, indicator clearance, as measured
by mixed end-expired concentration at the mouth, appears more uniform
with increasing VD.

In contrast to approximating the lung as a small finite number of
compartments, two major approaches have been proposed to model the lung

as a continuous distribution of some characteristic defining the

19

indicator clearance. The R compartment model (equation 3) is extended
to an integral equation by allowing k to approach infinity and compart-
ment size to approach zero. Nakamura et a1. (1966) and Okubo and
Lenfant (1968) have modelled the lung washout function as a continuous
distribution of clearance time constants. This technique makes the
initial assumption that indicator washout is continuous in time. All
models for indicator clearance assume that the parameters governing the
washout are constant. In the case of the difference equations 1-5, one
primary constraint is on the constancy of tidal volume (VT)° When the
washout is defined as a function in time rather than breath number,
both depth (VT) and respiratory rate must be constant. The spontaneous-
ly ventilating anesthetized horse exhibits greater variations in

respiratory rate than V during the washout maneuvers, making analysis

T
in time subject to large errors. The equations of Gomez (1963) take
into account the discontinuous nature of ventilation by modelling the
lung as a distribution of specific tidal volume (VT/V, E). Gomez and
Filler (1966) proposed approximating the expired indicator concentration

during an open circuit washout with the equation:

1

' =
F N2(n) (1+an)z (6)
where:
a,z = empirical constants
n = breath number

By solving for the parameters a and z with a generalized non-linear
least squares algorithm (Bevington, 1969) and evaluating the inverse

Laplace transform of equation 6, a continuous distribution in terms of

20

relative number of units as a function of E is obtained (Appendix).
Characteristics of the distribution (mean, mode, half-width) have been
used to determine changes in the distribution of ventilation in both
normal man and patients with pulmonary pathologies (Gomez et al., 1964;
Filler and Gomez, 1966). There have been no studies where open circuit
multiple breath indicator dilutions have been used to determine ventila-
tory inhomogeneities during anesthesia.

The lung models described above have been primarily concerned with
inhomogeneities of lung volume and tidal volume, and do not account for
regional differences in emptying rates; i.e., sequential or asynchronous
ventilation. Fowler (1949) investigated uneven ventilation using a

single breath of 100% O and noted three phases in expired FN :

2
l) deadspace gas (phase I), 2) a deadspace-alveolar boundary (phase II),
and 3) an alveolar plateau (phase III). Asynchronous ventilation results
in a non-zero slope of the alveolar plateau, and depends on regional
differences in both indicator concentration and emptying rates (Fowler,
1949). Milic-Emili et al. (1966), measuring regional inhomogeneities of
133Xe distributions, demonstrated that upper lung units are relatively
better ventilated than lower units near RV, with the opposite case
found near TLC. As a result, boli of 133Xe inhaled at RV are prefer-
entially distributed to upper lung regions during a vital capacity (VC)
inspiration (Anthonisen et al., 1970). The pattern of lung emptying
from TLC is reversed from the direction of filling, and a positive slope

of the alveolar plateau is seen as non-dependent, indicator rich regions

make progressively greater contributions to the expirate. By rotating

21

both standing and laterally recumbent subjects 1800 after receiving a
bolus of 133Xe during a VG inspiration, the vertical distribution of
indicator is inverted relative to the gravitational field. The slope
of phase III on the subsequent expiration is negative, and demonstrates
that regional asynchronous lung emptying proceeds from the bottom to
the top of the lung relative to the gravitational field, with the slope
of the alveolar plateau dependent on the vertical distribution of the
indicator at TLC (Anthonisen et al., 1970). Isolated dog lungs show
similar behavior, with the regional distribution of 133Xe dependent on
the vertical gradient of transpulmonary pressure (Glaister et al.,

1973B). In man, single breath N washouts performed under zero gravity

2
show little or no slope to the alveolar plateau (Michele and West,
1977), confirming that asynchronous ventilation is a result of the ver-
tical distribution of gravitational stresses within the lung.

The single breath dilution experiment has also been used to evalu-
ate small airway closure in dependent lung regions (Anthonisen et al.,

1969). Following a VG inspriation of 100% 02 from RV, regional N con-

2

centration varies through the lung, with N preferentially distributed

2
to the upper lung regions. During the subsequent exhalation to RV,
there may be a relatively abrupt change in the slope of the alveolar
plateau, termed phase IV (Dollfuss et al., 1967), with the lung volume
above RV at the onset of phase IV designated as closing volume (CV).
From measurements of regional lung volume with inspired 133Xe, dependent

lung regions have been shown to reach their regional RV prior to the

lung as a whole, indicating that basal lung units are either subject to

22

airway closure or their compliance approaches zero at low lung volumes
(Milic-Emili et al., 1966).

If airway closure occurs, regions subjected to closure will trap
gas and tend to develop atelectasis. The closed airway will also
exhibit a critical opening pressure necessary to reopen the airway
dependent on its radius and the surface tension of the airway fluid
(Mead et al., 1957). By removing N2 from any trapped gas, the develop-
ment of absorption atelectasis is enhanced. Burger and Macklem (1968)
measured increases in lung elastic recoil in human subjects following
complete N2 washout with 100% O2 and tidal breathing at RV or breath
holding at RV, and revealed a significant amount of airway closure below
20% VC. When the lung volume is equilibrated with an inspired indicator,
indicator concentration at equilibrium is dependent on the volume of
thoracic gas communicating with the mouth. Any gas trapped behind
closed airways will not equilibrate with the indicator. If the lung is
now inflated to a high lung volume such that the closed airways open,
the sequestered trapped gas will cause a proportionate decrease in
indicator concentration. Using this method, small airway closure has
been demonstrated in both conscious and anesthetized man (Burger and
Macklem, 1968; Don et al., 1972). If N20 is distributed throughout the
lung, Engel et al. (1975) observed that a 133Xe bolus at the mouth will
be drawn into the lungs as N

2
open glottis, and the 133Xe will be distributed according to regional

O is absorbed during a breath hold with an

blood flow provided the airways are patent. Regional 133Xe concentra-

tions following a bolus of 133Xe inhaled or given intravenously provide

23

indices of regional ventilation and perfusion respectively. At low
lung volumes, dependent lung regions were found to receive less

inhaled 133Xe than predicted on the basis of their regional blood flow.
This was attributed to the presence of airway closure preventing the
replacement of absorbed N20 with 133Xe (Engel et al., 1975).

While airway closure is a sufficient condition to account for the

presence of phase IV during the single breath N washout, the only

2
requirement for phase IV to occur is a change in regional contributions
to the expirate. The lung operates within a pressure-flow-volume rela-
tionship bounded by a maximal flow-volume curve. At any given lung
volume, expiratory flow is limited by dynamic airway compression and
becomes independent of effort (Fry et al., 1958). Since dependent lung
regions reach their regional RV first (Milic-Emili et al., 1966), for a
given expiratory flow rate, dependent lung units may begin operating on
their maximal flow-volume curve prior to the remainder of the lung, and
make progressively smaller contributions to the total expirate due to
flow limitation. Regional flow limitation was advanced by Hyatt et a1.
(1973) and Rodarte et al. (1975) to explain the flow dependence of CV
in man. In contrast to man, the onset of phase IV is relatively flow
independent in the dog (Lai et al., 1977), and the contribution of this
mechanism to the development of phase IV may be species specific.

However, the presence of phase IV during a single breath N washout,

2

either by virtue of small airway closure or regional flow limitation,

does reflect an impairment of ventilation in dependent lung regions.

24

Regional inhomogeneities of ventilation have been demonstrated
in both conscious and anesthetized man. These include regional differ-
ences in volumes and tidal volumes, and asynchronous ventilation between
lung regions. The distribution of ventilation is dependent on gravita-
tional stress distributions, lung-chest wall interactions, and local
stress distributions between mechanically interdependent lung units.
Uneven ventilation has been implicated as a factor influencing gas
exchange during anesthesia. In particular, airway closure has been
cited as a major cause of impaired arterial oxygenation in anesthetized
man. Further, a significant portion of the changes in gas exchange
with anesthesia in man are related to postural changes. Since the
mechanisms influencing gas exchange during general anesthesia may be
similar for man and horse, inhomogeneities of ventilation were investi-

gated using multiple and single breath N dilution tests, and the dis-

2

tribution of ventilation determined as a function of body position.

MATERIALS AND METHODS

Pressure—volume Experiments

 

Eight horses weighing 395-541 kg (E'= 485 kg) were restrained,
and a stomach tube with a 10 cm long esophageal balloon (Anode Rubber
Plating Co., Boling, Texas) attached to its outer wall was passed into
the esophagus via the nares. The balloon was positioned in the mid-
thorax by visually approximating the distance from the nares to the 10th
intercostal space. Animals were subsequently anesthetized (2% glycerol
guaiacolate and 0.2% sodium thiamylal, i.v.) and placed into lateral
recumbency. The trachea was intubated and anesthesia was maintained by
a continuous infusion of the anesthetic solution with depth adjusted to
maintain a palpebral reflex.

For measurement of esophageal pressure, the lumen of the stomach
tube was sealed and the volume of the esophageal balloon adjusted to
0.5 ml. Esophageal pressure was monitored and the balloon repositioned,
if necessary, to give good dynamic response and a negative excursion on
inspiration. Changes in pleural pressure (PPL), as reflected by
esophageal pressure, were measured with a differential pressure trans-
ducer (PM131, Statham Instruments, Hato Rey, Puerto Rico) calibrated

)

against a water manometer. Transpulmonary (FTP) and transthoracic (PTT

pressure were recorded by measuring P with respect to either mouth

PL
(PM) or atmospheric (PB) pressure (Figure 3).

25

 

26

.mumvuooou wlx o3u co cowumuucoocoo Nz was Ammlqmmv

ousmwoua oaomuonumamuu no AqmmIva whammmua unmaoaaammcmuu umonwm vmuuoam use
aoumzm soumuwmucalumosvmcmuu whommmualnnmuwonomuoadocm nouns: m nua3 unknowns
mama emanao> huoumaaucm> .cunoa mzu um mo>am> has 0:0 he vmumummom whoa mow
newsman cam Ame Nooav omufiamoH .amumxm xon a« man Hosp m ou nmnomuum uoamoao

Eonom>luoahowmamuu oaomwum> m :uH3 vaHouucoo mum3 coaumufiaxm was nOfiumuwmmaH

.musosmma Nz gunmen oawcfim vow mm>uno manao>lounmmmua oHumumIHmmsv mo weanuoomu

msomamDasawm now won: unmaaflsvo acucoafluomxm as» «o cofiumucmmoummu owumaonom

.m muswfim

 

 

 

 

 

 

Y

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

28

Nitrogen concentration (F ) was monitored at the mouth with a

N2

rapidly responding N2 analyzer (Cardio-pulmonary Instrument Corporation,
Houston, Texas) having a specified linearity of 1% full scale and cali-
brated to an electronic 0% N2 derived within the analyzer and 79% N2
using room air. Inspired and expired volumes were measured by electronic
integration of the flow signal generated by a pneumotachograph (Fleisch
#4, Dynasciences, Blue Bell, Pennsylvania) and differential pressure
transducer (PMS, Statham Instruments). The pneumotachograph was heated
to prevent vapor condensation and the pneumotachograph-transducer—
integrator system was calibrated using known flows (1-6 l/sec) delivered
via a rotameter (Fischer and Porter Co., Warminster, Pennsylvania).
Quasi-static pressure-volume (P-V) curves of the lung and thorax
were recorded while simultaneously performing a single breath N2 wash—
out. Horses were studied in left lateral, right lateral,dorsa1, and
sternal recumbency with the sequence of positions randomized by a latin
square design. At the end of a normal expiration (functional residual
capacity, FRC), the endotracheal tube was attached to a dual 100 1 bag
in box system (Figure 3) with one—way valves for separation of inspired
and expired gases. A vacuum cleaner controlled by a variable trans-
former was used to deflate the lungs to a PTP less than -30 cm H20
(defined as residual volume, RV), slowly inflate the lungs with 100% 02
to a PTP greater than 30 cm H20 (defined as total lung capacity, TLC),
and slowly return the lungs to RV. The complete maneuver was performed

in 60-90 secs. X-Y recorders (Hewlett-Packard Co., Palo Alto, Califor—

nia) were used to plot volume versus pressure and N2 concentration

29

versus volume. Five maneuvers were performed in each position, three

monitoring P and two monitoring P A sample record is shown in

TP TT’

Figure 4 with volume measured relative to RV, and PTP and PTT measured

relative to resting levels (PM = 0 cm H20). Expiratory reserve volume
(ER) was measured as the volume expired from FRC to RV, and chest wall
compliance (COW) was calculated from the slope of the recordings of

volume versus PTT over a tidal volume (5 1) started at FRC (Figure 4).

Due to differences in PTP at end inflation both between animals

and between successive maneuvers, a standardized treatment of the
lung P-V curve was adopted. Fifty to one hundred coordinates of lung
volume measured relative to RV and distending pressure measured rela-

tive to resting P were taken from the expiratory limb of three lung

TP
P-V curves and fitted to a general sigmoid function proposed by Murphy

and Engel (1977):

P=————-—-—+——————+k (7)

VMAX and VMIN represent the two volume asymptotes (TLC and RV respec-

tively) with k1, k2, and k3 defining the shape of the curve. Since

volume was measured with respect to RV ( the equation of Murphy

VMIN)’

and Engel was rewritten with V

MIN set to zero to express volume as a

function of pressure:

-b - V b2-4ac

V = 2a (8)

 

30

Figure 4. Representative recordings of quasi-static lung (PTP) and

thoracic (PTT) P-V curves and the single breath N2 washout.

Pressure—volume curves are plotted as volume in liters above

RV (V = 0 1) versus change in distending pressure (APTP or

APTT) in cm H 0 from the resting level (PM = 0 cm H20).

2
Expiratory reserve volume (ER) was taken as the volume
exhaled from FRC (PM = 0 cm H20) to RV (V = 0 1). Chest
wall compliance (ch) was taken as the slope of the chest
wall P—V curve over a tidal volume (5 l) started at FRC.
Respiratory efforts appear as large pressure excursions away
from the visually smoothed P-V curve (dotted line). Single

breath N washout is plotted as fractional N concentration

2 2

(FN ) versus lung volume in liters above RV (V = 0 1).
2
Phases I-IV, anatomic deadspace (VD), closing volume (CV),

and slope of the alveolar plateau (phase III) are designated.

31

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Horse Il/4 left lateral
40«
V
2()-‘
liters
as
C)- .1 r
-30
1 /
4O . 1
' /
V . ,
2 o - 1 0°“
: ,/
liters .1 ' I
(7" I r 1
-30 A RN 0 cm "20 30
F
N; .-
F“2 0.2
n V
SIJL v.4 i'TA A;
phase! ‘
I
r I I I I
“0 IO W 20 30 40
V liters

Figure 4

32

where:

a - vk3-
(PP k 3) - k -k

:VMAXacZ3) 1 2

the parameters of equation 8 were determined by minimizing the residuals
(measured V - predicted V) with a non-linear least squares algorithm
(Bevington, 1969) performed by a digital computer (LSI-ll, Digital
Equipment Corporation, Maynard, Massachusetts) (Figure 5). FORTRAN source
programs for the fitting algorithm are given in Bevington (1969). The
maximum lung volume (VMAX ), vital capacity (VC) defined as the pre-

dicted volume at a distending pressure of 30 cm H20, the coordinates of

the inflection point (d2V/dP2 = 0, P ), and lung compliance (CL)

IP’ vIP
defined as the slope of the function over a tidal volume (5 I) started

at PM = 0 cm H20 were determined from the parameters of equation 8 in

each body position.
Expired N2 concentration is subdivided into four phases (Figure 4):

phase I represents anatomic deadspace, phase II is the deadspace-

alveolar boundary, phase III is an alveolar plateau, and phase IV repre-

sents a departure in F from the earlier portion of the plateau. Five

N2

FN versus volume records taken during the P-V maneuvers were analyzed
2

to determine anatomic deadspace (VD), the slope of the alveolar plateau,
and the onset of phase IV (closing volume, CV) (Figure 4). Deadspace
was estimated by bisecting phase II with respect to N2 concentration.
The slope of phase III was determined by drawing the best straight line
through the initial portion of the plateau. Closing volume was then

determined as the first point at which FN made a permanent departure
2

 

33

.musmfiw
one CH uo>fiw w :ofiumavm mo mumumamuma usow mzu Scum wmumuocmw mm3
m>u=u euooEm mLH .mocmnanomu HmumumH umoa cw mono: m mo Aon Bu 0 n Emv

Hm>wH wCMDmmH mLu 80pm Aon So .mH

mqv munmmona xumcoaasamcmuu CH
mwcmco msmum> AH o u >v >m.m>onm muouHH aw madao> mama mm wmuuoaa mm>uso

>Im mesa mouse mo mafia mucumsflaxo mam aouw cmxmu muoB mucfioa xfimimumwm

.m :OHumsvm mo uww mwumsvm ummma Homafialcoc mo muHSmmm

.m munwwm

34

 

m muswam

of .5 t a
oe om n. o om-

- C - p

30:.

row

 

.33.... 30.. ex: 020....

35

from the alveolar plateau. The distending pressure at CV (PCV) was
determined using the estimated parameters of equation 8.
The influence of position on the P-V characteristics of the lung

and the parameters of the single breath N washout were analyzed by a

2
two-way analysis of variance (ANOVA) with differences between body

positions evaluated with Student—Newman-Keuls (SNK) procedure (Sokal

and Rohlf, 1969).

Closed Circuit Nq Experiments
l.

 

Closed circuit N2 equilibrations were performed on 8 anesthetized

horses with an average weight of 488 kg (range = 380-540 kg). Anesthesia
was induced as previously described and equilibrations performed in left
lateral, right lateral, dorsal, and sternal recumbency (sequence random-
ized). The closed circuit N2 equilibrations were performed with a 120 l
spirometer-CO2 absorbant circle system (Figure 6). A known volume of
100% 02 was added to the spirometer (V

a known quantity of N

O ) such that the system contains
2

contained in a known volume (V + V ). At the

2 S 02

end of a normal expiration, the lungs contain a known concentration of

N2 (F = 0.79) in an unknown volume (functional residual capacity, FRC).
2
By connecting the horse to the closed circuit spirometer system at end

N

expiration, and maintaining spirometer volume constant by the addition

of 100% 02, the equilibrium N concentration in the lungs and spirometer

2
may be used to determine FRC:

FN (an) V

0
2 2
- V (9)
F (0) - F (00) 8
N2 N2

 

FRC =

Figure 6.

36

Schematic diagram of equipment used to perform closed circuit

N2 equilibrations. The horse was attached to a 120 1 spirom-

eter via one-way valves and expired through a CD2 absorbant.

Spirometer volume and N concentration at the mouth were

2

recorded continuously on a strip chart recorder.

 

 

 

 

 

 

 

 

 

 

37

I20 I
Spirometer

 

 

 

 

 

 

 

 

 

ISodo

 

 

 

Vl.Hne
L—_ ,

/

 

 

1

Recorder

fié

\Q.F~2

 

 

 

 

 

 

Figure 6

\e

H

"2
Analyzer

38

where:

FN (0) = initial N concentration of system

2
2
FN (m)= equilibrium N2 concentration
2
vO2 = initial volume of 100% 02 added to
spirometer
VS = volume of system deadspaces

System deadspace was substantial due to the large diameter tubing

(approximately 6 cm) and large C0 absorbant canister necessary to

2
reduce the resistance to breathing. The volume of the system deadspace
was independently measured by replacing the animal with an empty 10 1
rebreathing bag (‘FRC‘ = 0 1) and performing the closed circuit equi-

libration (VS j;SE = 36.10 i 0.26 l, n = 6). Following equilibration,
the inspiratory line was disconnected from the one-way valve at the end
of a normal expiration and the washin of N2 monitored. The spirometer
system was subsequently reconnected to the expiratory one-way valve to
record expired tidal volume during the washin. A sample record showing
the N2 equilibration followed by the N
Expired N

2 washin is shown in Figure 7.

2 concentration at breath n was expressed as a dimensionless

N2 concentration (F'Nz, equation 2, Appendix) and analyzed using the

models for 1) two compartment, no deadspace (2 Comp), 2) two compart-
ment with equal, parallel deadspace (2 Comp + VD), 3) two compartment,
common deadspace (Weber-Bouhuys) (Weber and Bouhuys, 1959), and 4) the
continuous diStribution of specific tidal volume (Gomez) (Gomez, 1963).

The influence of position on FRC and the model results was determined

using a two-way ANOVA and SNK procedure.

39

ufisouwo ammo unencquSm mam coaumupaaanvo

mafia mamum> cowumuuaoocoo

N

.aanmma N2

N2 ufisouao memoao m wafiunv

z Hmcoauomuw mo msfivuooou m>Humucomouamm

.n muawfim

40

r

239:0

coco

 

 

n ounwwm

:35 2.3.0

SE _
DI D
I J-

”p:_.

.823 23m

0 ~_\N_ once:

56

I‘D

rod

[0.0

 

RESULTS

Pressure-volume characteristics of the horse's respiratory system
were determined from quasi-static pressure-volume (P-V) curves of the
lung and thorax (Figure 4). In the initial experiments, attempts were
made to induce apnea prior to recording a P-V curve. However, it was
impossible to entrain the horses to a respirator, and the maneuvers had
to be performed in the presence of occasional respiratory efforts.
Respiratory movements produced the largest artifacts on the P-V curves
of the thOrax, shown as large pressure 'loops' away from a smooth P-V
relationship (Figure 4). Cardiogenic artifacts were also superimposed
on the P—V curve. To standardize interanimal treatment of the data,
the expiratory limb of the lung P-V curve was first visually smoothed
and coordinates of the P-V curve were then subjected to a non-linear
least squares fit to equation 8 (Methods). The results of the curve fit
to experimental data are shown in Figure 5. The predictive capability
of equation 8 was good, with a mean residual (measured V-predicted V)
of 0.1 liters and average variance of 2.7 1 for all body positions in
all horses. Composite expiratory P-V curves were computed for each body
position and are shown in Figure 8.

Several parameters of the quasi-static P-V curves are reported in
Table l. The estimated maximum lung volume (V ) calculated from

MAX
equation 8 was not altered by body position. Vital capacity (VC),

41

42

2 me

$on 5 o u .3 m

mm Bo a“ AmHmdv msammmum mumaoaasmmamuu GH mwcmau onu umcfimwm

wcwumou Eoum o
omuuoan we AH o n >V >m o>onm muouHH aw manao> .mGOHuHmoa moon snow :H

mmmuo: m Eouw nonwauouow mo>uno >lm mesa ofiumumtammac mucumuflmxo muflmonaou

.m muswfim

43

m unawwm

 

of so tn 4
0» cm 0. o o... om-
F _ - p b b 10
\ no.
«.2:
TON
>
.330
.2204 3 10m
.32.... .m
_occo.m
10¢

 

44

Table 1. Lung volumes (liters) determined from quasi-static P-V curves
of the lung as a function of body position.

 

 

 

 

Position
1t lat rt lat dorsal sternal SE
VMAX 42. 44. 59. 48. 6.9
VC 30. 31. 28.* 36. 2.4
ER 6.* 5.* 6.* 8. 0.5

 

*
Significantly different (P<<0.05) from values obtained in sternal
recumbency

SE = standard error of mean based on error mean square from ANOVA n = 8

defined as the lung volume above residual volume (RV) at a distending
pressure of 30 cm H20, and expiratory reserve volume (ER) were signifi—
cantly increased in sternal recumbency over the lateral and dorsal
postures. Due to large interanimal variations, there were no signifi-
cant changes in lung (CL) or chest wall (CCW) compliance with posture,

averaging 1.3 :;0.1 and 0.6 :_0.1 l/cm H 0 respectively (R : SE, n = 8).

2

Values similar to these for CL and ch have been reported in both awake,

standing, and anesthetized, suspended upright horses (Gillespie et al.,
1966; Leith and Gillespie, 1971).

The effect of position on the single breath N2 washout in one

animal is shown in Figure 9. Parameters defined by the single breath N2
washout were analyzed as a function of posture. Anatomic deadspace (VD)

did not vary with body position, and averaged 2.1 :_O.1 1 (E :_SE,

n = 8). There was a significant decrease in the slope of the alveolar

45

.smoumaa

N
umaoo>ao can seem zm mo annoyance unocmahma umun oeu mm vochov .>H woman mo

ummco one mmumoflwcfl Bonus wee .smmumam onu wo maoam mnu mafiaumumw ou :mmumam

umaoo>am one cwsouau csmuw mm3 mafia ucwfimuum 4 .AH o u >mv muoufla ca oasao>

N .
mesa msmuo> A rev coaumuucoocoo Nz Hmaowuomuw mm vovuoomu mos cowumuucooaoo

N

N2 conflaxm .omuo: oco MOw usozmms z fiuwmun oawcflm mam so coauflmom mo uummwm

.m muawwm

46

m ounwwm

93:. >

on ow on ON 0.
_ _ P _ _\

Lo

   
 

.oacou

1’ II, F
I ‘l l

NAu

u
u.

_o.o.o_ «.0.

I ‘ ‘1‘

.ueco.u a/l/e\)ra

¢\: mmcoz

47

plateau (phase III) in sternal versus lateral and dorsal recumbency
(Table 2). The onset of phase IV (closing volume, CV) in liters above
RV and the distending pressure at CV (PCV) calculated from the para-
meters of equation 8 (Methods) did not change with posture (Table 2).
Closing volume (Table 2) was significantly greater than ER (Table 1) in

the lateral and dorsal postures.

Table 2. Values for parameters of the single breath N washout test in

 

 

 

 

different body positions. 2
Position

lt lat rt lat dorsal sternal SE
Slope of phase
111 (FN .102.1) 0.30* 0.36* 0.35* 0.13 0.03

2

CV (l) 9. 9. 9. 9. 0.8
VIP (1) 12. 12. 12. 14. 0.8
PIP (cm H20) 4.6 2.8 4.0 0.1 1.57

 

*
Significant difference (P‘<0.05) from sternal recumbency

SE = standard error of mean based on error mean square from ANOVA n = 8

Previous investigations have suggested a strong correlation be—
tween the distending pressure and lung volume at the onset of phase IV

(PCV’ CV) and the P-V coordinates at the inflection point of the lung

P-V curve (PI VIP) (Ingram et al., 1974; Glaister et al., 19733).

P’

The coordinates of the inflection point (P ) did not vary with

IP’VIP

48

body position (Table 2). Correlation coefficients (r) were computed

between PCV and PIF P

table of critical values for r (P<10.05) (Rohlf and Sokal, 1969).

, and CV and VI and significance determined from a

Significant correlations between CV and VIP were seen in all but sternal

recumbency, while a correlation between PCV and PIF was seen only in
left lateral and sternal recumbency. If the data is combined without

respect to body position, the correlations of CV on V (r = 0.76,

IP
n = 32) and PCV on PIP (r = 0.78, n = 32) were significant.

Mean values for functional residual capacity (FRC) in the various
postures determined from the closed circuit N2 equilibration are shown
in Table 3, and compared with an estimated FRC based on the open circuit
N2 washin as predicted by five models of the lung. One animal‘s results
were excluded from the closed circuit equilibration results due to a
leak in the spirometer system, and one animal's results were excluded
from the open circuit washin results due to a very erratic breathing
pattern in dorsal recumbency. Tidal volmme (VT) did not vary with body
position (E.= 6.41, SE = 0.5 1), and FRC was significantly increased in
sternal recumbency over the other postures (Table 3).

To assess the influence of position on the distribution of venti-
lation, data from the open circuit N2 washins were subjected to a non-
linear least squares curve fitting algorithm to obtain solutions to
three basic equations. For a two compartment lung with no deadspace

(2 Comp), the dimensionless N2 concentration (F'N ) at breath n given

2
by equation 3 (k = 2, Appendix) may be rewritten as:

F‘N2(“> = (A)(B)n + (1~A>(c>“ (10)

49

Table 3. Estimated functional residual capacity (FRC, liters) given by
closed and open circuit N2 dilution experiments.

 

 

 

Open +
Circuit
Closed 2 Comp Weber- Gomez

Position Circuit + VD Bouhuys + VD 2 Comp Gomez
1t lat 12.1 11.3 11.7 12.5 18.6* 20.0*
rt lat 11.8 14.4 14.5 16.2 21.8* 23.7*
dorsal 12.0 15.8 14.5 16.5 22.6* 25.0*
sternal 24 . 7* 25. 4* 25. 5* 30. 7* 38 . 2* 45 . 8**
SE 1.5 2.1 2.4 3.3 3.5 4.8

 

*
Significantly different (P <0.05) from closed circuit, 2 Comp + V

Weber-Bouhuys, and Gomez + VD models

*Significantly different from all other positions

D,

SE = standard error of mean based on error mean square from ANOVA n = 6

+Models for lung clearance are designated in Tables 3-6 as:

2 Comp: 2 compartment model with no deadspace

2 Comp + V : 2 compartment model with equal, parallel deadspace
Weber—Bouhuys: 2 compartment model with common deadspace

Gomez: continuous distribution of specific tidal volume
Gomez + VD: Gomez model based on the effective specific tidal

' 8 —
volume (V T VT VD)

50

where:
A = V /V
T1 T
B = Vl/(Vl + VT )
l
c = v2/(v2 + vT — vT )
1
and:
V = V - V
T2 T Tl

Using the mean V measured during the washin, the estimated parameters,

T
A, B, and C, may be used to estimate FRC (V1 + V2) (2 Comp, Table 3).
The same parameters, A, B, and C, may also be applied to a two compart-

ment model assuming parallel, equal deadspaces (k = 2, equation 20,

Appendix). The parameters of equation 10 are now redefined as:

A = v /v
T1 T
B = (V1 + VD/2)/(V1 + le)
C = (v2 + vD/2)/(v2 + vT — vT )

1

The mean value for VD obtained from the single breath N washout (2.1 l)

2

was taken as the best estimate of VD, and the average VT during the

washin calculated. Given V and V V

T D’ 1’
mate of FRC (V1 + V2 + VD) may be determined from the parameters of

V V : V 9 and an esti-
2’ T1 T2

equation 10. The model (2 Comp + VD) estimates for the compartment
volumes and tidal volume are given in Table 5, and estimated FRC is

given in Table 3.

51

The model of Weber and Bouhuys (equation 6, Appendix) was found
to converge on experimental data erratically when the minimization
algorithm manipulated either the volumes and tidal volumes of the com-
partments directly or the parameters k0, mo, p0, and qo. A possible
reason for this failure to converge on a solution is that the para-
meters of equation 6 were not constant during the washin. However,
this constraint is also imposed on the other lung models and indicates
that variability in the parameters may not be responsible for the
erratic behavior of the fitting algorithm. Since V must be considered

T

to be constant throughout the washin and V can be estimated from the

D
value determined by the single breath N2 washout, the model can be re-
duced to only three unknowns: VT , V1, and V2. It should then be

1
possible to define the Weber-Bouhuys model with only three parameters

which may be easier to solve for than the original parameters ko-qo.

By dividing k0 by mO and q0 by p0, three new parameters were defined:

<
3:4

 

l-l<
HF]

<1:
t-i
N
V
<
C

N

 

<1:
|-‘
F'] F] P]
N N
V V
< <
U

52

The curve fitting algorithm was able to manipulate A, B, and C, and

converge reliably on a solution. Again, given VT and VD,

may be solved for the compartment volumes, tidal volumes, and FRC

A, B, and C

(Weber-Bouhuys, Tables 3 and 5).

Finally, the parameters a and z of equation 7 (Appendix) were
manipulated directly and the distribution of specific tidal volume
defined by equation 28 (Appendix) computed. The mean value of this

distribution ( E') corresponds to VT/FRC. Thus, knowing V FRC may be

T,
calculated (Gomez, Table 3). This model does not explicitly account for
deadspace, but by using an effective tidal volume (V'T = VT - VD) in
place of VT, a second estimate of FRC based on'E is obtained (Gomez +

V Table 3).

D’
The results of applying these models to an experimental washin
is shown in Figure 10. Visually there was no difference between the
functions used to approximate the washin. To evaluate which model best
describes the behavior of the lung, the variance of the residuals of
F'N2 were compared with an F ratio test (Sokel and Rohlf, 1969). Mean
values for the residual variance from the solutions of the three basic
equations used to define the N2 washin were not significantly different
(Table 4) indicating that all the models predict the washin function
equally well. Models extended to three and four compartments never

resulted in a better approximation of the dilution phenomena. Also,

single compartment models never gave a satisfactory fit to the data.

53

.A0 cowumnvov manao> Hmvfiu ofimwooam mo coausnwuumwo moosawucoo one AN
use .Am COHumsvov Q> coaaou :uflz vac .Aom coaumnvo .N u xv n> Hoaamumn

fiasco sues .Am cowumsco .N u xv Q> usonufiz uaoauumaaoo 03u ad "How

N
maowoe oxu cu wcwuuflw mopmnvm ammoa Homafialaoc >9 cm>Hw z.m vmuoflooun

ecu mucmmwummu w>u:o :uooEm .xocmnabomu Honouma unwfiu QH mono: o How

N

Acv monasc numoun umcwmwm vmuuoaa A z.mv coaumuuamocoo Nz mmeGOHmaoEHQ

.3 32w;

54

OH 0.3me

 

 

 

we on om o. o
a, . _ _ 00.
.3.
row.
-ns.
00..

.223 22.. m «.3. male:

55

Table 4. Variance of residuals (measured FN2 - predicted FNZ) X 104

for the three basic lung models investigated.

 

 

 

ModelI
Position 2 Comp Weber-Bouhuys Gomez
1t lat 1.19 1.18 1.44
rt lat 1.97 1.96 2.32
dorsal 2.94 2.85 4.27
sternal 1.10 1.10 1.28

 

TModel designations as given in Table 3.

Comparisons between the model estimates of FRC and FRC determined

by closed circuit N2 equilibration are given in Table 3. In all cases,

FRC was significantly elevated in sternal recumbency. The importance of

considering deadspace in modelling the open circuit N dilution function

2
is illustrated by the agreement in estimated FRC with a standard method

for determination of FRC (closed circuit equilibration) even though a
common value of VD was assumed for all animals. Without the inclusion
of a deadspace volume (2 Comp, Gomez) FRC was significantly greater than

the estimates obtained from the closed circuit N2 equilibration and lung

models incorporating VD (2 Comp + V Weber-Bouhuys, Gomez + VD).

D,

Because of the agreement between the closed circuit N equilibra-

2
tion FRC and model estimates of FRC, the effect of position on individual

compartment volumes and tidal volumes was only evaluated in the models

incorporating VD. The smaller volumed compartment (V1) of the

56

Weber-Bouhuys model demonstrated a significant increase in volume and

decrease in tidal volume (V ) in sternal recumbency over the other

T

l
postures. There was a significant increase in V1 in sternal recumbency,
but no change in VT given by the two compartment parallel, equal dead—
1

space model (2 Comp + VD). The larger compartment (V2) demonstrated no
shift in tidal volume in either two compartment deadspace model, but
had a larger estimated volume in dorsal and sternal recumbency than the
lateral postures (Table 5).

The continuous distribution function of specific tidal volume (E)
given by the model of Gomez (1963) (equation 28, Appendix) was highly
variable and showed no consistent changes with posture. Mean values
for a and 2 (equation 7, Appendix) were used to determine average dis-
tributions of E in three different body positions (Figure 11).

In assessing the distribution of ventilation, Specific tidal
volumes (VT/V’ E) were compared using the effective tidal volume
(V'T = VT — VD) for the three models incorporating VD (Table 6). The
fast compartment (largest E) was designated 51. In the parallel, equal

deadspace two compartment model, V'T is computed as VT
k k
1,2. In the Weber—Bouhuys model, V'T is computed by weighting the
k
relative contribution of V to each compartment (V' = V -
D Tk Tk
(VT /VT)(VD), k = 1,2). A mean specific tidal volume (E) is given
k

explicitly by the model of Gomez (equation 29, Appendix). There was no

- VD/2 for k =

significant shift in estimated specific tidal volumes given by the Gomez
or Weber-Bouhuys models with body position (Table 6). Specific tidal

volume of the slow compartment was significantly elevated in left lateral

57

Table 5. Estimated compartment volumes and tidal volumes (liters) as
a function of position given by the 2 compartment models+
incorporating deadspace.

 

 

 

Position
lt lat rt lat dorsal sternal SE
V
T1
2 Comp + VD 3.7 3.9 4.2 2.8 0.77
Weber-Bouhuys 4.2 6.2* 4.6 3.4 0.67
V1
2 Comp + VD 1.9* 4.0* 4.6* 5.2 0.79
Weber-Bouhuys 2.6* 4.2* 3.8* 7.2 0.89
V
T2
2 Comp + VD 2.7 3.0 2.0 3.3 0.55
Weber-Bouhuys 2.2 0.6 1.6 3.0 0.58
V2
2 Comp + VD 7.0* 7.6* 11.5 17.1 0.69
Weber-Bouhuys 6.5* 7.5* 11.2 15.4 2.09

 

*
Significantly different (P< 0.05) from values obtained in sternal
recumbency

SE = standard error of mean based on error mean square from ANOVA n = 7

+Model designations as given in Table 3

58

.va masfio> Hangs UHMfiumam umcfimwm emuuofia Axwvm >\o>v mafia:
mo mamas: w>wumamu zufiB Amomav Noaoo mo um3u ou Hmoflucmwfl ma coaumuoz
.Axwwcmam< .wm coaumnvwv mcowuwmoa xuon mmusu CH mmmuoa u now Nwaou

mo Hmvofi mam mm Cm>Hw va manfio> Hmnwu owwflommm mo mCOHuanuumfiv cmwz

.HH muswfle

 

59

HH muswqm

 

Av; mgo mgo .Vgu Ndv 030
_ L p P _ _AUAU
/ 10.0

  

_oacoo

  

_o.o.oq —4

   

_acto.m

 

60

recumbency for the two compartment parallel, equal deadspace model

(2 Comp + V Table 6).

D,

Table 6. Specific tidal volume (5, V' /V) predicted from the model
parameters using the effective tidal volume (V' = V — VD)

 

 

 

as a function of posture. T T
Position

Modelf 1t lat rt lat dorsal sternal SE
2 Comp + VD

51 1.73 1.45 0.82 0.58 0.35

g2 0.15* 0.09 0.06 0.07 0.02
Weber—Bouhuys

g1 1.29 0.95 0.90 0.60 0.22

52 0.10 0.06 0.05 0.06 0.02
Gomez + VD

E 0.38 0.31 0.27 0.23 0.06

 

*
Significant difference (P<<0.05) from all other postures

SE = standard error of mean based on error mean square from ANOVA
n = 7

.1.

Model designations as given in Table 3

By combining the results of the quasi-static P-V and single
breath N2 experiments with the closed circuit determination of FRC, sub-
divisions of lung volume in the four positions may be estimated
(Figure 12). Residual volume (RV) was obtained by subtracting ER

(Table 1) from the closed circuit determination of FRC (Table 3).

61

.>m wan Ao>v zuflumamo Hmufl> mo 83m use ma muflommwo mesa Hmuou
m>m mam A>uv mesao> wCHmoau mo Sam mnu we muwomamu waflmoau ”Ammv madfio>
m>ummmu mucumuwaxm mam Aommv mufiommmo Hmsmflmmu HmGOHuocsm :mmBumn

mocmpmmmfiv mcu mm wmcflsumumw mmB A>mv mabao> Hmnvfiwmm .COHuHmoa Avon

mo coHuocnw m mm muwuHH cfl A>v madao> mesa mo mGOflmfi>HvQDm mmumEHumm

.NH mpswam

62

_oceo.m

_oacoo

NH «Human
_uca.64
.=u_m

_oco.uu
~*oq

 

m.o.:

\/

63

Total lung capacity (TLC) and closing capacity (CC) were calculated by
adding RV to VC and CV respectively. In sternal recumbency, there was
a large increase in RV and TLC from the lateral and dorsal postures,

with FRC approaching CC.

DISCUSSION

These investigations have shown alterations in the distribution
of lung volume and ventilation as a function of body position in the
anesthetized horse. The major posturally related changes are a 50%
reduction in functional residual capacity (FRC, Table 3), a three-fold
increase in asynchronous ventilation (slope of phase III, Table 2), and
a significant increase in closing volume (CV, Table 2) relative to
expiratory reserve volume (ER, Table 1) as the animal moves from sternal
into lateral or dorsal recumbency.

General anesthesia results in increased alveolar-arterial O

2
tension differences (PA_302), increased venous admixture (08/0T) and
a reduction in FRC in both man and horse (Nunn, 1969; Don et al., 1972;
Hall et al., 1968; McDonell and Hall, 1974). In both conscious and
anesthetized man, impaired gas exchange occurs when resting lung volume
is less than the lung volume at which small airways are thought to
close. Airway closure results in a lung region with ventilation-
perfusion ratio (V/0) of zero, contributing to venous admixture and
hypoxemia. Positive airway pressures raise FRC in conscious and
anesthetized subjects, and improve arterial oxygenation in patients
whose FRC was reduced during anesthesia to 60% of the normal predicted

value (Craig et al., 1972; Wyche et al., 1973). Improved gas exchange

resulting from positive end-expiratory pressure has been attributed to

64

65

the increase in FRC. When FRC exceeds the lung volume at which airways
close, V/0 inequalities are reduced and arterial oxygenation aided.

In the present study, the relationship of resting lung volume to
small airway closure in the anesthetized horse was assessed by simul-
taneously performing a quasi—static pressure—volume (P-V) maneuver and
a modified single breath N2 washout. Following a vital capacity (VC)

inspiration of 100% O , regional N concentration varies throughout the

2 2
lung. This regional variation depends on differences in preinspiratory
regional lung volume and on regional filling rates. During the subse-
quent expiration, expired N2 concentration follows four characteristic
phases (Figure 4); 1) phase I due to anatomic deadspace, 2) phase II,
a deadspace-alveolar boundary, 3) phase III, an alveolar plateau, and

4) phase IV, any permanent deviation of N concentration from phase III.

2
The lung volume, in liters above residual volume (RV), at the onset of
phase IV has been termed the closing volume of the lung (Dollfuss et al.,
1967). Phase IV represents a significant alteration in the relative
contribution of some lung region to the expirate. This change in
relative regional emptying rate (phase IV) may be the result of:
1) small airway closure, 2) a regional compliance approaching zero, or
3) regional flow limitation.

The presence of small airway closure at low lung volumes has been
well documented (Milic-Emili et al., 1966; Burger and Macklem, 1968;
Engel et al., 1975). If a significant portion of the lung exhibits

airway closure at some critical lung volume, those units undergoing

closure cease to contribute to the expirate when lung volume falls below

66

the volume at which closure occurs. At this point, expired N concen-

2
tration makes an abrupt departure from the alveolar plateau. The onset
of phase IV corresponds to the start of true airway closure.

When an airway closes, the region undergoing closure reaches its
residual volume. Regional residual volume may also be determined by
chest wall rigidity and local intrapulmonary stress distributions such
that the effective regional compliance approaches zero (Leith and Mead,
1967; Mead et al., 1970). Again, the lung region involved ceases
contributing to the expirate, resulting in the onset of phase IV. In
this case, all airways remain patent.

The final mechanism advanced to explain phase IV of the single
breath N2 washout is regional flow limitation (Hyatt et al., 1973;
"Rodarte et al., 1975). As regional volume is reduced, regional flow
may approach the local maximal expiratory flow. At this point, expira-
tory flow from the region becomes independent of expiratory effort and
the relative contribution to the expirate from the flow limited segment
is reduced, resulting in the onset of phase IV.

The relative contribution of each of these mechanisms to the onset
of phase IV in this study is unknown. In addition, reporting a single
critical lung volume at which 'closure' may occur (closing volume, CV,
Table 2) should be interpreted with caution. If the lung were composed
of two unique compartments, the onset of phase IV, by whatever mechanism,
would appear as a sharp discontinuity in the alveolar plateau. As the

lung deviates from behaving as two well—defined compartments, 'closure'

takes on a more continuous nature throughout expiration, and smoothes

67

the onset of phase IV into a gradual transition between phases III and
IV. Therefore, the onset of phase IV is best interpreted as that lung
volume at which an appreciable portion of the lung undergoes some change
in its relative pattern of emptying. In addition, despite the mechanism
responsible, the presence of phase IV does represent a significant
impairment of ventilation to some lung region when resting lung volume
falls below CV. Since CV exceeds ER in all but sternal recumbency, a
significant impairment of ventilation to dependent lung regions may
exist during tidal breathing in the laterally and dorsally recumbent
horse.

The contribution of regional flow limitation to phase IV was
advanced to explain flow dependence of CV in man (Hyatt et al., 1973;
Rodarte et al., 1975). These investigations found CV to vary with
expiratory flow rates from 0.2 to 4.0 l/sec. However, Travis et a1.
(1973) did not find flow dependence of CV at respiratory flows below
1.4 l/sec in man, and Lai et al. (1977) reported a constant CV in the
dog with flows up to 1.0 1/sec. In the present study, the single breath
N2 maneuver was performed in 60-90 secs, and required the movement of
approximately 80 l of gas, giving a mean expiratory flow of 1-2 l/sec.
At FRC, maximal expiratory flows in man and dog range from 6-8 l/sec
(Macklem and Mead, 1968) as compared to 65-90 1/sec in the anesthetized
upright horse (Leith and Gillespie, 1971). In sternal recumbency, CV
was not significantly different from resting lung volume, indicating
that the horse was operating at only 2/65 of its dynamic flow range

(expiratory flow/maximal flow) at the onset of phase IV.

68

Flow independence of CV is seen in man and dog when the lung is deflat—
ing at less than 1/8 of its maximal expiratory flow. While regional
flow limitation cannot be ruled out, by operating at lower fractions of
the horse's dynamic flow range than those producing flow independence

of CV in man and dog, the contribution of this mechanism to the develop-
ment of phase IV in the horse is probably negligible.

While the contribution of true airway closure to the onset of
phase IV is unknown, the quasi—static lung P—V curve provides more
evidence as to the mechanism of phase IV in the horse. If an airway
closes, a fluid meniscus forms in the closed airway. Both fluid surface
tension in the meniscus and airway diameter determine a critical opening
pressure necessary to recruit the closed segment. The loss of lung
units due to airway closure also increases lung elastic recoil (Burger
and Macklem, 1968). In the present study, lung compliance (CL) was not
significantly altered by changes in position. However, there was an
increase in VC and ER (Table l) in sternal recumbency, and the mean
lung P-V curves were shifted toward lower recoil pressure in the sternal
posture. Furthermore, in many horses in dorsal and lateral recumbency,
lung volume was still increasing with distending pressures in excess of
+45 cm H20 when inflation was terminated. These findings support the
contention that there were lung regions which were closed and exhibit
high critical opening pressures. The empirical equation (equation 8)
used to predict lung P-V behavior provides an estimate of maximal lung
volume (V

MAX)'

rather than surface tension forces, and the independence of VMAX and

Maximal lung expansion is limited by tissue forces

69

posture may reflect identical lung tissue properties and the potential
for complete alveolar recruitment with higher inflation pressures.

From the single breath N2 test, the onset of phase IV exceeded
resting lung volume in all but sternal recumbency, indicating the
presence of ventilatory inhomogeneities in dependent lung regions during
tidal breathing in the lateral and dorsal postures. There was also an
associated increase in lung elastic recoil in lateral and dorsal recum-
bency. Previous investigations have reported a large shunt fraction
(08/0T) in laterally recumbent anesthetized horses (Hall et al., 1968).
These data suggest that true small airway closure is present in depend-
ent lung units in laterally and dorsally recumbent horses.

The onset of phase IV has been correlated with the pressure-

volume coordinates of the inflection point (P VIP) of the lung P-V

IP’
curve. In excised dog and monkey lungs, the pressure at CV (PCV) and
the pressure at the inflection point (PIP) were strongly correlated
(Glaister et al., 1973B). Ingram et al. (1974) reported a close agree-
ment between CV and the volume at the inflection point (VIP) in man.
These correlations suggested that the inflection point of the P—V curve
reflected the presence of small airway closure. However, Demedts et al.
(1975) and Lai et al. (1977) did not find as strong a relationship
between the P-V inflection point and the P—V coordinates at CV in man

or dogs, and in general found V P to exceed CV. In this study, there

I

was a significant correlation of P on P and CV on V but this

CV IP IP’

relationship was not as strong as the correlation reported in the stud—

ies of Glaister or Ingram. Also, the coordinates of the P-V inflection

70

point were consistently greater than the P-V coordinates at the onset
of phase IV.

The discrepency between investigations examining the relationship
of the quasi-static P-V curve to the onset of phase IV may reflect
species and interanimal differences in the regional P-V curve. The
shape of the regional P-V curve is determined by a combination of lung-
chest wall interactions, mechanical interdependence and regional surface
tension forces. If the lung P-V curve is considered as a sum of indi-
vidual regional P-V relationships the inflection point of the lung as a
whole (dV/dP = maximum) occurs when a majority of individual units are
also at their regional P—V inflection points. Glaister et al. (1973A)
modelled the excised lung P-V curve as a rising exponential function.
If regional P-V behavior follows this form, as regional pressure ap-

proaches 0 cm H 0, the regional rate of volume change (dV/dP) approaches

2

its maximum, i.e., the regional inflection point. At regional pressures
below 0 cm H20, the unit is at its regional RV, and is effectively

'closed'. Therefore, the agreement between P and PI would be expect—

CV P

ed since a majority of units reach their ‘closing‘ pressure (regional
pressure = 0 cm H20) and regional inflection point simultaneously,
forcing the onset of phase IV and an inflection point in the overall
lung P—V curve. However, if regional P—V behavior takes on a more
sigmoid shape, the maximal rate of regional volume change no longer
occurs at regional RV. As a result, the lung P—V inflection point is
no longer directly related to regional ‘closure‘ and a dissociation

between PIP and PCV’ and VIP and CV occurs. The data of Table 2 and

71

poor correlations of P on PIP’ and CV on VI reveals this dissocia-

CV P

tion, such that the lung P-V inflection point cannot be used as an
accurate predictor of small airway closure in horses.

The single breath N washout also provides an index of the amount

2
of regional inhomogeneities in lung emptying. The slope of phase III
is determined by the magnitude of differences in regional concentration
of a marker gas, and differences in regional emptying rates. Asynchron-
ous lung emptying is significantly reduced in the sternally recumbent
horse (slope phase III, Table 2). Similar changes in phase III with
position are seen in man and dog, with the slope of phase III reduced
in prone as compared to the lateral or supine postures (Cortese et al.,
1976; Lai et al., 1977).

The major factor influencing asynchronous lung ventilation is the
distribution of gravitational stresses within the lung (Anthonisen
et al., 1970; Glaister et al., 1973B; Michels and West, 1977). Glazier
et al. (1967) measured alveolar size in dog lungs frozen in situ in
both upright and inverted animals. In upright dogs, there was a 4:1
vertical gradient in alveolar size from apex to base, while inverted
dogs showed a uniform distribution of alveolar sizes. Regional distend—
ing force due to gravity in any plane taken through the lung normal to
the gravitational field is proportional to the cross sectional area of
the lung in the plane and the mass of lung below the plane. Measurements
of cross sectional area and lung weights showed a large vertical gradient
of regional distending force in the upright (anterior-posterior) dog

lung, but only a small gradient in the inverted lung (Glazier et al.,

72

1967). These findings were due to the conical shape of the dog lung,
with relatively more lung mass residing in the posterior regions.

The majority of the equine lung resides in the dorsal portion of
the thorax (Getty, 1975). In sternal recumbency, as in the inverted
dog, upper regions will have large cross sectional areas and tend to
be distended by the bulk of the lung. Successively inferior regions
will have smaller cross sections, but support proportionately less lung
mass. The net result will be a relatively homogenous distribution of
gravitational stress throughout the lung as evidenced by the small
slope to phase III (Table 2). With lateral and dorsal recumbency, the
degree of asynchronous ventilation is increased. This increase in the
slope of phase III reflects either an increase in regional N2 concentra—
tion differences or an alteration in the pattern of lung emptying. If
the pattern of lung emptying is similar between postures, increased
regional stratification of N2 must be involved. Regional filling will
be influenced by the presence and amount of airway closure, with more
closure causing greater regional differences in N2 concentration. Thus,
the increased slope of phase III may be directly linked to an increased
amount of small airway closure in the lateral and dorsal postures, sup-
porting the evidence that small airway closure is present in the lateral
and dorsal positions but reduced in sternal recumbency. Dynamic factors
such as diaphragmatic contraction (Roussos et al., 1976) and thoracic
muscle tone (Rehder et al., 1977) have been shown to alter intrapulmon-

ary regional gas distributions and may also play a role in determining

the slope of phase III in the different body positions.

73

Subdivisions of lung volume in the horse have not been well
investigated. In anesthetized, suspended upright horses, FRC deter-
mined plethysmographically averaged 19 l, with an ER of 9 1 and VC of
33 1 (Leith and Gillespie, 1971). In the sternally recumbent horse in
the present study, FRC measured by indicator dilution averaged 25 l,
with an ER of 8 l and VC of 36 1 (Tables 1 and 3). Functional residual
capacity is determined by the equilibrium position of the lung and
thorax. The weight of the horse on the sternum may cause dorso—ventral
compression and lateral thoracic expansion, accounting for the larger
FRC in the sternally recumbent animal over the suspended upright horse.
Other factors that may also contribute to the difference in FRC between
this study and the data of Leith and Gillespie are differences in recent
lung volume history, in anesthetic agent, and in plane of anesthesia.
Despite the elevated FRC, there is good agreement between ER and VC,
indicating that the sternal position may closely approximate the stand-
ing animal. Further, since plethysmographic FRC, which measures total
thoracic gas, is not markedly greater than the FRC from the N2 dilution
method, which measures gas in communication with the mouth, there is
apparently little trapped gas. This supports the conclusion that small
airway closure may not be present during tidal breathing in the sternal—
ly recumbent horse.

Functional residual capacity is reduced by 50% in the lateral and
dorsal postures as compared with awake standing or sternally recumbent
animals (Table 3, McDonell and Hall, 1974). This may result from 1) a

change in the equilibrium position of the lung and thorax, or 2) an

74

artifact due to a failure to measure gas trapped behind closed airways.
Radiographs of the human and equine thorax in lateral recumbency show
reductions in lung area due to compression of the down lung by the
viscera and mediastinal structures (Nunn, 1969; McDonell, personal com-
munication). In man, the change from seated to supine posture accounted
for 60% of the reduction in FRC seen during anesthetia, with 40% occur-
ring after induction of anesthesia (Craig et al., 1971A,B; Don et al.,
1972). There was very little trapped gas in conscious subjects, and the
increase in trapped gas with anesthesia represented less than a 10%
artifact in the determination of FRC (Don et al., 1972).

Since FRC and ER were measured under different conditions and with
a different population of horses, the extrapolation of absolute lung
volumes from the two groups of horses must be cautioned (Figure 12).
Although it appears that there is a large increase in RV in sternal
posture, it is not possible to determine how much of this increase is
artifactual. Because of the effects of lung volume history and P-V
hysteresis on lung elastic recoil, FRC may correspond to any lung volume
between RV and the resting lung volume on the deflation limb of the
P-V curve. For this reason, comparison of RV, TLC, and the absolute
lung volume at the onset of phase IV (closing capacity, CC) between
positions is difficult. However, the effect of position on FRC, and
estimates of RV, CC, and TLC probably reflects changes in regional
stress distributions through the lung and lung-chest wall interactions

in the different positions.

75

Regional inhomogeneities of lung volumes and tidal volumes were

assessed by multiple breath open circuit N2 washins. Three basic equa—

tions were solved by non-linear least squares minimization of the

residuals (measured—predicted) to predict the N dilution phenomena.

2
The estimated FRC given by five models of the lung were compared with

the FRC determined by closed circuit N equilibration (Table 3). The

2
incorporation of a deadspace volume (VD) was essential for accurate

prediction of FRC. The three models accounting for V 1) two compart-

D’
ment with parallel, equal VD, 2) two compartment with common VD (Weber
and Bouhuys, 1959), and 3) continuous distribution of specific tidal
volume (Gomez, 1963), are equivalent as judged by their ability to

reduce the variance of residuals (Table 4). Mixed expired N2 concentra-
tion is apparently relatively insensitive to functional differences in
lung clearance, pointing out a major problem in interpreting the results
from the N2 washin: identical mathematical equations may be interpreted
in markedly different fashion (two compartment with and without VD), and
very different mathematical formulations predict the data equally well
(compartmental versus a continuous distribution). Due to this limita-
tion, the significance of the volume and tidal volume estimates given

in Table 5 is unclear. The models do however provide a means to accur—
ately estimate FRC by monitoring FN2 for 20 breaths without the necessity

for a closed, leak proof system or the collection of all expired gas

for a long period of time during a N washout.

2

The distribution of ventilation in each position is given by the

effective specific tidal volume (5), and represents the volume of fresh

76

= V -V ) delivered to the initial

. . I
gas (effective tidal volume, V T T D

compartment volume with each breath (Table 6). The difference in
predicted E between the common deadspace and parallel deadspace two
compartment models is small and does not show an appreciable shift in
distribution of the tidal volume with body position.

While the models employing two compartments give similar results,
the common mixing chamber of the Weber—Bouhuys model is more anatomical-
ly correct. However, this model does not include a term for asynchron-
ous lung filling and emptying, and assumes that each compartment receives
a constant fraction of deadspace gas on inspiration. If asynchronous
ventilation exists, the partitioning of deadspace gas will be altered,
with the compartment leading inspiration making a greater contribution
to deadspace gas concentration and also reinspiring a greater percentage
of deadspace gas on the subsequent breath. The limit of this case
would be two parallel units with the unit lagging inspiration not start-
ing to inspire until all the deadspace gas was reinspired by the unit
leading inspiration. This is equivalent to two independent compartments;

one with deadspace equal to V and the other with no deadspace. While

D
the bulk distribution of ventilation determined by the N washin was

2
unchanged with body position, the increased asynchronous ventilation in
lateral and dorsal recumbency does indicate a shift in regional intra-
pulmonary gas distributions.
The model of Gomez (1963) attempts to solve the dilution function

for a continuous distribution of specific tidal volumes, and is based

on an extension of the two compartment model to an infinite number of

77

parallel, equal volume units. Mean distributions of E predicted by
this model are shown in Figure 11. Parameters of the distribution
(mean, mode, half-width) have been used to evaluate changes in lung
function (Filler and Gomez, 1966). In several horses, N2 washins were
predicted by distributions that approached their maximal value as E
approached zero. This condition has been seen in obstructive lung dis-
ease (Gomez et al., 1964), and may reflect the addition of an airway
closure component to the dilution function. In horses which did not
exhibit this type of distribution, the half-width and mode were not a
function of body position. This supports the compartmental model
results, indicating no major shift in the pattern of ventilation with
posture.

From these investigations, sternally recumbent anesthetized horses
show respiratory system pressure-volume characteristics and subdivisions
of lung volume similar to those in the upright animal. In the sternal
position, there is only a small degree of asynchronous ventilation with
the possibility of small airway closure occurring only at end-expira-
tion. In lateral and dorsal recumbency, there is a 50% reduction in
FRC, a decrease in VC, and an increase in asynchronous ventilation.

The change in VC and reduction of ER below CV indicate the possibility
of small airway closure occurring during tidal breathing. There was no
significant change with position in the distribution of lung volume or
tidal volume determined by the multiple breath N washin. The increased

2

asynchrony of ventilation and potential for small airway closure during

tidal breathing may increase V/Q inequalities and be the predominant

78

factor involved in the increased alveolar-arterial 02 tension differ-

ences and intrapulmonary shunt associated with general anesthesia in

the horse.

LIST OF REFERENCES

LIST OF REFERENCES

Agostoni, E., E. D'Angelo, M. V. Bonanni. The effect of the abdomen on
the vertical gradient of pleural surface pressure. Resp. Physiol.
8:332—346, 1970.

Agostoni, E., E. D'Angelo. Topography of pleural surface pressure dur-
ing simulation of gravity effect on abdomen. Resp. Physiol.
12:102-109, 1971.

Anthonisen, N. R., J. Danson, P. C. Robertson, W. R. D. Ross. Airway
closure as a function of age. Resp. Physiol. 8:58-65, 1969/70.

Anthonisen, N. R., P. C. Robertson, W. R. D. Ross. Gravity dependent
sequential emptying of lung regions. J. Appl. Physiol. 28:589-
595, 1970.

Bergman, N. A. Components of the alveolar-arterial oxygen tension
difference in anesthetized man. Anesthesiology 28:517-527, 1967.

Bergman, N. A. Pulmonary diffusing capacity and gas exchange during
halothane anesthesia. Anesthesiology 32:317-324, 1970.

Bevington, P. R. Data reduction and error analysis for the physical
sciences. McGraw Hill, New York, 1969.

Brown, E. S., R. P. Johnson, J. A. Clements. Pulmonary surface tension.
J. Appl. Physiol. 14:717-720, 1959.

Burger, E. J., P. Macklem. Airway closure: Demonstration by breathing
100% 02 at low lung volumes and by N2 washout. J. Appl. Physiol.
25:139-148. 1968.

Campbell, E. J. M., J. F. Nunn, B. W. Peckett. A comparison of artifi-
cial ventilation and spontaneous respiration with particular
reference to the ventilation-blood flow relationship. Br. J.
Anaesth. 166-175, 1958.

Clements, J. A., R. F. Hustead, R. P. Johnson, I. Gribetz. Pulmonary
surface tension and alveolar stability. J. Appl. Physiol.
16:444-450, 1961.

Cortese, D. A., J. R. Rodarte, K. Rehder, R. E. Hyatt. Effect of pos-
ture on the single-breath oxygen test in normal subjects.
J. Appl. Physiol. 41:474-479, 1976.

79

80

Craig, D. B., D. S. McCarthy. Airway closure and lung volumes during
breathing with maintained airway positive pressures. Anesthesi-
ology 36:540-543. 1972.

Craig, D. B., W. M. Wahba, H. F. Don. Closing volume and its relation-
ship to gas exchange in seated and supine positions. J. Appl.
Physiol. 31:717—721, 1971A.

Craig, D. B., W. M. Wahba, H. F. Don. Airway closure and lung volumes
in surgical positions. Can. Anes. Soc. J. 18:92-98, 1971B.

Darling, R. C., A. Cournand, D. W. Richards, Jr., B. Domanski. Studies
on intrapulmonary mixture of gases. V. Forms of inadequate
ventilation in normal and emphysematous lungs, analyzed by means
of breathing pure oxygen. J. Clin. Invest. 23:55-67, 1944.

Demedts, M., J. Clement, D. C. Stanescu, K. P. Van de Woestijne.

Inflection point on transpulmonary pressure-volume curves and
closing volume. J. Appl. Physiol. 38:228-235, 1975.

Dollfus, R. E., J. Milic—Emili, D. V. Bates. Regional ventilation of

the lung studied with boluses of 133 Xenon. Resp. Physiol.
2:234-246, 1967.

Don, H. F., B. D. Craig, W. M. Wahba. The measurement of gas trapped
in the lungs at functional residual capacity and the effects of
posture. Anesthesiology 35:582-590, 1971.

Don, H. F., W. M. Wahba, D. B. Craig. Airway closure, gas trapping,
and the functional residual capacity during anesthesia.
Anesthesiology 36:533-539, 1972.

Engel, L. A., A. Grassino, N. R. Anthonisen. Demonstration of airway
closure in man. J. Appl. Physiol. 38:1117-1125, 1975.

Filler, J., D. Gomez. The continuous distribution of specific tidal
volume throughout the lungs of children. Amer. Rev. Resp. Dis.
93:9-16, 1966.

Fowler, W. S. Lung function studies. III. Uneven pulmonary ventila-
tion in normal subjects and patients with pulmonary disease.
J. Appl. Physiol. 2:283-299, 1949.

Fowler, W. S., E. R. Cornish, S. S. Kety. Lung function studies. VIII.
Analysis of alveolar ventilation by pulmonary N2 clearance curves.
J. Clin. Invest. 31:40-50, 1952.

Fry, D. L. Theoretical considerations of the bronchial pressure flow
volume relationship with particular reference to the maximum
expiratory flow-volume curve. Phys. Med. Biol. 3:174-194, 1958.

81

Getty, R. Anatomy of Domestic Animals. Vol. 1. W. B. Saunders Co.,
Philadelphia, 1975.

Gillespie, J. R., W. 8. Tyler, V. E. Eberly. Pulmonary ventilation and

resistance in emphysematous and control horses. J. Appl. Physiol.
21:416-422, 1966.

Gillespie, J. R., W. 8. Tyler, L. W. Hall. Cardiopulmonary dysfunction
in anesthetized, laterally recumbent horses. Am. J. Vet. Res.
30:61—72, 1969.

Glaister, D. H., R. C. Schroter, M. F. Sudlow, J. Milic—Emili. Bulk
elastic properties of excised lungs and the effect of a transpul-
monary pressure gradient. Resp. Physiol. 17:347-364, 1973A.

Glaister, D. H., R. C. Schroter, M. F. Sudlow, J. Milic-Emili.
Transpulmonary pressure gradient and ventilation distribution in
excised lung. Resp. Physiol. 17:365-385, 1973B.

Glazier, J. B., M. B. Hughes, J. E. Maloney, J. B. West. Vertical
gradient of alveolar size in lungs of dogs frozen intact.
J. Appl. Physiol. 23:694-705, 1967.

Gomez, D. A mathematical treatment of the distribution of tidal volume
throughout the lung. Proc. Nat. Acad. Sci. U. S. 49:312-319,
1963.

Gomez, D., W. A. Briscoe, G. Cumming. Continuous distribution of
specific tidal volume throughout the lung. J. Appl. Physiol.
19:683-692, 1964.

Gomez, D., J. Filler. On the continuous distribution of specific tidal
volume throughout the lungs. Amer. Rev. Resp. Dis. 93:1-8, 1966.

Hall, L. W., J. R. Gilespie, W. S. Tyler. Alveolar—arterial oxygen
tension differences in anesthetized horses. Br. J. Anaesth.
40:560—567, 1968.

Hashimoto, T., A. C. Young, C. J. Martin. Compartmental analysis of the
distribution of gas in the lungs. J. Appl. Physiol. 23:203-209.
1967.

Hyatt, R. E., G. C. Okeson, J. R. Rodarte. Influence of expiratory
flow limitation on the pattern of lung emptying in normal man.
J. Appl. Physiol. 35:411-419, 1973.

Ingram, R. H., C. F. O'Cain, W. W. Fridy. Simultaneous quasi-static

lung pressure-volume curves and "closing volume" measurements.
J. Appl. Physiol. 36:135-141, 1974.

82

Kaneko. K., J. Milic-Emili, M. B. Dolovich, A. Dawson, D. V. Bates.
Regional distribution of ventilation and perfusion as a function
of body position. J. Appl. Physiol. 21:767-777, 1966.

Krueger, J. J., T. Bain, J. L. Patterson, Jr. Elevation gradient of
intrathoracic pressure. J. Appl. Physiol. 16:465-468, 1961.

Lai, Y., J. R. Rodarte, R. E. Hyatt. Effect of body position on lung
emptying in recumbent anesthetized dogs. J. Appl. Physiol.
43:983-987, 1977.

Leith, D. E., J. Mead. Mechanisms determining residual volume of the
lungs in normal subjects. J. Appl. Physiol. 23:221-227, 1967.

Leith, D. E., J. R. Gillespie. Respiratory mechanics of normal horses
and one with chronic obstructive lung disease. Fed. Proc.
30:556, 1971.

Macklem, P. T., J. Mead. Factors determining maximum expiratory flows
in dogs. J. Appl. Physiol. 25:159-169, 1968.

Marshall, B. E., P. J. Cohen, C. H. Klingenmaier. Pulmonary venous
admixture before, during and after halothane-oxygen anesthesia in
man. J. Appl. Physiol. 27:653-657, 1969.

McDonell, W. N., L. W. Hall. Functional residual capacity in conscious
and anesthetized horses. Brit..J. Anaesth. 46:802-803, 1974.

Mead, J., T. Takishima, D. Leith. Stress distribution in lungs:
a model of pulmonary elasticity. J. Appl. Physiol. 28:596-608,
1970.

Mead, J., J. L. Whitenberger, E. P. Radford, Jr. Surface tension as a
factor in pulmonary volume-pressure hysteresis. J. Appl. Physiol.
10:191-196, 1957.

Mellemgaard, K. The alveolar-arterial oxygen tension difference: its
size and components in normal man. Acta Physiol. Scand. 67:10-20,
1966.

Michels, D. B., J. B. West. Pulmonary function during short periods of
zero gravity. Fed. Proc. 36:468, 1977.

Milic-Emili, J., A. M. Henderson, M. B. Dolovich, D. Trop, K. Kaneko.
Regional distribution of inspired gas in the lung. J. Appl.
Physiol. 21:749-759, 1966.

Murphy, B. C., L. A. Engel. Model of the human lung pressure volume
curve. Fed. Proc. 36:468, 1977.

83

Nakamura, T., T. Takishima, T. Okubo, T. Sasaki, H. Takahashi.
Distribution function of the clearance time constant in lungs.
J. Appl. Physiol. 21:227-232, 1966.

Nunn, J. F. Factors influencing the arterial oxygen tension during
halothane anesthesia with spontaneous respiration. Brit. J.
Anaesth. 36:327—341, 1964.

Nunn, J. F. Applied Respiratory Physiology. Butterworth and Co.,
London, 1969.

Nunn, J. F., N. A. Bergman, A. J. Gleman. Factors influencing the
arterial oxygen tension during anesthesia with artificial venti—
lation. Brit. J. Anaesth. 37:898-914, 1965.

Okubo, T., C. Lanfant. Distribution function of lung volume and venti—
lation determined by lung N2 washout. J. Appl. Physiol. 24:658-
667, 1968.

Raine, J. M., J. M. Bishop. A-a difference in 02 tension and physio-
logical deadspace in normal man. J. Appl. Physiol. 18:284-288,
1963.

Rehder, K., A. D. Sessler, J. R. Rodarte. Regional intrapulmonary gas
distribution in awake and anesthetized-paralyzed man. J. Appl.
Physiol. 42:391-402, 1977.

Rodarte, J. R., R. E. Hyatt, D. A. Cortese. Influence of expiratory
flow and closing capacity at low expiratory flow rates. J. Appl.
Physiol. 39:60-65, 1975.

Rohlf, F. J., R. R. Sokal. Statistical Tables. W. H. Freeman and Co.,
San Francisco, 1969.

Roussos, C. S., D. I. M. Siegler, L. A. Engel. Influence of diaphragma-
tic contraction and expiratory flow on the pattern of lung
emptying. Resp. Physiol. 27:157-167, 1976.

Roussos, C. S., R. R. Martin, L. A. Engel. Diaphragmatic contraction
and the gradient of alveolar expansion in the lateral posture.
J. Appl. Physiol. 43:32-38, 1977.

Sokal, R. R., F. J. Rohlf. Biometry. W. H. Freeman and Co.,
San Francisco, 1969.

Travis, D. M., M. M. Green, H. F. Don. Expiratory flow rate and closing
volumes. J. Appl. Physiol. 35:626-630, 1973.

84

Weber, J., A. Bouhuys. Theoretical considerations on lung clearance.
Acta Physiol. Pharmacol. Neerlandica 8:121-136, 1959.

Wyche, M. Q., R. L. Teichner, T. Kallos, B. E. Marshall, T. C. Smith.
Effects of continuous positive pressure breathing on functional.
residual capacity and arterial oxygenation during intraabdominal
operations. Anesthesiology 38:68-74, 1973.

APPENDIX

APPENDIX

The quantity of nitrogen (QN ) in a single, well—mixed compartment

2
at any breath n is given by:
Q = F (n) ° V (11)
N2 N2
where:
FN = fractional N2 concentration at breath n
2
V = volume of compartment

Following inspiration of a tidal volume (VT) with FIN = 0, the nitrogen

2
concentration on the subsequent expiration is given by:

'—

V
F (n+1) = F (n) ---—- (1)
N2 N2 LV+VT

In general, the dilution function obeys:

v n
F (n) = F (0) -—— (12)
N2 N2 [W—VT 1

When inspired nitrogen concentration (FIN ) is not zero:

 

2
FN2(n) ' V + F1N2 - VT
F (n+1) = (13)
+
N2 V VT
Since FIN2 = FN2(00), substituting FN2(oo) for F1N2 and subtracting
FN (co) from both sides of the equation yields:
2
V
F (n + l) - F (00) = [F (n) - F (00)] -—-—- (14)
N2 N2 N2 N2 V+VT

85

86

 

Expressing F as a dimensionless N concentration (F' ), F' (n)
N2 2 N2 N2
at any breath n is then given in general form by:
FN (n) - FN (GD) - n
F' (n) = 2 2 = ——V—— (2)
N F (0) - F (00) V+V
2 N2 N2 T

Extending equation 2 to k parallel, independent compartments, the dilu-

tion of each compartment i is given by equation 1 and contributes

FiN (n) . VT nitrogen to the expirate (equation 11). The total quantity
2
of N2 expired at breath n (QtotN2(n)) is,
R
Q (n) ’ X F- (n) ° V (15)
t°th i=1 1N T1
2
and mean expired FN (n) will be
2
QtotN2(n)
FN (n) = v (16)
2 T
where: k
V = Z V
T i=1 Ti

Assuming all k compartments start at the same initial condition, and

substituting for Q (n) and FiN (n), the dilution in terms of the
2

totN2

dimensionless expired nitrogen concentration follows:

k v n
l 1
N2 VT i=1 T1 [ V1 + VT 1]

87

The addition of a deadspace volume (VD) to the single compartment

model yields:

[v+vD]
F (n + l) = F (n) -—--*-—- (17)
N2 N2 V + VT

Where FN (n) now represents the N2 concentration in the compartment, or
2

end tidal FN , and not mean expired FN . It follows:
2 2

[‘14-‘70] n
F' (n) = ‘—"‘——*— (4)
N2 v+vT

When extending this model to k independent compartments with parallel

deadspaces (VD ), the mixed expired FN (n) can be defined several ways.

i 2
By summing the quantity of N2 contributed by each compartment, the total

expired N2

F, (n) - (v .-v ) (18)

By dividing Q by VT’ the mean expired FN is obtained. However,
2 2

it is convenient to simply measure an average 'alveolar' concentration,

totN

neglecting the initial deadspace gas contribution to mean expired FN .

2
Average 'alveolar' or end tidal FN (n) is given by:
2

k
X F (n) (VT - VD ) (19)
=1 N i i

88

where:
k
VT = 1:1 VTi
k
VD - 1:1 vDi

Substituting F'

N (n) from the single compartment case for FiN (n), the

 

2 2
dimensionless mixed end tidal FN will be:
2
n
1 k Vi + VDi
F' (n) = _ Z (V - V ) (20)
N2 VT VD i=1 T1 Di Vi + VTi

The two compartment case in the presence of a common VD has been

developed by Weber and Bouhuys (1959). Each compartment is assumed to

contribute a constant fraction to the deadspace gas (FDN ) on expiration:
2
VTl VT2
FDN2<n> = T,— FiN (n) + — F2»; (n) (21)

T 2 VT 2

At the n+1 inspiration, the two compartments receive a fraction of gas

at FDN (n) with the remainder of the inspirate at FN = 0. The quantity
2 2

of gas at end inspiration in either compartment is the quantity at

expiration n plus the fraction of deadspace gas reinhaled:

V .V
T1 D
FlN (n-kl)'(VT + Vi) = FIN (n)-Vi + FDN (n) __17_-

2 i 2 2 T

; i=l,2 (22)

and :

Solving for Fi (n + 1) and
N

89

substituting for FDN (n):
2

 

 

 

 

 

2
FlN (n + l) = kO FIN (n) + m0 FZN (n)
2 2 2
FZN (n + l) = pO FIN (n) + qO FZN (n)
2 2 2
where: V \ 2
T1
v + ———-/ v
VT / D
k0 =
V + V
1 T1
\‘V
(T1<T2> VD
/
m = VT \VT
0 V + V
1 T1
V
T1 T2) v
p = VT VT,/ D
0 V + V
2 T2
(VTz 2
+ V
q 2 \VT D
+
0 V2 VT2

In matrix form the general solution can be shown to be:

n
F1N2(n) = Re m0 F1N2(0)
F2N2 (n) 00 qo F2N2 (0)/

(23)

(24)

(25)

90

The solution for the dimensionless end tidal N2 concentration

(F' = FDN ) follows:

N2 2

F' (n) =
N2

.\ q

(1:1 _ ko/
po 0;] \\1

The distribution of specific tidal volume (E) model of Gomez

(1963) arises from the general k compartment model with each compartment

having the same initial volume (V0):

E:1 = VT./Vo
l
k v v k E
1 o o i
F' (n)=— z v ————— = — 2: ———-—n (26)
N2 VT i=1 T1 Vo + "T1 VT i=1 (1 + £1)

By allowing k"@ and Vo+ 0, £1 becomes a continuous variable 5 and
yields a linear integral equation with f(€) defining the distribution

function of E:

V E.

. __g 2 €+(E.'

FN2(n)-VT J; n+5)“ ‘15 ‘27)
1

The solution for f(€) requires evaluating the inverse Laplace transform

of F'N (n). Gomez and Filler (1966) proposed approximating the washout
2

function by the empirical equation:

. _ 1
F N2(n) - (l + a-n)z (6)

 

where:

a,z = empirical constants

91

After determining a and z, the inverse Laplace transform and normaliz-
ing coefficients may be evaluated to yield Vo/V f(€) (relative number
of units) such that the area under the distribution is unity (notation

of Gomez, 1963):

 

 

V 2 - l
_g_ = ln(l + g)
where:
a = i +1
a
l
a = m
l
F(z) E --—-3
i=1 (a + i)
F(z) = gamma function of z
and the mean of the distribution (E, VT/V) is given by:
E M (29)

 

MICHIGAN STATE UNIV. LIBRARIES
\IWWWIWIW‘IHHIHIW{IHINIWIHWIWI
31293106461191