AMERICAN JOURNAL OF PHYSIOLOGY
Vol. 229, No. 4, October
1975.
Prinied
in U.S.A.
Factors affecting 0, transfer in sheep and
rabbit placenta perfused in situ
GORDON
Department
G. POWER
AND FITZGERALD
JENKINS
of Physiology, School of Medicine, Loma Linda University, Loma Linda,
POWER,GORDON G., AND FITZGERALD JENKINS.
Factors a$ecting
US transfer in sheep and rabbit placenta perfused in situ. Am. J. Physiol.
229(4):
1147-1153.
1975-111
these experiments
we have studied
three factors that affect placental
02 transfer.
The fetal artery
of an
isolated
cotyledon
of the sheep placenta
(or one of the umbilical
arteries
in flow studies in rabbits)
was perfused
in situ with blood
of varying
Pop and at different
flow rates while
the ewe was administered
varying
inspired
02 concentrations.
Measurements
were made of the Paz of inflowing
and outflowing
umbilical
blood,
and the 02 transfer
rate was calculated
by the Fick principle.
Changes
in individual
factors
could
be studied,
since most compensations
tending
to maintain
normal
02 delivery
were
not
operative
in the isolated
preparation.
Results
indicate
that a 20%
change
in umbilical
arterial
Paz, a 14q/‘, change in umbilical
blood
flow, and a 20 yO change in maternal
arterial
Paz would
be equivalent in causing
a 10% change
in placental
02 transfer.
Small
changes
in umbilical
arterial
Pea are sufficient
to maintain
the
rate of placental
02 transfer
equal to the rate of fetal consumption.
Maternal
arterial
Paz becomes
progressively
more critical
to fetal
oxygenation
as its level falls. The experimental
results
are compared
to those predicted
by a mathematical
model
of placental
exchange.
isolated
cotyledon;
umbilical
and uterine
blood
flow
and
Peg
SINCE THE PIONEERING
WORK
of Barcroft, Barron, and Huggett and their co-workers
(1-3, 13), a variety of factors
affecting placental exchange have been studied in sheep and
other animals as well as man. Many investigators
have obtained samples of blood flowing to and from the placenta
under physiologic
conditions
and sought to describe placental transfer as it operates during normal fetal life (14, 21).
These studies have contributed
substantially
to our understanding
of the basic mechanisms
of placental
transfer.
Oxygen transfer, for example, has been shown to depend on
the inflowing
oxygen tensions in uterine and umbilical
arteries, maternal
and fetal blood flow rates, and the placental
diffusing
capacity
(17). The importance
of these factors,
furthermore,
has been found to vary in different species and
at different times during gestation (I).
There
are two major drawbacks
using results of such
studies in attempting
to understand
placental
exchange
quantitatively.
In the first place, the importance
of individual factors is difficult to assess because compensations
in the
intact animal tend to mask the effects of primary
changes.
For example, at high altitude not only will maternal arterial
oxygen tension fall, but other factors, such as hemoglobin
levels and blood flow rates, will change as well, thereby
California
92354
partially
masking the efYects of the primary
change (4, 15).
And second, there is the difficulty
that venous blood samples-critical
in the analysis of placental transfer-constitute
collections
of blood draining
different
areas of the entire
placenta
which do not necessarily
accurately
reflect the
transfer taking place in any single exchange
unit. Thus,
admixture
of shunted blood (22) as well as that blood draining from units with different relationships
of maternal
and
fetal blood flows (25, 26) and diffusing capacity (23), and
oxygen consumption
(6), will distort the accuracy of venous
values (18, 24). While changes of a given factor will affect
the placental
end-capillary
values in a single unit,
the
change may not be accurately
reflected
by the venous
values+
We have sought to minimize
these problems in these experiments
by isolating a cotyledon of the placenta and perfusing its umbilical
circulation
in situ with blood of known
and accurately
controlled
flow rates and 02 tensions. We
have also varied maternal
arterial
02 tensions by administering
various
concentrations
of inspired
02. Thus, the
present data describe the consequences
of varying three of
the primary
determinants
of placental
02 transfer one at a
time while avoiding the effects of compensatory
changes in
the system. Although
the problem
of uneven distribution
could not be avoided entirely, it was reduced by studying a
single cotyledon of the placenta.
METHODS
Cross-bred
ewes were studied at 110-135 days gestation
(term about
147 days) under spinal anesthesia
(20 mg
Pontocaine)
and barbiturate
sedation
(6 mg/kg
pentobarbital).
A catheter
was inserted
into a carotid
artery
under local anesthesia
(Xylocaine)
to monitor
maternal
blood pressure and to collect samples for arterial
oxygen
tension measurements.
A tracheostomy
was performed
and
the ewe allowed to breathe air or other gas mixtures spontaneously from a Douglas bag. The abdomen
was opened
and one of the uterine horns
through
a midline
incision,
containing
a number of cotyledons was exposed. (The sheep
placenta
consists of 60-120 discrete cotyledons
implanted
over the uterine caruncle&)
The uterus was incised, and the
uterine wall lying opposite the incision was invaginated
into
the wound so as to expose one of the cotyledons
lying some
distance from the incision. Membranes
were gently stripped
from the vessels. A plastic catheter (PE-90) was inserted into
the artery, and a second catheter
(PE-90 or PE-60) was
placed in the vein supplying the single cotyledon.
The preparation is illustrated
schematically
in Fig. 1.
G.
1148
MATERNAL
PRESSURE
ART.
PUMP
11
OUTFLOW
UMBILICAL
INFLOW
AND
OUTFLOW
PRESSURE
SENSORS
1. Schematic
diagram
of perfusion
of isolated
cotyledon.
A
syringe
pump
perfuses maternal
blood at constant
rate through
an
umbilical
artery
supplying
a single cotyledon.
Inflowing
and outflowing
samples
of blood
are taken for analysis
of respiratory
gas
tensions, and pH and vascular
pressures
are monitored
with transducers. Bulk of placenta is undisturbed
and fetus remains in utero.
G.
POWER
AND
F. JENKINS
lated as the product of fetal perfusion rate (known from the
pump setting and calibration)
and the arteriovenous
02
content
difference.
Hemoglobin
dissociation
curves corrected for pH and temperature
were used to convert Paz to
oxyhemoglobin
saturation.
An average curve for the two
types of sheep hemoglobin
was used (19), and the oxygen
tension at half-saturation
for rabbit hemoglobin
was taken
as 31.6 mmHg (pH 7.4,38”C).
RESULTS
AND
DISCUSSION
FIG.
Heparinized
maternal
blood was used to perfuse the
cotyledon
after it had been equilibrated
with various gas
mixtures in swirl tonometers
at 37 “C. Perfusion was begun
at 1.6 ml/min
from a calibrated
syringe infusion
pump
(Harvard
Instrument
Co., Millis, Mass., model 972). This
flow rate, roughly one-third
of normal, was chosen as standard because it resulted
in inflow
blood pressures in the
normal range of 60 mmHg, i.e., the resistance to blood flow
was about 3 times greater than normal in the artificially
perfused organ. Settling of red cells was prevented by rotating a
magnetically
driven stirring bar inside the perfusion syringe.
Before entering
the placenta,
blood was passed through
a
cotton wool filter to remove small particulate
matter.
Collateral
and small secondary
vessels were tied off to
isolate the circulation
to the cotyledon.
This was verified by
I) temporarily
stopping
the perfusion
to test whether
the
outflow of blood ceased completely
and 2) demonstrating
in
separate control
experiments
that the rate of outflow was
equal (+5 %) to the rate of pump perfusion.
The isolated
cotyledon was returned
to within the uterine cavity and the
abdominal
incision closed.
Rabbits were studied at 27-29 days of gestation (term 30
days) under barbiturate
anesthesia (20 mg/kg pentobarbital). The fetus was delivered
through
an incision
in the
antimesenteric
border of the uterus, and one of the two
umbilical
arteries was perfused in situ while effluent blood
was collected
from an umbilical
vein. Other
umbilical
vessels were tied off. Blood was drawn from a second adult
rabbit
to provide
adequate
quantities
for perfusion.
In
rabbits, blood was passed through a temperature-controlled
water jacket at 38°C before entering
the placenta,
but in
sheep warming
of blood was not considered
necessary, since
the cotyledon lay within the maternal abdomen.
Samples of blood flowing to and from the cotyledon were
collected
in glass capillary
tubes of 0.18 ml volume,
and
maternal
arterial
samples were collected
in heparinized
glass syringes. Samples were placed in ice and later analyzed
for Paz, Pcoz, and pH with a Radiometer
blood gas analyzer
(model BMS3).
Control studies showed less than 1 mmHg
Pop change per hour in the capillary
tubes while awaiting
analysis. The reason this change was small may be attributed
to the low range of Paz under study. Corrections
were applied where necessary. Hemoglobin
content was determined
as cyanmethemoglobin.
The net 02 transfer rate across the placenta was calcu-
Eff& of varying matmid arterial 0xygcn tension. Oxygen, air,
a.nd vari .OUS Nz-air mixtures were administered
in random
sequence to six ewes for 3- to IO-min intervals. The inspired
gases achieved maternal
arterial
02 tensions ranging
from
27 to 340 mmHg. During this interval, the isolated cotyledon
was perfused at constant flow rate (1.6 ml/min)
with blood
of constant Paz (8-10 mmHg).
At the end of the interval,
maternal
arterial,
umbilical
arterial,
and umbilical
venous
blood samples were taken simultaneously
and analyzed for
respiratory
gases and pH.
Figure 2 shows the effect of varying
maternal
arterial
PO* on oxygen transfer rate. Figure 3 shows the effect of
10
02
EXCHANGE
RATE
t
8ML/MIN
l
.
.
*
L
100
MATERNAL
ARTERIAL
300
I
200
Pa2
I
(MM HGI
on maternal
arterial
POZ.
FIG. 2. Dependence
of 02 exchange
Thirty-eight
determinations
were obtained
on 6 sheep, an average of 6
determinations
from each. Curvilinear
regression
shown as solid line
shown has form:
was fit to data to minimize
sum of squares. Curve
‘ciro2 = -22.5
+ 9.38 (In Mat POZ) - 0.733 (In Mat Po~)~, which fit
data significantly
better than simple linear or quadratic
forms.
750
OUTFLOW
I2
pO2
(MM HG)
-40
MAT. PO2
0.5
-loi
l
c
1
MATERNAL
100
*
ARTERIAL
200
1
PO2
I
300
(MM HG)
FIG. 3. Dependence
of outflow
Paz on maternal
arterial
POZ. Regression curve fit to data has form: outflow
POT = - 75.2 + 36.9 (ln
Mat Paz) - 3.04 (In Mat Po~)~, a form which fit data significantly
better than linear or quadratic
regressions.
O2 TRANSFER
IN PERFUSED
1149
PLACENTA
maternal
Po2 on outflowing
(umbilical
venous) POT.Results
for all experiments
are shown as data points. Curvilinear
regressions were found to fit the data significantly
better
than linear regressions (27). The curves shown in the figures
were fitted through the data to minimize
the sum of squares,
and the form of each curve is given in the legends. A comparison of the data curves with predictions
of a theoretical
model is given below.
The average oxygen transfer rate for the single cotyledon
rose roughly 40 % as maternal PO:! increased from 100 to 340
mmHg.
Umbilical
venous Pox rose about 25 % for this
maternal
change.
These increases are not exceptionally
large, probably
due to the shape of the oxyhemoglobin
saturation
curve.
Decrements
in maternal
PO;!below normal levels became
progressively
more detrimental
to 0, transfer. At a maternal
Po2 of 50 mmHg (73 % saturation),
for example, umbilical
venous POT was about 23 mmHg, and the 02 exchange rate
had decreased to less than 0.03 ml/min,
a 40-50 % reduction. These relatively
large changes are probably
greater
than would
be observed in the intact animal or human,
since most compensatory
mechanisms
that maintain
02
transfer were not operative
in the isolated cotyledon.
Control experiments
to examine whether
these results could be
attributed
to progressive deterioration
in the function
of the
cotyledon are described below.
Effects of varying umbilical Jkw rate. Fetal vessels supplying
the cotyledon were perfused at Aow rates ranging from 0.22
to 4.0 ml/min
in sheep and from 0.05 to 2.3 ml/min
in rabbits. About
10 different flow rates were studied randomly
in each of 10 ewes and 5 rabbits.
Figure 4 shows the effects of varying fetal flow on the 02
tensions of blood leaving the cotyledon in sheep; results for
rabbits were similar. To consolidate
the results of changes
in flows in the various preparations-some
of which were
more efficient in gas exchange
than others-we
have normalized the data by plotting
02 tension measured relative
to the 02 tension measured in that preparation
at a standard
flow. Regression curves are again shown.
-2
OUTFLOW
1.5 ’
PO2
PO2 MEASURED
Po2AT
STANDARD
FLOW
:
.
l
05
FETAL
I
1
COTYLEDONARY
2
I
BLOOD
FLOW
ML/MIN
3
1
4
I
FIG. 4. Dependence
of outflow
POT on fetal cotyledonary
flow. An
average of 10 determinations
are shown for each of 10 sheep. Data are
normalized
so that different
preparations
may be compared.
Outflow
POT is expressed relative
to outflow
PO* found in that preparation
when
fetal flow was 1.6 ml/min.
Curvilinear
regression
shown through
data
is of form: outflow Po2 (relative
to standard)
= 1.33 - 0.73 (Fet
flow) + 0.18 (Fet Aow)~.
02
EXCHANGE
RATE
io2
MEASURED
to2 AT
STANDARD FLOW
FETAL
COTYLEDONARY
FLOW ML/MIN
5. Dependence
of 02 exchange
on fetal cotyledonary
flow. 02
exchange
is expressed
relative
to standard
for that preparation
when
flow was 1.6 ml/min,
a time when 02 exchange
averaged
0.05 ml
02/min.
Regression
curve is of form:
v02
(relative
to standard)
=
0.074 + 2.55 (Fet flow) - 1.62 (Fet flo~)~.
FIG.
Figure 4 shows that outflowing
Paz tended to increase at
slower flow rates. This was an anticipated
finding,
since
fetal end-capillary
Paz should increase as the ratio of maternal to fetal flow increases (17). This follows since more 02
is supplied relative to the capacity of fetal blood to transport
it away from the placenta, and therefore final equilibrium
is attained at a higher Po2. In fact, ignoring
placental
02
consumption,
fetal outflowing
PO:! should
theoretically
approach
maternal arterial Paz at very slow fetal flow rates.
This result was observed in two of the rabbits studied. In
two sheep, results were anomalous;
however,
in that outflowing PO:! fell at low flow rates. This may have been because I) al .arger fraction of the total OxYge n delivered to the
platen ta 1s metabolized
by placental tissue at low flows, and
2) distribution
becomes more uneven between fetal flow and
surface area for exchange at low flows, with a consequent
fall in the efficiency of O2 exchange. Thus, a larger fraction
of fetal flow might pass through
regions remote from the exchange areas.
Blood remains for a relatively short interval of time in the
exchanging
vessels at high flow rates. If a diffusional
limitation to 02 exchange
were present, it should have become
apparent
once a certain flow rate had been exceeded by an
abruptly
decreasing
Paz in fetal effluent blood (17), other
things being equal. The present data gave no hint of such a
change and hence do not support any diffusional
limitation
of 02 transfer.
Figure 5 shows the effect of fetal flow on the 02 exchange
rates. Again the data are normalized
so the results from
different preparations
can be included
together.
Effects of varying umbilical arterial 02 tension. The cotyledon
was perfused simultaneously
with blood from two infusion
pumps. The output
from each of the pumps was mixed
together before entering the placenta. One syringe contained
blood with a low oxygen tension (PO, = 10 mmHg, Pco:! =
40 mmHg,
pH = 7.4) and the other blood with a higher
oxygen tension
(PO, = 50 mmHg,
Pcoz = 40 mmHg,
pH = 7.4). By varying
Aow rates of the two pumps independently,
the PO, of the inflowing
blood could be varied
while total flow and Pcoz were held constant.
G. G. POWER
l
.
.
1.5
02
EXCHANGE
RATE
i0,
MEASURED
$AT
STANDARD
INFLOW
PO2
INFLOW
PO2
IOL
20I
l
30
MM HG
SO
40
I
FIG. 6. Dependence
of 02 exchange
on inflow
(umbilical
arterial)
Paz. An average
of 14 determinations
is shown from each of 5 sheep.
02 exchange
is expressed relative
to average 02 exchange
for that preparation
when inflow Paz is 15 mmHg.
A linear regression
fit data as
well as other curves, with form: vo2 (relative
to standard)
= 1.25 0.0305 (inflow PO&
OUTFLOW
PO2
PO2
AND F. JENKINS
rate. The same type of divergent
response was noted with
changes in umbilical
flows. Indeed, the only change in fetal
hemodynamics
that would seem able to increase both outflowing Paz and 02 transfer rate simultaneously
would be a
more uniform
distribution
of fetal placental
blood flow,
such that it more closely parallels maternal flow distribution
and/or the available surface for gas exchange.
Control experiments. The sequence of studying
d&rent
flow rates and 02 tensions was random and varied from one
preparation
to another in an effort to minimize
systematic
errors. To further evaluate possible errors which might have
been caused by progressive deterioration
in the function
of
the cotyledon,
we carried out a series of control experiments
in another group of five sheep. The cotyledon was perfused
at constant flow rate with the same blood for 1.5 h while the
ewe breathed
air. Inflowing
and outflowing
blood samples
were collected about every 15 min.
Results for one sheep are shown in Fig. 8 and, as can be
seen, outflowing
Paz and 02 exchange rate remained
generally constant. As an average for all sheep, the PO:! of outflowing blood remained
within 4 mmHg (k SD) of initial
levels during
the l-2 h of perfusion.
Thus, it would seem
that the 02 transfer function
of the cotyledon
was maintained within 20 % of initial levels during the time period of
the study.
After 2-3 h of perfusion, signs of extravasation
and edema
appeared
in some of the cotyledons
and inflow pressure
increased.
The experiment
was ended when these changes
were considered
to have become significant.
MEASURED
PD2 AT STANDARD
INFLOW
PO2
FIG. 7, Dependence
of outflow
POT on inflow Paz. Outflow
Paz is
expressed
relative
to average
outflow
Pas for that preparation
when
inflow Pea is 15 mmHg.
Best fit data curve is of form: outiow
POS
(relative
to standard)
= 0.85 + 0.0047 (inflow POZ) + 0.0004 (inflow
PO,) .2
An average of 14 different
02 tensions were studied in
random sequence in each of 5 sheep. The effects of varying
umbilical
Paz on transfer rate are shown in Fig. 6 and the
effect on placental outflowing
Paz is shown in Fig. 7, It may
be seen that both 02 transfer rate and outflowing
POZ are
sensitive to small changes in umbilical
arterial
Po2, For
example, a 20 % fall in umbilical
arterial Po2 below normal
(15 mmHg) resulted in a 13 % increase in 02 transfer rate.
Since umbilical
arterial
PO:! varies a few millimeters
Hg
from day to day in unanesthetized
sheep (7), the data
suggest that these changes would be associated with sizeable
variations in the transfer function of the placenta for oxygen.
It is of interest that changes in umbilical
arterial
PO:!
resulted in changes in both 02 transfer rate and outflowing
POTwith the responses occurring
in opposite directions- That
is to say, decreases in umbilical
arterial Pop increased umbilical venous Paz at the expense of lowering
the 02 transfer
COMMENT
Rdatiw
imfortance of factors aecting placental 02 transfer.
This will depend upon how sensitive 02 transfer is to changes
in various factors, their normal range of variation,
and the
time course of their change. The present experiments
characterize the dependence
of 02 transfer on maternal
arterial
Po2, fetal flow rate, and fetal inflowing
Paz. Of these three,
02 transfer is found to be most sensitive to a given percent-50
MM HG
-30
-10
-to
02
EXCHANGE
RATE
ML/MlN
.,/‘-*-‘-.,-
0-m
X IO2 5
c
01
L
20
MINUTES
I
8. Control
study showing
that
exchange
rates do not vary appreciably
FIG.
sion.
4.
L
L
inflow and outflow
during
initial 60
60
!
and 02
min of perfuPOT
O2 TRANSFER
IN
PERFUSED
1151
PLACENTA
age change in umbilical
blood flow. For example,
to accomplish a 10 % rise in 02 transfer, there would be required
a 20 % change in umbilical
artery PO*, a 20 % change in
maternal
arterial
Paz and a 14 % change in umbilical
flow
(values estimated from the slopes of the data fit curves).
Of the three factors umbilical
arterial Paz is of particular
interest because it is affected by the metabolic
rate of the
intact fetus. That is to say, inflowing
Paz in the umbilical
arteries is a consequence
of oxygen extraction
by the fetal
tissues, on the one hand, and a major determinant
of placental 02 transfer, on the other hand. The level of umbilical
arterial
Pop thereby directly links fetal needs and 02 delivery.
The present data show that a change of only 3 mmHg in
umbilical
arterial
Paz will cause a 10 % change in transfer
rate. As noted above, moment-to-moment
and day-to-day
fluctuations
in Pea of this magnitude
are measured
(7, 2 1).
These fluctuations
suggest the possibility
that umbilical
arterial
Paz is continually
seeking a level that causes placental 02 transfer to equal fetal 02 consumption,
as, of
course, it must in the long term. If this is so, adjustments
in
placental
02 transfer would require no more than the 5-10 s
required
for blood to recirculate
between
placenta
and
peripheral
tissues. Such a homeostatic
mechanism
would
not require
energy expenditure
by the fetus nor neural
control reflexes. The data, therefore, permit the hypothesis
that umbilical
arterial
Pas is a critical link in the passive
regulation
of placental
02 transfer.
While other factors are of undoubtedly
great importance
over the long term, they are probably
of lesser importance
over the short term for a variety of reasons. An example is
umbilical
blood Aow. Elsner
(personal
communication)
finds in radiotelemetry
experiments
in unanesthetized
sheep
that umbilical
blood flow remains
remarkably
constant
during
rest as well as exhausting
maternal
exercise. This
constancy
speaks against a short-term
regulatory
role for
umbilical
blood flow. Other arguments
supporting
a constancy of flow have been discussed by Faber (11) A constant flow and hence constant pressure would tend to maintain water balance between mother and fetus, other things
being equal. In brief, the fetus would avoid becoming
dehydrated during periods of increased 02 transport.
Over the
long term, of course, the situation
is altogether
different.
Here umbilical
flow increases several orders of magnitude
as
the fetus grows and 02 transport
increases in accord with
fetal needs.
Other factors affecting
placental
02 transfer probably
change too slowly to have an important
regulatory
function. For instance, changes in hemoglobin
concentration
and the oxygen affinity of fetal blood would require several
hours or days to affect 02 delivery appreciably.
Nor is there
any presently known mechanism for maternal placental flow
or other maternal adaptations
to respond to the level of fetal
oxygenation.
If oxygen requirements
of the fetus do vary over the short
term while umbilical
flow is reasonably
constant, then fetal
tissues must withstand
intervals of time when they receive
blood with less than average 02 content. One might suspect
adaptations
to maintain
02 utilization
in vital organs,
including
a redistribution
of blood flow throughout
the
body and changes in the number of functioning
capillaries
per unit of tissue. In contrast to adult tissues, these adaptations have not been extensively studied in the fetus.
Comparison with computer model predictions. We used a marhematicaf model to compare
experimental
results with those
predicted
theoretically.
The model assumes difftision of 02
between parallel
capillaries
with uniform
diameter,
equal
length, and uniform concurrent
maternal
and fetal flow as
described previously
( 12). We varied each factor one at a
time in the model while keeping others constant to assess its
effects without
compensatory
changes occurring
in the rest
of the system
In constructing
the predicted
model curves, we used
averages for the experimental
values for maternal
and fetal
hemoglobin
concentrations
(10.7 and 10.8 g/100 ml, respectively), maternal
inflowing
Paz (90 mmHg),
fetal inflowing
POT (12 mmHg) and fetal flow rate (1.6 ml/min),
placental
diffusing
capacity
(0.04 ml/min
X mmHg)
and placental tissue 02 consumption
(6 ml/min
X kg). These
values were assumed ‘to remain constant during an experiment. Because maternal
flow to the isolated cotyledon was
unknown,
we predicted
a family of curves covering a range
of maternal
flows from 0.5 to 3 ml/min,
some lo-50 % of
estimated
normal.
It was in this low range that data and
predicted results tended to agree, a not unreasonable
finding
since uterine surgery is known to lower maternal
blood flow
to the uterus. Since the computer
predictions
were for
different,
arbitrary
(and low) maternal
flows’, any agreement between
experimental
and theoretical
results is not
meaningful
in absolute terms. The predicted
trends as contrasted to experimental
data remain valid points of comparlson.
In Fig. 9 the curves fit to the experimental
data are
plotted again as solid lines, and they may be compared
to
the results predicted
by the theoretical
model shown as
interrupted
lines. In Fig. 9, A, B, and D the model predictions may be seen to follow the data satisfactorily.
In these
instances, in fact, the model predictions
fit the data as well
as the regression
curves within
statistical
limits
(27) if
maternal
flows were such as to give the best model fit. In
other instances, however,
the model proved somewhat
inadequate quantitatively.
For example, as shown in Fig. 9 E,
the model predicted
that outflowing
Paz should rise more
\L --.
1 :
I?0
MATERNAL
-
CURVE
2op
1
I,6
3.;
FET. COTYLEDON
FLOW
**.**.-.
FIT TO DATA
MODEL
PD2
r
c
1
115
FET. INFLOW
3:
PO2
PREDfCTlON
9; Comparison
of data-fit
curve,s with predictions
of a mathematical
model of placental
oxygen
exchange.
Slopes of interrupted
and continuous
lines should be compared
as a test of agreement
between experimental
data and model; comparison
in absolute terms is
not possible because arbitrary
(and low) maternal
flows were assumed
in making model predictions.
FIG.
1152
G.
sharply at low flow rates and fall more sharply at high flow
rates than it actually did. There was a small but significant
reduction
(P < 0.05) in the sum of squares between the best
data fit curve and the best model prediction
curve (assuming
maternal
flow of 1.3 ml/min).
There was also a small but
significant
difference between model and experimental
data
in Fig. 9, C and F.
Because of these differences there are likely to be errors of
inadequacies
in the assumptions
of the model. These might
include
the facts that the model takes no account
of I)
changes in the distribution
pattern of flows at different flow
rates, 2) differences in 02 con sumption
by p lacental t issue at
vario us oxygen tensions, and 3) changes in maternal flow as
fetal flow varies. Additional
studies will be needed before the
model can be improved in these regards.
Figure 9 shows that differences
between
predicted
and
experimental
results were generally small. By and large , the
present data were found compatible
with a model assuming
a concurrent
pattern of flow in exchan ge vessels with little or
limitation.
The data were no t sufficiently
n .o diffusional
closely grouped,
however,
that it was possible to exclude
crosscurrent,
countercurrent,
and other flow patterns combined with diffusional
limitations.
Thus, other models may
fit the data equally well, and many previous
theoretical
analyses have been presented (5, 9, 10, 16, 20) whose results
may be compared
with the experimental
data given here.
Advmtuges and disadvantages of isolated, perfused cotyledon.
Advantages
include 1) the blood used for perfusion contains
a constant and readily measured concentration
of respiratory gases, electrolytes,
hormones,
and vasopressors.
Thus,
the preparation
is free from systemic changes caused by
release of catecholamines,
falling pH, and low glucose levels
that are known to follow fetal surgery. 2) Umbilical
flow rate
is constant, since it comes from an infusion pump. Changes
in vascular
pressure,
therefore,
are a reliable
index of
changes in resistance to placental
blood flow, an important
advantage
in analyzing
pressure-flow
relationships.
3) The
fetal lamb remains in utero and within the maternal
abdomen, reducing
traction
on the uterine vessels and thereby
minimizing
the changes in maternal
blood flow due to
G.
POWER
AND
F. JENKINS
surgery. This is an advantage
over isolated perfused preparations
reported
previously
(8). All umbilical
flow is
directed to an “end-capillary”
vascular bed and collateral
flow, such as that supplying
surrounding
membranes,
is
eliminated.
There are several disadvantages
and these include:
1)
there is no method known to measure maternal
flow to the
isolated
cotyledon.
We attempted
to insert catheters in
these tiny maternal
vessels but were unsuccessful. Rather it
is necessary to assume that maternal
flow remains constant
throughout
an experiment,
an assumption
justified
by the
finding of nearly constan t res pira tory gas tensions in outflowing fetal blood during long con trol periods. 2) Umbilical
vessels constrict
where catheters are placed in them, and
there is a high resistance to blood flow. High flow resistance
may also have been caused by poorly understood
changes in
the microvasculature
as reported
in other artificially
perfused organs. In fact, we had to limit fetal flows to about
30 % of normal
in order to keep vascular pressures in a
normal
range. 3) When maternal
blood is used to perfuse
the umbilical
circulation,
allowance
is required
for its lesser
oxygen affinity.
With due caution
regarding
these disadvantages,
the
preparation
should prove helpful in studying transplacental
movements
not only of oxygen but also a variety of other
organic
and inorganic
substances, the possible hormonal
controls involved, and the changes during gestation.
The authors
express their appreciation
to John
Bankhead
and
Peter Yuen for expert technical
assistance.
This study was supported
by the Lalor Foundation,
United
Cerebral
Palsy Foundation,
and the National
Institute
of Child Health
and
Human
Development
Grant
HD 04394. Computational
aid was given
by the Scientific
Computation
Center,
School
of Medicine,
Loma
Linda University,
which is partially
supported
by National
Institutes
of Health
Grant
FR-00276.
This study was presented
in part at the Meeting
of the Federation
of American
Societies for Experimental
Biology
at Chicago,
Ill., in
April,
1971,
G. G. Power
is the recipient
of Public
Health
Service
Research
Career Development
Award
J-K4 H.D. 02,253.
Received
for publication
13 May
1974.
REFERENCES
J., J, A. KENNEDY,
AND M. F. MASON.
Oxygen
in the
blood of the umbilical
vessels of the sheep. J. Physiol.,
London
97:
347-356,
1940.
2. BARRON,
D. H. Some aspects of the transfer
of oxygen
across the
syndesmochorial
placenta of the sheep. Yale J. 113ioZ. Med. 24:
169490,
1951-1952.
Supplementary
observations
3. BARRON, D. H., AND G. ALEXANDER.
on the oxygen
pressure gradient
between
the maternal
and fetal
bloods of sheep. Yale J. Biol, Med. 25 : 61-66,
1952.
4, BARRON, D. H., J, METCALFE,
G. MESCHIA,
W. HUCKABEE,
A.
HELLEGERS,
AND EL PRYSTOWSKY. Adaptations
of pregnant
ewes
and their fetuses to high altitude.
Symp* Bysiol.
Effects High Altitude,
1. BARCROFT,
Interkaken,
Switzerland,
1963.
5. BARTELS,
H., AND W. MOLL.
Passage of inert substances
and oxygen in the human placenta.
P’uegers Arch. 280: 165-l 77, 1964.
6. CAMPBELL,
A. G. M., G. S. DAWES,
A. P. FISHMAN, A. I. HYMAN,
AND G, B. JAMES. The oxygen
consumption
of the placenta
and
foetal membranes
in the sheep. J. Physiol.,’
London
182: 439-464,
1966.
7. COMLINE,
R. S., AND M. SILVER. Daily
changes
in foetal and
maternal
blood of conscious
pregnant
ewes, with catheters
in
umbilical
and uterine
vessels. J. Physiol.,
London
209 : 567-586,
1970.
8. DANCIS,
J. The
perfusion
Proc, 23 : 701-784,
1964.
9. FABER,
J+ J. Application
transfer
1969.
of inert
materials
of guinea
pig placenta
of the theory
of heat
in placentas.
Circulation
in situ.
exchangers
Res. 24:
Federation
to the
22 l-234,
10. FABER, J. J., AND F. M. HART. The rabbit placenta
as an organ of
diffusion al exchange.
Circulation
Res. 19 : 8 16-833,
1966.
11. FABER,
J. J. Regulation
of fetal placental
blood flow. In : Respit-atory Gas Exchange
in the Placenta,
edited by L. D. Longo
and H.
BarteIs. Bethesda,
Md. : National
Institutes
of Health,
1972.
12. HILL,
E,, P., G, G. POWER,
AND L. D. LONGO. A mathematical
model of placental
02 transfer
with consideration
of hemoglobin
reaction
rates. Am. J. Physiol.
222 : 72 l-729,
1972.
13, HUGGETT, A. ST.G. Foetal blood-gas
tensions and gas transfusion
through
the placenta
of the goat. J. Physiol.,
London
62: 373-384,
1927.
14. KAISER, I. H., AND J* N. CUMMINGS.
Hydrogen
ion and hemoglobin concentration,
carbon dioxide
and oxygen content
of blood
of the pregnant
ewe and fetal lamb. J. A#.
Physiol.
10: 484-492,
1957,
15. KAISER, I. H., J, N. CUMMINGS,
S. R. M. REYNOLDS,
AND J. P.
MARBARGER.
Acclimatization
response of the pregnant
ewe and
O2 TRANSFER
16.
17.
18.
19.
20.
21.
IN
PERFUSED
1153
PLACENTA
fetal lamb to diminished
ambient
pressure.
J. &Z.
Physiol.
13 :
171-178,
1958.
LAMPORT,
H. The transport
of oxygen
in the sheeps placenta:
the
diffusion
constant
of the placenta.
Yale J. Biol. Med. 27 : 26-34,
1954-1955.
LINGO,
1;. ID., E. P. HILL, AND G. G, POWER. Theoretical
analysis
of factors
affecting
placental
02 transfer.
Am. J. Physiol.
222:
730-739,
1972.
LONGO,
L. II., M. DELIVORIA-PAPADOPOULOS,
G. G. POWER,
E. P.
HILL,
AND R, E. FORSTER
II. Diffusion
equilibration
of inert gases
between
maternal
and fetal placental
capillaries.
Am. J. physiol.
219: 561-569,
1970.
MESCHIA,
G., A. HELLEGERS,
J. N. BLECHNER,
A. S. WOLKOFF,
AND D. H. BARRON.
A comparison
of the oxygen
dissociation
curves of the bloods of maternal,
fetal and newborn
sheep at
various pHs. Quart. J, Exptl.
Physiol.
46 : 95- 100, 196 1.
MESCHIA,
G., F. C. BATTAGLIA,
AND P. D. BRUNS. Theoretical
and
experimental
study of transplacental
diffusion.
J. A#d. Physiol.
22 :
1171-l 178, 1967.
MESCHIA,
G., J. R. COTTER,
C. S. BREATHNACH,
AND D. H.
BARRON.
The hemoglobin,
oxygen,
carbon dioxide
and hydrogen
22.
23.
24.
25.
26.
27.
ion concentrations
in the umbilical
bloods of sheep and goats as
sampled
via indwelling
plastic catheters.
Quart. J. Exptl.
Physiol.
50: 185-195,
1965,
METCALFE,
J., W. MOLL,
H. BARTELS,
P. HILPERT,
AND J. T.
PARER.
Transfer
of carbon
monoxide
and nitrous
oxide in the
artificially
perfused
sheep placenta,
Circulation
Res.
16 : 95- 101,
1965.
POWER,
G. G., E:, P. HILL, AND L. D. LONGO. Analysis
of uneven
distribution
of diffusing
capacity
and blood flow in the placenta.
Am. J. Physioi.
222: 740-746,
1972.
POWER,
G, G., AND L. D. LONGO. Graphical
analysis of maternal
and fetal exchange
of 02 and CQ. J. Appl. Physiol.
26 : 38-47,
1969.
POWER,
G. G., L. D. LONGO,
N. N. WAGNER,
JR., D. E. KUHL,
AND R. E. FORSTER
II. Uneven
distribution
of maternal
and fetal
placental
blood flow, as demonstrated
using macroaggregates,
and
its response to hypoxia.
J, Clin. Invest. 46 : 2053-2063,
1967.
RANKIN,
J., G. MESCHIA,
E. L. MAKOWSKI,
AND F. C. BATTAGLLZA.
Macroscopic
distribution
of blood flow in the sheep placenta.
Am.
J. Physiol.
219: 9-16, 1970.
SNEDECOR,
G. W., AND W. G. COCHRAN.
Statistid
Methods
(6th
ed.). Ames: Iowa State Univ. Press, 1967, p. 455.