British Journal of Obstetrics and Gynaecology

November 1984, Vol. 91, pp. 1096-1102

Permeability of the human placenta to bicarbonate :

in-vitro perfusion studies
J. G. AARNOUDSE, N. P. ILLSLEY, P. PENFOLD, S. E. BARDSLEY, P. RISPENS &
F. E. HYTTEN Division of Perinatal Medicine, MRC Clinical Research Centre, Harrow, UK

Summary. The effect of maternal acidosis on fetal acid-base balance was

studied in a dual circuit perfusion of a single cotyledon in normal, term,
human placentas. Both the fetal and maternal (intervillous) circulations were
perfused with a Krebs-Ringer solution adjusted to pH values between 7.35
and 7.45. After a control period, the perfusate in the maternal circulation
was replaced by an acidified medium (mean pH 7.06) for 30 min. This was
followed by a second control period of 30 min during which the acidified
maternal perfusate was replaced with the original medium. During the 30
min of maternal acidosis, fetal vein pH was not significantly altered despite
the large decrease in maternal artery pH, but there was an efflux of total CO,
(tCO,) from the placenta into the maternal circulation which was not
matched by an influx of tCO, from the fetal circulation. The tCO, trans-
ferred was in the form of bicarbonate rather than dissolved CO,, but the
maximal rate of tCO, transfer of in the form of bicarbonate was lower than
the rate of placental transfer of tCO, necessary in vivo. It is probable there-
fore that bicarbonate does not play a major role in placental CO, transfer
but the placental tissue bicarbonate pool may play an important part in
buffering the fetus against changes in maternal pH or blood gas status.

During the second stage of labour maternal blood
pH tends to fall and maternal ketosis may also
cause a fall in blood pH. It is uncertain whether
maternal acid-base status has any influence on
fetal pH but such an effect could become import-
ant if the fetus also has a metabolic acidosis
caused by hypoxia. Blechner et al. (1 967) showed
that maternal acidosis induced by ammonium
chloride given intravenously during anaesthesia
before caesarean section had a negligible effect on
fetal plasma bicarbonate and pH and they con-
cluded that the fetus is well protected against
maternal metabolic acidosis. Longo et al. ( 1 974)
found no appreciable transfer of bicarbonate
across the placenta in sheep experiments. How-
Correspondence: N. P. Illsley, Division of Perinatal
Medicine, Medical Research Council Clinical Research
Centre, Watford Road, Harrow HA1 3UJ.
ever, other workers (Rooth 1964; Newman et al.
1967; Jacobsen 1970; Chang & Wood 1976;
Kastendieck & Kunzel 1979) reported bi-
carbonate transfer across the human placenta and
showed that maternal metabolic acidosis was
accompanied by fetal acidosis. Goodlin & Kaiser
(1957) reported a 4-h time lag between maternal
acidosis, induced with ammonium chloride before
delivery, and evidence of fetal acidosis.
Studying the inter-relationship between
maternal and fetal acid-base balance in vivo
presents several problems. The first is the impos-
sibility of separating the influence of maternal
acid-base disturbances on placental bicarbonate
transfer, from acid-base variations caused by
changes in fetal and maternal metabolism. The
second is the difficulty in assessing net transfer
because of the problems of accurate measure-
ment of umbilical and uterine blood flows. Finally

1096

there is the complicating problem of the added
equilibration steps between maternal and fetal
erythrocytes and plasma, superimposed on the
measurements of placental exchange. Animal
studies on maternofetal acid-base balance are not
necessarily instructive because of the wide species
differences in placentation.
To overcome these problems, we studied the
effect of maternal acidosis on placental bi-
carbonate transfer using an in-vitro placental per-
fusion system with dual circuits as described
previously by Penfold et a1 (1981) and Illsley et
al. (1984).
Methods
The flux of CO, across the placenta was assessed
by measuring the umbilical and uterine artery and
vein pH and Pco, before, during, and after
acidification of the maternal circulation. This
could be achieved while measuring the flow rates
in both fetal and maternal circulations.
Six placentas from normal pregnancies were
obtained immediately after delivery and the fetal
and maternal (intervillous) circulations were
cannulated and cleared of blood as described
previously (Penfold et al. 1981; Illsley et al.
1984). After clearance of blood, the fetal per-
fusate was recirculated using a total circuit
volume of 300-400 ml, while the maternal
medium was perfused on a single-pass basis. We
used a recirculating fetal circuit for comparison
with previous studies (Illsley et al. 1984). The per-
fusion medium in both circuits was a
Krebs-Ringer buffer containing NaCl (6.92 g),
NaHCO, (2.10 g), KCI (0.35 g), KH,PO,
(0.16 g), MgSO, (0.29 g) and CaCI, (0.37 g) per
litre plus 30 g/l of dextran (av.mol.wt approxi-
mately 63000) and 1 g/1 of glucose. Both
perfusates were gassed with 95% oxygen, 5%
carbon dioxide. The perfusates and the placental
lobules were maintained at a constant tempera-
ture of 37OC during perfusion.
Fetal inflow pressure and the flow rates for
both circulations were measured continuously.
Maternal flow rates were kept between 15 and 20
ml/min. The mean fetal rate was 6.2 ml/min (SEM
0.6). If the fetal inflow pressure was >SO mmHg at
a flow rate of G3.5 ml/min, the perfusion was
discarded. Fetal to maternal leakage of perfusate
was detected by measuring the volume of the re-
circulated fetal medium. Experiments in which the
volume altered were discarded. In addition, the
Permeability ofplacenta to HCO, 1097
integrity of the fetal vasculature was assessed by
adding dye to the fetal perfusate at the end of each
experiment. If dye was detected in the maternal
perfusate, the experiment was discarded.
Metabolic viability was assessed by measuring the
placental oxygen consumption as described
previously (Illsley et al. 1984). If during the
perfusion, oxygen consumption fell below 180
pmol/min/kg, the perfusions were not included in
the study. The mean wet weight of the lobules was
30.2 g (SEM 0.8).
After 10 min, artery and vein samples were
taken anaerobically and simultaneously from four
sampling ports located next to the maternal and
fetal cannulation points; the samples were immedi-
ately analysed for pH, PCO,and Po, on a Corning
165 pH-blood gas analyser which also calculated
bicarbonate concentrations. (The blood gas
analyser was routinely calibrated by standard
techniques using calibration gases of known
composition.) Samples were taken at regular
intervals of 10 min thereafter. Thirty minutes after
the first sample was taken, the maternal perfusate
was replaced by an acidified Krebs-Ringer
medium for 30 min. The pH of this medium was
adjusted to between 6.9 and 7.2 with HCI. Finally,
the acidified medium on the maternal side was
replaced by the original one for a further 30 min.
In three experiments, antipyrine clearance from
the maternal circulation was measured as
described previously (Illsley et al. 1984), before,
during and after acidification. Lactate pro-
duction was measured in the perfusate as
described by Engel & Jones (1978). Tests for
statistical significance were made using one- or
two-sample t-tests. The results are expressed as
means (SEM).
Results
Perfusion conditions
The antipyrine clearance was not altered by acidifi-
cation of the maternal perfusate or during the
recovery period after acidification. The mean
clearance value (2.3 ml/min, SEM 0.9) indicated
an appropriate rate of transfer at these flow rates
(Schneider et al. 1972) and demonstrated that the
manipulations of the maternal circulation had not
altered the relation between maternal and fetal
circulations.
Lactate release into maternal and fetal circu-
lations is shown in Table 1, before, during and after
1098 J. G. Aarnoudse et al.
Table 1. Release of lactate into fetal and maternal circulations and tissue oxygen consumption before, during and
after maternal acidification

Phase of perfusion

Before During After

Lactate release (pmol/min/kg)

Fetal 37 (7) 33(10) 36 (9)
Maternal 194 (25) 188 (50) 218(42)
Placental oxygen consumption (Fmol/min/kg) 284 (39) 27 1 (47) 239 (32)

Results are means (SEM) for six perfusions.

acidification respectively. There was no signifi- Perfusion and acidosis

cant change in lactate production between the
three phases of the perfusion. The oxygen con-
sumption of the perfused tissue is also given in
Table 1. No significant differences in oxygen con-
sumption were detected between the three phases.
The mean fetal and maternal pH, pco, and bi-
carbonate values from the inflow and outflow
samples before, during and after simulated
maternal metabolic acidosis are summarized in
Table 2. The differences in inflow pH, PCO,and bi-

Table 2. pH, Pco, and bicarbonate values before, during and after maternal=

Phase of perfusion

Time after start of acidotic perfusion (min)

Before After
10 20 30
~~

PH
Maternal
inflow 7.418 (0.016) 7.072* (0.051) 7,054* (0.034) 7.005* (0.045) 7.339(0.034)
outflow 7.305 (0.025) 7.085 (0.045) 7.020 (0.027) 6.980 (0.024) 7.224(0.051)
Feral
inflow 7.374 (0.014) 7,378 (0.024) 7.373 (0.015) 7,388 (0.026) 7.382 (0.024)
outflow 7.261 (0.028) 7.188 (0.008) 7.178 (0.027) 7.225 (0.023) 7.233 (0,046)

Pco, (kPa)
Maternal
inflow 4.03 (0.35) 6.16**(0.95) 6,03**(0,88) 7.11**(1.03) 4.99 (0.59)
outflow 5.22 (0.50) 7.66 (0.74) 7.26 (0.67) 7.72 (0.61) 6.46 (0.78)
Fetal
inflow 4.52 (0.39) 4.70 (0.37) 4.88 (0.29) 5.16 (0.24) 4.86 (0.41)
outflow 5.72 (0.63) 6.87 (0.16) 6.80 (0.28) 6.68 (0.56) 6.13 (0.92)
HCO, (a)
Maternal
inflow 19.0 (2.0) 12.7* (1.5) 11.9* (1.2) 12.1* (1.2) 19.4 (1.5)
outflow 18.6 (1.7) 15.8 (1.0) 13.4 (0.8) 13.2 (0.9) 18.9 (2.1)
Fetal
inflow 18.9 (1.4) 20.0 (1.0) 20.5 (1.1) 22.6 (0.3) 21.5 (0.9)
outflow 18.4 (1.9) 18.9 (0.8) 18.2 (0.9) 20.1 (0.5) 18.5 (1.7)

a Results are means (SEM) for six perfusions, values for time points during initial and final phases have been combined.
Significance of differences * P<O.O1 compared with maternal inflow values before and after acidosis and with fetal
inflow values during acidosis; ** P< 0.05 compared with maternal inflow values before acidosis.

carbonate values between maternal and fetal circu-
lations are a result of gas diffusion out of the
perfusates in differing lengths of perfusion line at
differing flow rates. During the period of acidifica-
tion, maternal pH and bicarbonate values dropped
significantly (P(0.05) and Pco, values rose
(P(0.05). Fetal pH, Pco, and bicarbonate values
were not significantly altered during the acidifica-
tion period. Before acidification, fetal inflow pH,
Pco, and bicarbonate were similar to maternal
values. During acidosis, a significant gradient in bi-
carbonate developed between maternal (m) and
fetal (f) circulations (f>m; P(O.01). At the same
time the Pco, gradient was aligned in the opposite
direction (m> f).
Flow rates, lobule weight, Pco, and bi-
carbonate concentrations were used to calculate
the net flux of total CO, (tCO,) into or out of the
placenta from maternal and fetal circulations. The
tCO, flux is illustrated in Fig. 1. During the
period of maternal acidification there was a
significant gain in maternal tCO, during passage
through the placenta (P(O.01 at 40, 50 and 60
min). This was paralleled by a smaller but signifi-
cant loss of tCO, from the fetal circulation
(P<0.05 at 50 and 60 min).
The net effect of tCO, changes in the maternal
and fetal circulations is shown in Fig. 2. This
demonstrates an apparent washout of placental
Permeability ofplacenta to HCOr 1099
tissue tCO, during the period of acidosis, followed
by an uptake of tCO, by the tissue during the
recovery phase. The apparent depletion of
placental tissue tCO, during acidosis was not
constant but rather peaked and diminished there-
after, possibly reflecting a decreased gradient of
tCO, between the tissue and the maternal
circulation.
Discussion
The in-vitro perfused placenta originally
described by Schneider et al. (1972) provides an
attactive model for transfer studies. Using this
system it is possible to produce considerable
concentration gradients across the placenta. It
also allows acute replacement of the perfusate.
This provided the opportunity to study a control
period before and after acidification of the
maternal circulation.
The study reported here differs from previous in
vivo investigations on bicarbonate transfer across
the human placenta because only with an in-vitro
perfusion system can one study the effect of
maternal acidosis in isolation from fetal and
maternal interactions. The viability of the perfused
placental preparation has been discussed
previously (Illsley et al. 1984). Rigid criteria for

Acidosis

*r T

-1 L I I I I I I I
20 30 40 50 60 70 80 90
Time (min)
Fig. 1. Changes in total CO, (AtCO,) in fetal (0)and maternal (0)circu-
lations during placental passage before, during and after maternal acidifica-
tion (mean & SEM for six perfusions). Maternal A tCO, at 40,50 and 60 min
significantly greater than at 20,30,80 and 90 min (P<O.Ol). Fetal AtCO, at
50 and 60 min significantlyless than at 10,20,30,70 and 80 min (P<0,05).
 1100 J. G.Aarnoudse et al.
Acidosis
lr
-2 L I I I
20 30 40
Fig. 2. Net changes in placental tissue tCO, (AtCO,). Positive values
equivalent to tissue uptake of tCO,, negative values equivalent to tissue loss
of tCO, (mean i-SEM for six perfusions). AtCO, at 40 min significantly
less than at 30 min (P<O.Ol).
mechanical as well as metabolic variables ensure
the integrity of the preparation. The values of anti-
pyrine clearance, oxygen consumption and lactate
production attest to the stability of the prepara-
tion during the perfusion period.
The absence of any significant buffering in the
Krebs-Ringer perfusate apart from bicarbonate
means that changes in PCO, will have little or no
influence on bicarbonate concentrations. This
effectively makes bicarbonate and dissolved CO,
concentrations independent of each other. Thus in
the acidotic period, dissolved CO, will tend to
travel down its gradient from maternal to fetal
circulations. The bicarbonate gradient during the
same period, however, was in the reverse direction
(f>m). The net tCO, transfer (Fig. 2) was there-
fore the sum of these two processes. Since the net
tCO, transfer into the maternal circulation took
place in the opposite direction to dissolved CO,
transfer, it seems reasonable to suggest that the
tCO, transfer into the maternal circulation was
actually transfer of bicarbonate.
During the period of maternal acidosis, the gain
in tCO, by the maternal circulation was greater
than that which could be accounted for by the CO,
lost from the fetal circulation plus that which might
be derived from substrate oxidation. This is despite
the 10 m~ fetal-to-maternal gradient in bi-
carbonate, larger than that found in vivo. It may be
advantageous therefore to view CO, transfer
I I I I
50 60 70 80 90
Time (min)
across the placenta in terms of three compart-
ments; a placental tissue compartment in addition
to the fetal and maternal ones. In the acidotic
period, transfer from the fetal to placental
compartment does not attain the higher rate of
transfer observed between maternal and placental
compartments. The rate of transfer from the fetal
circulation therefore appears to be the slowest step
in placental bicarbonate transfer.
The placental to maternal transfer of bi-
carbonate did not remain constant during the
period of acidosis. Rather, it diminished, pre-
sumably reflecting a decreasing gradient between
the tissue and maternal circulation and re-
inforcing the suggestion that the bicarbonate was
being washed out of the tissue. The transfer of
tCO, into the placenta in the recovery phase would
also seem to bear out the idea of depletion of
placental tissue bicarbonate during acidosis.
Although maternal pH was restored to a level
approaching that observed in the initial perfusion,
transfer of tCO, out of the circulations and into the
placenta continued to take place.
These data would tend to confirm the suggestion
that placental transfer of CO, takes place pri-
marily by the transfer of dissolved CO, rather than
of the bicarbonate form. This seems likely because
the maximal rate of bicarbonate transfer from the
fetal circulation to the placenta observed in the
present study is below that required to remove

fetally produced CO, from the umbilical circu-
lation in vivo. The required rate of tCO, transfer
for the average term fetus (3.2 kg in weight with a
0.5 kg placenta) is likely to be of the order of
500-1000 pnol/min (for references see Lorijn &
Longo 1980). This is greater than the maximal
transfer rate observed in our experiments (300
pmol/min), despite the 10 m fetal-to-maternal bi-
carbonate gradient. The maximal rate of bi-
carbonate transfer was also well below the
probable placental capacity for transfer of dis-
solved CO, (>2500 ~mol/min/mmHgdifference in
transplacental Pco,; Longo et al. 1974). It is un-
likely therefore that bicarbonate plays a major role
in the transfer of CO, across the placenta.
It is possible that the bicarbonate lost from
placental tissue is in fact replaced by or
exchanged for another anion. The stable rates of
oxygen consumption and lactate production
suggest that placentally produced lactate was not
replacing bicarbonate. On the other hand, an
exchange of chloride for bicarbonate, such as that
observed in the erythrocyte, is a possibility and
requires further investigations.
The decrease in fetal vein pH after maternal
acidification was not significant, despite a pro-
found and protracted decrease in maternal artery
pH. It is probable therefore that severe maternal
acidosis in vivo, unless prolonged beyond 30
minutes, will not adversely affect fetal pH. In the
second stage of labour both maternal and fetal
blood pH values are lower than normal. Our
results suggest that the low pH levels are not
linked but rather that the falls in pH occur as a
result of independent processes in mother and
fetus.
The placental bicarbonate pool, though not
important in overall CO, transfer, may protect
the fetus against changes in maternal pH and
blood gas status by acting as an additional buffer
between the two circulations. Thus bicarbonate
transfer could take place between tissue and the
maternal circulation while minimizing the trans-
mission of maternal pH and blood gas changes to
the fetal circulation. The effect would be to buffer
the fetus against events such as maternal acidosis,
as suggested by Blechner et al. (1967).
In the present study maternal acidosis was
induced by adding hydrochloric acid to the per-
fusate, whereas, during labour, particularly in the
second stage, maternal acidosis is caused by lactic
acid production. It has been suggested that lactic
acid in the maternal circulation, is exchanged
against bicarbonate from the fetal circulation
Permeability of placenta to HCO; 1101
under these circumstances (Kastendieck & Moll
1977), but as lactic acid is also produced in the
placenta, even under aerobic conditions (Illsley et
al. 1984), it is not possible to measure in vivo the
effect of maternal-to-fetal lactic acid transfer per
se. The question of lactate transfer and its influence
on maternofetal acid-base balance is presently
under investigation.
Acknowledgment
The authors are indebted to Professor H. J.
Huisjes, Professor W. G. Zijlstra and Dr. T. E.
Stacey for critical review of the manuscript. Dr
Aarnoudse was a holder of a Royal Society
Research Fellowship.
References
Blechner, J. N., Stenger, V. G., Eitzman, D. V. &
Prystowsky, H. (1967)Effects of maternal acidosis
on the human fetus and newborn infant. Am J Obstet
Gynecol99,46-54.
Chang, A. & Wood, C. (1976)Fetal acid-base balance.
Interdependence of maternal and fetal Pco, and bi-
carbonate concentration. Am J Obstet Gynecol 125,
61-63.
Engle, P. C. & Jones, J. B. (1978) Causes and
elimination of erratic blanks in enzymatic assays
involving the use of NAD+ alkaline hydrazine
buffers: improved conditions for the assay of L-
glutamate, L-lactate and other metabolites. Anal
Biochem 88,475-484.
Goodlin, R. C. & Kaiser, I. H. (1957)The effect of
ammonium chloride induced maternal acidosis on
the human fetus at term: I. pH, hemoglobin, blood
gases. A m JMed Sci 233,662-674.
Illsley, N. P.. Aarnoudse, J. G., Penfold, P., Bardsley, S.
E., Coade, S. B., Stacey,T. E. & Hytten, F. E. (1984)
Mechanical and metabolic viability of an in vitro
placental perfusion system under oxygenated and
anoxic conditions.Placenta 5,213-226.
Jacobsen, L. (1970)Studies on acid-base and electro-
lyte components of human fetal and maternal blood
during labour. PhD. Thesis, University of London,
pp.99-118.
Kastendieck, E. & Moll, W. (1977) The placental
transfer of lactate and bicarbonate in the guinea-pig.
Pfliisers Arch 370,165-171.
Kastendieck, E. & Kunzel, W. (1979)Der Einfluss des
diaplazentaren Bicarbonattransfers auf die meta-
bolische Azidose des Feten. Z GeburtshPerinat 183,
35-44.
Longo, L. D., Delivoria-Papadopoulos,M. & Forster,
R. E. (1974) Placental CO, transfer after fetal
carbonic anhydrase inhibition. A m J Physiol 226,
703-710.
1102 .
I.Aarnoudse et al.
G.
Lorijn, R. H. W. & Longo, L. D. (1980) Clinical and Rooth, G. (1 964) Early detection and prevention of fetal
physiologic implications of increased fetal oxygen acidosis. Lancet i, 290-293.
consumption. A m JObstet Gynecoll36,45 1-457. Schneider, H., Panigel, M. & Dancis, J. (1972) Transfer
Newman, W., Mitchell, P. & Wood, C. (1967) Fetal across the perfused human placenta of antipyrine,
acid-base status. Relationship between maternal and sodium and leucine. A m J Obstet Gynecol 114,
fetal blood bicarbonate concentrations. A m J Obstet 822-828.
Gynecol91,52-57.
Penfold, P., Drury, L., Simmonds, R. & Hytten, F. E.
(1981) Studies of a single placental cotyledon iri
vitro: I. The preparation and its viability. Placenta 2, Received 19 October 1983
149.154. Accepted 22 February 1984