- Email: [email protected]
Blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit
Ferroni et al.
5. Lowensohn RI, Gabbe SG. The value of lecithin/sphingomyelin ratios in diabetes: a critical review. AM J 0BSTET GYNECOL 1979; 134:702. 6. Hallman M, Wermer D, Epstein BL, Gluck L. Effects of maternal insulin or glucose infusion on the fetus: study on lung surfactant phospholipids, plasma myoinositol, and fetal growth in the rabbit. AM J 0BSTET GYNECOL 1982;142:877. 7. Aubry RH, Rourke JE, Almanza R, Cantor RM, Van DorenJE. The lecithin/sphingomyelin ratio in a high-risk obstetric population. Obstet Gynecol 1976;47:21. 8. Chik L, Sokol RJ, Kooi R, Pillay S, Hirsch VJ, Zador I. A perinatal database management system. Methods Inf Med 1981;20:133. 9. Pedersen J, Pedersen LM, Andersen B. Assessors of fetal perinatal mortality in diabetic pregnancy. Diabetes 1974; 23:302. 10. Hertz RH, Sokol RJ, Knoke JD, Rosen MG, Chik L, Hirsch VJ: Clinical estimation of gestational age: rules for avoiding preterm delivery. AM J 0BSTET GYNECOL 1978;131:395. 11. Ballard JL, Novak KL, Driver J. A simplified score for assessment of fetal maturation of newborn infants. J PEDIATR 1979;95:769. 12. Hallman M, Kulovich M, Kirkpatrick E, Sugarman RG, Gluck L. Phosphatidylinositol and phosphatidylglycerol
October I, 1984 Am J Obstet Gynecol
13.
14.
15.
16.
17.
in amniotic fluid: indices of lung maturity. AM J 0BSTET GYNECOL 1976;125:613. Gross TL, Wilson MV, Kuhnert PM, Sokol RJ. Clinical laboratory determination of phosphatidylglycerol: oneand two-dimensional chromatography compared. Clin Chern 1981;27:486. Gross TL, Sokol RJ, Kwong MS, Hirsch V, Kuhnert PM. Fetal pulmonary maturation associated with maternal diabetes mellitus: amniotic fluid phospholipids and clinical outcome. In: Patrick J, ed. Proceedings of the ninth international symposium on fetal breathing and other measurements, London, Ontario, Canada: University of Western Ontario, 1982. Gabbe SG, MestmanJH, Freeman RK, Goebelsmann UT, Lowensohn RI, Nochimson D, Cetrulo C, Quilligan EJ. Measurement and outcome of pregnancy in diabetes mellitus, Classes B to R. AM J 0BSTET GYNECOL 1977; 129:723. Kitzmiller JL, Cloherty JP, Younger MD, Tabatabau A, Rothchild SB, Sosenko I, Epstein MF, Singh G, Nelf RK. Diabetic pregnancy and perinatal morbidity. AM J ORSTET GYNECOL 1978;131:560. Gluck L, Gould JB, Kulovich MV. The acceleration of neurological maturation in high stress pregnancy and its relation to fetal lung maturity. Pediatr Res 1972;6:335.
Blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit Fred D. Fumia, M.D.,* Daniel I. Edelstone, M.D., and Ian R. Holzman, M.D. Pittsburgh, Pennsylvania The purpose of our experiments was to relate blood flow and oxygen delivery (blood flow x arterial blood oxygen concentration) to fetal organs as functions of fetal hematocrit. In 12 chronically catheterized fetal lambs, we observed two patterns of responses of fetal organs and tissues to isovolemic alterations in fetal hematocrit from 12% to 55%. In group 1 organs (brain, heart, adrenal glands), blood flows increased as hematocrit was either raised or lowered from normal such that oxygen delivery to these organs was stable over the entire range of hematocrits studied. In group 2 organs (gastrointestinal tract organs, spleen, kidneys, placenta, and carcass), blood flows varied little over the range of hematocrits from 12% to 40% or 45% but decreased at hematocrits 2!40% to 45%. Because of these flow responses, oxygen delivery to these organs and tissues was maximal at hematocrits ranging from 32% to 38%. Our data indicate that the various organs of the unanesthetized fetal lamb respond in different ways to alterations in hematocrit. It is of particular interest that, in the great majority of the organs of the fetus, oxygen delivery is maximal at hematocrits considered normal for the fetal lamb in utero. (AM J OesTET GVNECOL 1984;150:274-82.)
From the Departments of Obstetrics and Gynecology and Pediatrics, University of Pittsburgh School of Medicine, Magee-Womens Hospital. This work was supported by Grant No. RO 1-163 68 from the National Institute of Child Health and Human Development. Presented at the Twenty-ninth Annual Meeting of the Society for Gynecologic Investigation, Dallas, Texas, March 24-27, 1982. Received for publication December 27, 1983; revised March 13, 1984; accepted April13, 1984. Reprint requests: Daniel I. Edelstone, M.D., Magee-Womens Hospital, Forbes and Halket St., Pittsburgh, PA 15213. *Fellow in Maternal-Fetal Medicine.
274
In adult animals, anemia is accompanied by increases in tissue perfusion, whereas polycythemia is accompanied by decreases in perfusion. 1 - 4 These circulatory changes are primarily due to two factors: (I) the ability of tissues to regulate perfusion in response to alterations in blood oxygen concentration or oxygen supply resulting from anemia or polycythemia and (2) the effects of altered blood viscosity on blood flow. Whereas alterations in blood viscosity uniformly affect all organs and tissues, vasomotor adjustments to changes in oxy-
Volume 150 Number 3
gen concentration or oxygen supply vary considerably among tissues. 2 Because of these vasomotor differences, the effects of hematocrit variations on blood flow and oxygen supply in adult animals show regional variations. Blood flows to the brain and heart vary inversely with hematocrit in such a way that oxygen delivery (blood flow X arterial blood oxygen concentration) is kept relatively constant. 2 • 4 In other organs, such as the intestines, liver, and kidneys, blood flows vary only at the extremes of hematocrit alterations. Thus, in these latter organs, oxygen delivery is maximal at certain hematocrits (that is, optimal hematocrits 1 ) and decreases when hematocrits are raised or lowered from the optimum. 2 • 3 Although the relationships of organ blood flows and oxygen deliveries to hematocrit have been extensively studied in the adult animal, there are very few data on the effects of anemia or polycythemia on blood flows and oxygen deliveries to fetal organs. Tenenbaum et al. 5 found that polycythemia caused reductions in umbilical blood flow and fetal oxygen delivery; effects of polycythemia on individual fetal organs were not assessed. Numerous fetal conditions are known to be associated with alterations in fetal hematocrit. These conditions are as diverse as rhesus isoimmunization, intrauterine growth retardation, twin-to-twin transfusion, or diabetes mellitus. The resultant anemia and polycythemia may produce serious, sometimes lifethreatening, insults to the brain, heart, kidneys, and intestines. 6 The purpose of our experiments was to study the relationships of blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit. We were also interested in determining whether optimal hematocrits existed for any or all of the organs of the intact fetus in utero. Methods
Preparation of animals. We studied 12 fetal lambs of gestational ages ranging from 112 to 129 days (term = 147 days). Seven of the lambs were one of a set of twins and the other five were singletons. After an epidural or a spinal anesthetic (tetracaine hydrochloride, 1%) was administered to the mother, we opened the maternal abdomen and pregnant uterine horn and inserted catheters into two fetal hind limb arteries and one hind limb vein. 7 These catheters were advanced into the fetal descending aorta and abdominal inferior vena cava, respectively. Anesthesia for catheter placements in the fetus was provided by local infiltration with lidocaine hydrochloride, 0.5%. A second uterine incision was made over the area of the fetal neck, and a catheter was placed into one carotid artery and advanced into the brachiocephalic artery. We placed a catheter in the amniotic sac and closed the second uterine incision. When twins were present, the second twin had a catheter inserted into a fetal hind limb artery or
Blood flow and oxygen delivery to fetal organs
275
vem. The second twin was used as a plasma or red blood cell donor. The maternal abdominal incision was closed and all catheters were brought out subcutaneously to the ewe's flank, where they were stored in a canvas pouch. Two to four days after operation, we studied the fetuses of the unanesthetized ewes as the ewes stood quietly in an open transport cage. Experimental protocol. In each fetus, we measured blood flows and oxygen deliveries to all organs at various randomly determined levels of fetal hematocrit. To change the fetal hematocrit, we performed isovolemic exchange transfusions by slowly infusing plasma or packed red blood cells through one fetal catheter while withdrawing fetal blood at the same rate from a second fetal catheter. Hematocrits were changed slowly over 1 to 2 hours during which time fetal aortic blood pressure (obtained with P23Db Statham pressure transducers) and heart rate (with a cardiotachometer) were continously monitored and recorded. Both variables were stable throughout the study. We corrected mean aortic blood pressure to a common reference point, intrauterine pressure, by subtracting the amniotic fluid pressure that was measured simultaneously. Plasma for exchange transfusion was freshly obtained either from the second, or donor, twin or from a nonpregnant adult sheep. Packed red blood cells were obtained either from the donor twin or from newborn lambs 1 to 3 days old. In these neonatal lambs, blood oxygen affinity was the same as that of fetal blood. 8 Whole blood drawn from donor animals was placed in transfer bags containing citrate-phosphate-dextrose as anticoagulant. Centrifugation of the bags yielded plasma and packed red blood cells for exchange transfusion. As the exchange transfusions were progressing, we periodically obtained blood samples for hematocrit determination. After changing hematocrit to a previously selected level, we discontinued the exchange transfusion and allowed the fetus 30 to 60 minutes for stabilization. When the hematocrit and circulatory variables were stable for at least 30 minutes, fetal blood flows and oxygen deliveries were measured as described in the next paragraph. Each fetus had an average of four such measurements (range, two to six) of organ flows and oxygen deliveries made at different fetal hematocrits. At each hematocrit studied, we withdrew 0.6 ml blood samples simultaneously from the brachiocephalic artery and the descending aorta for the analyses of blood gases and pH, hematocrit, and oxygen concentration. Blood samples were kept on ice and were analyzed within 15 minutes. Blood gases and pH were determined at 38° C with standard electrodes and a blood gas meter (Instruments Laboratories); hematocrit was measured in duplicate with the microcapillary technique; and blood oxygen concentration was measured with an oxygen analyzer (Lex-02 -Con, Lexington In-
276
Fumia, Edelstone, and Holzman
October I, 1984 Am J Obstet Gynecol
Table I. Fetal blood gases, pH, heart rates, and mean blood pressures at various fetal hematocrits Feta! hematocrits
n Descending aortic pH Po 2 (mm Hg) Pco 2 (mm Hg) Brachiocephalic arterial P0 2 (mm Hg) Heart rate (bpm) Mean arterial blood pressure (mm Hg)
<20%
20%-29%
30%-39%
?.40%
10
12
II
10
7.36 17 46 20 192 50
± ± ± ± ± ±
0.05 I 4 I 10 3
7.42 18 45 21 181 49
± ± ± ± ± ±
0.06 I 6 I 17 2
7.44 19 43 22 185 48
± ± ± ± ± ±
0.03 2 7 2 10 3
7.34 15* 50 17* 175 55
± ± ± ± ± ±
0.09 3 9 3 10 6
Values are means ± SD. *p < 0.05, analysis of variance, Neuman-Keuls test. 12
struments). 7 Immediately after the blood samples were obtained, we injected approximately 1 to 2 X 106 microspheres (15 p,m diameter, labeled with iodine 125, cerium 141, chromium 51, strontium 85, niobium 95, or scandium 46, 3M Co. or New England Nuclear) into the fetal inferior vena cava over 20 seconds. 9 Blood reference samples were withdrawn from the brachiocephalic artery and descending aorta with withdrawal pumps set at 7.9 mllmin, beginning 10 seconds before the microspheres were injected and ending 60 seconds after the injection catheter was flushed with saline. We used high withdrawal rates to ensure that blood flow measurements made in the anemic animals were accurate.10 The volume of fetal blood withdrawn for the reference samples and for the oxygen analyses was replaced with fetal plasma or red blood cells to maintain a constant intravascular blood volume. In all cases, blood and amniotic fluid pressures and fetal heart rates were stable before, during, and after the microsphere injections. When the oxygen and blood flow measurements were completed at this hematocrit, we changed the fetal hematocrit to a new level and repeated the oxygen and flow measurements. Of the first four fetuses studied, two fetuses had increased hematocrits first and decreased ones second; this order was reversed in the other two fetuses. There were no differences in blood flow responses referable to the order or direction in which hematocrits were altered. Because of these observations and because of the length of time involved in performing the exchange transfusions, we either raised or lowered the hematocrit in the remaining fetuses; none of these fetuses was studied at both high and low hematocrits. Preparation of tissues and calculation and analyses of data. At the end of the study, the sheep and fetus were killed with an overdose of sodium pentobarbital. All catheter locations were verified visually. The fetus was dissected into its component organs and each organ was weighed. Fetal organs and carcass tissues were incinerated at 280° C for 48 to 72 hours, after
which the ashed tissues were packed into counting vials. The amounts of each radionuclide in the tissues and reference arterial blood samples were measured with a well-type gamma scintillation counter (Model 1185R, Searle Analytical) and a 1024-channel pulseheight analyzer (Ultima II, Norland). 7 • 9 Nuclide isotope separation was performed by computer (Nova 3, Data General; average error in the calculation of corrected radioactive counts per minute for each isotope <5%). Left and right kidneys and left and right brain hemispheres were evaluated separately; paired radioactive counts per minute per 100 gm of left and right brain hemispheres and of left and right kidneys each differed by <10%, which indicated that microspheres were well mixed with blood in the arterial circulation.9 Reference arterial blood samples and all organs and tissues, except the adrenal glands, contained more than 1500 microspheres to ensure that the error in any blood flow measurement was <5% of the 95% confidence limits. 11 The adrenal glands contained more than 400 microspheres (potential error 10%; 95% confidence limits) in all but seven instances; these seven measurements were not included in the data analyses because the potential error in these flow calculations because of the inadequate numbers of microspheres was too great. 11 We computed blood flow (Q0 ) and oxygen delivery (Do 2.) to fetal organs and tissues as follows: Qo = (cpmo/cpmrer) <:2rer Do2o = Q 0 • Cao, where cpm 0 is the radioactive counts per minute in the organ or tissue, cpmrer is the radioactive counts per minute in the reference arterial blood sample (brachiocephalic arterial reference sample for upper body organ blood flows, and descending aortic reference sample for lower body organ and placental blood flows 9 ), Qrer is the reference arterial blood flow determined by calibrating the withdrawal pump, and Cao, = arterial blood oxygen concentration (brachiocephalic
Blood flow and oxygen delivery to fetal organs
Volume 150 Number 3
277
BOO y-381-17.911.+ 0.26x2
• I
Ocerebrol
( •
••
ml/min)
( IOOg
•
6Coronary ml/min.\ ( IOOg I
•• 400
•
y = 5.2 ... 0.0611
002
(
cerebrol
ml 02 /min)
IOOg
10
. .. • •
•;a,
•
••
•
'•14.6·0.0711.
•
••
002 Coronary (ml 0 2 /min)
IOOg
• • • •• •• • • •• • . • • ••• • • • I • • •• •••• •.: • •• ••• • • •
•
•
•I •••• •• •• •
Hematocrit (%)
Fig. 2. Top: Coronary blood flow (Qcoronaru) versus fetal heFig. I. Top: Cerebral blood flow (Qcerebraz) versus fetal hematocrit. Bottom: Cerebral oxygen delivery (D0 2 c"''"') versus
hematocrit.
arterial blood sample for upper body tissues and descending aortic blood sample for lower body and placental tissues). Blood flows and oxygen deliveries (dependent variables) were plotted against fetal hematocrit (independent variable). Data analyses were based on 46 measurements of blood flow and oxygen delivery for upper body organs and carcass and on 43 measurements of flow and oxygen delivery for lower body organs, carcass, and placenta (except for the adrenal glands where n = 36 measurements). We analyzed the data statistically by means of a weighted least-squares method based on covariance analysis, which allowed us to generate regression lines that took into account the responses seen in individual animals. 12 With this method of analysis, we could define mathematically the relationships that existed between the dependent variables and the independent variable. In most cases, quadratic equations had the greatest correlation coefficients with the least residuals, which meant that they best described these relationsY When quadratic functions did not provide the best fit, linear correlation analyses were done on the data. We also identified the hematocrits associated with maximal oxygen delivery (that is, optimal hematocrit) to the various organs studied. We calculated the optimal hematocrit by determining the first-order differential (dy/ dx) of the equations relating organ oxygen delivery (y) to hematocrit (x) and by solv-
matocrit. Bottom: Coronary oxygen delivery (D0 2 c"'"""'") versus hematocrit. ing these equations for x when dy/dx = 0. Optimal hematocrits could be calculated only for those organs in which oxygen delivery versus hematocrit was best illustrated by quadratic functions. Results
In the 12 fetal lambs, gestational age was 122 ± 6 (mean ± SD) days on the day of study, and fetal body weight was 2.4 ± 0.4 kg. All fetuses had a descending aortic pH 2::7.33, Po 2 2::19 mm Hg, and Pco 2 :s45 mm Hg at the onset of the study. Fetal hematocrit before the start of the studies was 32% ± 1%. Arterial blood pH was not significantly altered by changes in hematocrit (Table I), although at hematocrits 2::40% or <20%, there was a tendency for pH to fall slightly in most of the fetuses studied. Brachiocephalic arterial and descending aortic blood Po 2 did not change significantly as fetal hematocrits were lowered from 40% to 12% (Table I); when hematocrits were increased above 40%, however, arterial blood Po 2 decreased significantly. The oxygen concentrations in brachiocephalic and descending aortic blood were greatest at a fetal hematocrit of about 33%. At higher and lower hematocrits, arterial blood oxygen concentrations decreased. There was no relationship of fetal heart rate or arterial blood pressure (mean, systolic, or diastolic) to fetal hematocrit (Table I). Fetal heart rate and aortic blood pressure varied < 15% over the entire range of hematocrits studied (from 12% to 55%). Analysis of our data showed that fetal organs and
Fumia, Edelstone, and Holzman
278
.. a
Adrenal
ml/min) 400 ( IOOQ
October I, 1984 Am J Obstet Gynecol
1600
y•t426-68•••z
.:· .. ..
Y • 125 + 2LI'l -0.39x2
QSplenic
(ml/minJ
800
.. . . .. .• . •. ... . . •. .• .~
IOOg
y•261-022x
DO
40
•
2Adrenal
(ml O,Jm;nJ IOOg
20
.·
10
30
20
40
50
o__j
60
Y • -101 + 8.5x -0.13x 2
....
. ••, .•
~
0 0
.-._·-~
•
•
•
002
Splen•c
(ml~/min) IOOg
Hematocrit (%)
Fig. 3. Top: Adrenal blood flow (QAdrenal) versus fetal hematocrit. Bottom: Adrenal oxygen delivery (D0 2 """"') versus hematocrit.
• 0
400
p 124 + 4.S7x -0.082x 2
•
... .__j
30
40
50
60
Hematocrit (%)
Fig. 5. Top: Splenic blood flow (Qsp~en;c) versus fetal hematocrit. Bottom: Splenic oxygen delivery (D0 28.,,,) versus hematocrit. (Optimal hematocrit = 33%.)
..
QRenal
0
. •
•• .. ..
ml/min\ ( IOOg/ 200
y=-25.4 T 2.22x-0.03tx 2
(
... .
DOzRenal ml o,Jmin)
IOOg
.• c•...
•
'
Hematocrit (%)
Fig. 4. Top: Renal blood flow (QRenal) versus fetal hematocrit. Bottom: Renal oxygen delivery (D0 2 H,,,.,) versus hematocrit. (Optimal hematocrit= 36%.)
tissues could be classified into two groups depending on the blood flow and oxygen delivery responses observed as functions of fetal hematocrit. Group 1. This group included the brain, heart, and adrenal glands. Figs. 1 to 3 illustrate the relations of cerebral, coronary, and adrenal blood flows and oxy-
gen deliveries to fetal hematocrit. Blood flows to these organs increased as hematocrit was either raised or lowered from normal. Because of the flow responses seen in the brain, heart, and adrenal glands, oxygen delivery to these organs was stable over the whole range of hematocrits studied. For these organs, we could not identify optimal hematocrits because the equations that best described the relations of oxygen delivery to hematocrit for group 1 organs were linear. Group 2. This group included kidneys, spleen, stomach, small and large intestines, placenta, and carcass tissues. Figs. 4 to 10 illustrate the relations of blood flow and oxygen delivery to these organs and tissues as functions of fetal hematocrit. In general, organ blood flows varied little when hematocrit was changed from 12% to about 40% or 45%. At higher hematocrits, blood flows to group 2 organs and tissues decreased. Oxygen delivery to these organs as a function of hematocrit was best described by quadratic functions that had optimal hematocrits ranging from 32% to 38% (the optimal hematocrit for each organ or tissue is given in the legends to Figs. 4 to 10). Although we calculated a specific optimal hematocrit for each organ or tissue, visual inspection of the data in Figs. 4 to 10 showed that oxygen delivery to these tissues was stable over a range of hematocrits. Fig. 11 illustrates the relation of cardiac output (com-
Blood flow and oxygen delivery to fetal organs
Volume 150 Number 3
279
600 y=202+ 2.77x-0,069x2
QPlacental
(ml~~'")
•
.. •
I
•
051......,
·...
400
. .•... . •
ml/min\ ( IOOq
I
•
~· : ·- _. . ~
••
~
200
....· •
40
oo..........
2 y = -22.9 + 2.22x -0.031x
002
Placental
(ml~min) 20
...,
.. •.
.
20
2.5
•• 40
30
Hematocrit
....... '•.•
.. 60
• • • 10
ml
·'
50
60
Fig. 7. Top: Stomach blood flow (Qsromach) versus fetal hematocrit. Bottom: Stomach oxygen delivery (D02s.'m""') versus hematocrit. (Optimal hematocrit= 38%.)
(%)
Fig. 6. Top: Placental blood flow (QPzacental) versus fetal hematocrit. Bottom: Placental oxygen delivery (D0 2 p'"'"'",) versus
hematocrit. (Optimal hematocrit= 36%.)
Qlntestine
ml/min) 200 ( IOOg
bined ventricular output) and whole body oxygen delivery to fetal hematocrit. These relationships were similar to those seen in the group 2 organs, primarily because the great majority of fetal combined ventricular output was distributed to group 2 organs and tissues. Comment
The data presented in this report describe blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit. We observed two general responses of fetal organs to changes in hematocrit. In the group I organs (brain, heart, and adrenal glands), blood flow varied in such a way that oxygen delivery was maintained over a wide range of hematocrits. With anemia, and its consequent reduced blood oxygen concentration, organ blood flows increased. Since perfusion pressure did not change, the calculated vascular resistance (roughly approximated as mean arterial blood pressure divided by blood flow) in the blood vessels of these organs decreased during isovolemic anemia. Since blood viscosity also is known to decrease with anemia, 1 • 2 a portion of the increased flows must have been due to reduced viscosity. With polycythemia, however, blood flows to these organs also increased. The reason for the increase in perfusion when blood viscosity was also increasing relates to the development of fetal hypoxemia
..
.. y= -6.9 + l.llx -0017x2
DOzlntestine
(ml
.....
....
...
Fig. 8. Top: Intestinal blood flow (Contestine) versus fetal hema-
tocrit. Bottom: Intestinal oxygen delivery (D0 21""''"J versus hematocrit. (Optimal hematocrit= 33%.) at these higher hematocrits. Among fetuses studied at high hematocrits, almost all developed arterial hypoxemia as hematocrit was raised above 40%. Even though blood viscosity must have increased, 5 the circulations of these organs dilated sufficiently to enable oxygen delivery to be constant during the polycythemia. Thus, the blood vessels of the brain, heart, and adrenal glands dilated in response to reductions in oxygen concentration regardless of whether blood viscosity in-
280
Fumia, Edelstone, and Holzman
October 1, 1984 Am J Obstet Gynecol
800
.... ....
y • 477+ 4.76x -0.132x 2
y•zo5-3.4a
Oc~.,
••
• •
200
ml/min) ( IOOg
Cardiac Output
••
..
400
(ml~;in)
•
y=-3.9 + 070l-0011aZ
00
....... .. .....
D02Colon
ml04/min\ ( IOOg )
10
20
40
30
50
60
80 y= -70.9 +6.6411- 0.098x 2
• DQ2Totol (ml o::min)
4
.• ....•.. ..
0
Hematocrit(%)
Fig. 9. Top: Colon blood flow (Qcolon) versus fetal hematocrit. Bottom:Colon oxygen delivery (D0 2 c,,,) versus hematocrit. (Optimal hematocrit= 32%.)
' 00
10
20
.. •• • 40
30
50
60
Hematocrit (%) y= 26.2 + 0.085x -0.006x 2
Fig. 11. Top: Cardiac output (combined ventricular output) versus fetal hematocrit. Bottom: Total fetal oxygen delivery (D0 2 r,) versus hematocrit. (Optimal hematocrit 34%.)
.... . . . - .-.. . -. . . . .~
. •
QCorcass
ml/min) ( IOOg
-
~
-
• 40
30
20
10
50
60
4
y = -4.1
002 (
corcoss
+
0.38x -0.00611.2
•• • •
.... .. •• . ••
ml 0 2/min)
IOOg
•
•
.
I
•
10
20
30
40
50
when fetal hematocrits were increased above about 40% is unknown. It is possible that fetal polycythemia adversely affected the relationship of maternal and fetal blood flows within the placenta (that is, the fetal equivalent of an altered ventilation/perfusion relationship of the adult lung). Similar observations have been made in adult animals 15 in which variations of hematocrit produced changes in ventilation/perfusion relationships in the lung, which led to substantial changes in arterial blood Po 2 • A second possibility is that fetal hypoxemia during polycythemia is a reflection of increases in oxygen extraction by the fetus that can occur as compensation for reductions in fetal-placental blood flow and fetal oxygen delivery. 5 By considering the following relationships, we can see how arterial hypoxemia can develop during fetal polycythemia: Fetal oxygen delivery (Doz)
=
Qplacental •
Cuvo 2
Hematocrit ( "'o)
Fig. 10. Top: Carcass blood flow (Qcarcass) versus fetal hematocrit. Bottom: Carcass oxygen delivery (D0 2 c""",.) versus hematocrit. (Optimal hematocrit = 32%.)
creased or decreased. This response is not surprising since these three organs are known to vasodilate as a result of fetal hypoxemia induced by maternal hypoxemia.I3, 14 The cause of the fetal hypoxemia that occurred
Fetal oxygen extraction Fetal oxygen consumption
(Cuv 0 , - Cao,) Cuv 02
=
(fetal Do2 ) (fetal oxygen extraction) (
where
. Qplacemal
Qptacental
•
CuVo2 )
(Cuv 02 - Cao2 ) Cuv Oz
= fetal-placental blood flow and Cuv 02
Volume 150 Number 3
and Cao, = umbilical venous and arterial blood oxygen concentrations. If fetal oxidative needs are to be met during polycythemia (if fetal oxygen consumption is to be maintained) as fetal Do 2 decreases, then fetal oxygen extraction must increase. If the increase in Cuvo, resulting from the greater fetal hemoglobin concentration is proportionally less than the decrease in Qplacentai because of increased viscosity, then Cao, may also decrease. We have preliminary experiments that support these theoretical considerations. 16 In these experiments, increasing the fetal hematocrit to about 50% with fetal red blood cells did not affect oxygen consumption by the fetal lamb. The net result was that Cao, decreased as both Qpiacentai and fetal Do 2 fell. Tenenbaum et al 5 obtained similar results when fetal hematocrit was increased to 46% with adult red blood cells. In contrast to responses observed in group 1 organs, blood flow responses observed in group 2 organs and tissues (gastrointestinal organs, placenta, kidneys, carcass and spleen) indicated the relative inability of these tissues to regulate perfusions in response to reduced blood oxygen concentrations occurring with anemia as well as with polycythemia. Blood flows were relatively unaffected by increases in hematocrit until hematocrits exceeded 40% to 45%, a point at which viscosity has been shown to increase rapidly. 1· 2 • 5 Thus, in group 2 organs and tissues, changes in blood flows observed over the range of hematocrits studied reflected adjustments resulting primarily from altered blood viscosity.1· 2 • 4 Our observations are supported by the experiments of Tenenbaum et al. 5 who found that increases in fetal hematocrit from 33% to 46%, produced by isovolemic exchange transfusion with adult red blood cells, resulted in 58% increases in blood viscosity and 32% reductions in fetal-placental (umbilical) blood flow. The responses of the other fetal organs, however, were not measured. Because of the limited vasodilatory responses of group 2 organs and tissues to changes in hematocrit, tissue oxygen deliveries were not maintained over as wide a range of hematocrits as were those for group 1 organs. This difference in oxygen delivery responses among fetal organs is especially important when we consider that, for the heart, for example, maintenance of aerobic metabolism is critically dependent on a stable oxygen delivery, because oxygen extraction by the heart varies little with changes in blood oxygen concentration.14 In contrast, for organs and tissues having only limited vasodilatory responses to changes in blood oxygen concentration (kidneys, spleen, gastrointestinal organs, placenta, and carcass 13 ), aerobic metabolism can be maintained by adjustments in oxygen extraction.7 These tissues do not require a stable oxygen delivery to meet their oxygen demands. It is of particular
Blood flow and oxygen delivery to fetal organs
281
interest that the hematocrits associated with maximal oxygen delivery (optimal hematocrits) to group 2 organs were the hematocrits generally considered to be normal for the fetal lamb. 5 • 7· 14 The responses of fetal organs to changes in hematocrit were similar to responses that have been observed for the corresponding organs of the adult. 1· 4 In the adult dog, perfusion to the brain and heart varies in such a way that oxygen supply remains relatively constant at hematocrits ranging from 10% to 65%. 2 • 4 In contrast, blood flows to the intestines, liver, kidneys, and spleen vary little as hematocrit is changed from 10% to 60%, although at hematocrits >60% perfusion decreases. 2 Because of these perfusion relationships, oxygen supplies to the intestines, liver, kidneys, and spleen of the adult are stable over narrower ranges than those observed for the brain and heart. Furthermore, optimal hematocrits for these organs are similar to hematocrits normally found in the adult. 2 ~ 4 Thus, for the adult as well as for the fetus, maintaining a normal hematocrit at or near the optimum for oxygen delivery may be an important way of ensuring that oxygen supplies to tissues are sufficient for the metabolic demands of those tissues. We wish to thank Anthony Battelli, Patrick Moran, and Betty Steranka for their skillful technical assistance, and Lynn Heddinger for her assistance in the preparation of this manuscript. REFERENCES 1. Crowell JW, Smith EE. Determinant of the optimal hematocrit. J Appl Physiol 1967;22:501. 2. Fan F-C, Chen RYZ, Schuessler GB, Chien S. Effects of hematocrit variations on regional hemodynamics and oxygen transport in the dog. Am J Physiol 1980;238:H545. 3. Shepherd AP, Riedel GL. Optimal hematocrit for oxygenation of canine intestine. Circ Res 1982;51:233. 4. Jan K-M, Chien S. Effect of hematocrit variations on coronary hemodynamics and oxygen utilization. Am J Physiol 1977;233:H106. 5. Tenenbaum DG, Piasecki GJ, Oh W, Rosenkrantz TS, Jackson BT. Fetal polycythemia and hyperviscosity: effect on umbilical blood flow and fetal oxygen consumption. AMJ 0BSTET GYNECOL 1983;147:48. 6. Ramamurthy RS, Brans YW. Neonatal polycythemia. I. Criteria for diagnosis and treatment. Pediatrics 1981 ;68: 168. 7. Edelstone DI, Holzman IR. Fetal intestinal oxygen consumption at various levels of oxygenation. Am J Physiol 1982;242:H50. 8. ListerG, WalterTK, Versmold HT, Dallman PR, Rudolph AM. Oxygen delivery in lambs: cardiovascular and hematologic development. AmJ Physioi1979;237:H668. 9. Heymann MA, Payne BD, Hoffman JIE, Rudolph AM. Blood flow measurements with radionuclide-labeled particles. Prog Cardiovasc Dis 1977;20:55. 10. Rosenberg AA,Jones MD Jr, Koehler RC, Traystman RJ, Lister G. Precautions for measuring blood flow during anemia with the microsphere technique. Am J Physiol 1983;244: H308. 1 1. Buckberg GD, Luck JC, Payne DB, Hoffman JIE, Archie JP, Fixler DE. Some sources of error in measuring re-
282
Fumia, Edelstone, and Holzman
gional blood flow with radioactive microspheres. J Appl Physiol 1971;31:598. 12. Zar JH. Biostatistical analysis. Englewood Cliffs, New Jersey: Prentice-Hall, 1974. 13. Peeters LLH, Sheldon RE, Jones MD Jr, Makowski EL, Meschia G. Blood flow to fetal organs as a function of arterial oxygen content. AM J 0BSTET GYNECOL 1979; 135:637. 14. Fisher DJ, Heymann MA, Rudolph AM. Fetal myocardial oxygen and carbohydrate consumption during acutely induced hypoxemia. Am J Physiol 1982;242:H657.
October I, 1984 Am J Obstet Gynecol
15. Michalski AH, Lowenstein E, Austen WG, Buckley MJ, Laver MB. Patterns of oxygenation and cardiovascular adjustment to acute, transient normovolemic anemia. Ann Surg 1968;168:946. 16. Edelstone DI, Fumia F. Fetal 0 2 consumption as a function of fetal hematocrit [Abstract]: In: Scientific abstracts of the thirty-first annual meeting of the Society for Gynecologic Investigation, San Francisco: 1984.
5. Lowensohn RI, Gabbe SG. The value of lecithin/sphingomyelin ratios in diabetes: a critical review. AM J 0BSTET GYNECOL 1979; 134:702. 6. Hallman M, Wermer D, Epstein BL, Gluck L. Effects of maternal insulin or glucose infusion on the fetus: study on lung surfactant phospholipids, plasma myoinositol, and fetal growth in the rabbit. AM J 0BSTET GYNECOL 1982;142:877. 7. Aubry RH, Rourke JE, Almanza R, Cantor RM, Van DorenJE. The lecithin/sphingomyelin ratio in a high-risk obstetric population. Obstet Gynecol 1976;47:21. 8. Chik L, Sokol RJ, Kooi R, Pillay S, Hirsch VJ, Zador I. A perinatal database management system. Methods Inf Med 1981;20:133. 9. Pedersen J, Pedersen LM, Andersen B. Assessors of fetal perinatal mortality in diabetic pregnancy. Diabetes 1974; 23:302. 10. Hertz RH, Sokol RJ, Knoke JD, Rosen MG, Chik L, Hirsch VJ: Clinical estimation of gestational age: rules for avoiding preterm delivery. AM J 0BSTET GYNECOL 1978;131:395. 11. Ballard JL, Novak KL, Driver J. A simplified score for assessment of fetal maturation of newborn infants. J PEDIATR 1979;95:769. 12. Hallman M, Kulovich M, Kirkpatrick E, Sugarman RG, Gluck L. Phosphatidylinositol and phosphatidylglycerol
October I, 1984 Am J Obstet Gynecol
13.
14.
15.
16.
17.
in amniotic fluid: indices of lung maturity. AM J 0BSTET GYNECOL 1976;125:613. Gross TL, Wilson MV, Kuhnert PM, Sokol RJ. Clinical laboratory determination of phosphatidylglycerol: oneand two-dimensional chromatography compared. Clin Chern 1981;27:486. Gross TL, Sokol RJ, Kwong MS, Hirsch V, Kuhnert PM. Fetal pulmonary maturation associated with maternal diabetes mellitus: amniotic fluid phospholipids and clinical outcome. In: Patrick J, ed. Proceedings of the ninth international symposium on fetal breathing and other measurements, London, Ontario, Canada: University of Western Ontario, 1982. Gabbe SG, MestmanJH, Freeman RK, Goebelsmann UT, Lowensohn RI, Nochimson D, Cetrulo C, Quilligan EJ. Measurement and outcome of pregnancy in diabetes mellitus, Classes B to R. AM J 0BSTET GYNECOL 1977; 129:723. Kitzmiller JL, Cloherty JP, Younger MD, Tabatabau A, Rothchild SB, Sosenko I, Epstein MF, Singh G, Nelf RK. Diabetic pregnancy and perinatal morbidity. AM J ORSTET GYNECOL 1978;131:560. Gluck L, Gould JB, Kulovich MV. The acceleration of neurological maturation in high stress pregnancy and its relation to fetal lung maturity. Pediatr Res 1972;6:335.
Blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit Fred D. Fumia, M.D.,* Daniel I. Edelstone, M.D., and Ian R. Holzman, M.D. Pittsburgh, Pennsylvania The purpose of our experiments was to relate blood flow and oxygen delivery (blood flow x arterial blood oxygen concentration) to fetal organs as functions of fetal hematocrit. In 12 chronically catheterized fetal lambs, we observed two patterns of responses of fetal organs and tissues to isovolemic alterations in fetal hematocrit from 12% to 55%. In group 1 organs (brain, heart, adrenal glands), blood flows increased as hematocrit was either raised or lowered from normal such that oxygen delivery to these organs was stable over the entire range of hematocrits studied. In group 2 organs (gastrointestinal tract organs, spleen, kidneys, placenta, and carcass), blood flows varied little over the range of hematocrits from 12% to 40% or 45% but decreased at hematocrits 2!40% to 45%. Because of these flow responses, oxygen delivery to these organs and tissues was maximal at hematocrits ranging from 32% to 38%. Our data indicate that the various organs of the unanesthetized fetal lamb respond in different ways to alterations in hematocrit. It is of particular interest that, in the great majority of the organs of the fetus, oxygen delivery is maximal at hematocrits considered normal for the fetal lamb in utero. (AM J OesTET GVNECOL 1984;150:274-82.)
From the Departments of Obstetrics and Gynecology and Pediatrics, University of Pittsburgh School of Medicine, Magee-Womens Hospital. This work was supported by Grant No. RO 1-163 68 from the National Institute of Child Health and Human Development. Presented at the Twenty-ninth Annual Meeting of the Society for Gynecologic Investigation, Dallas, Texas, March 24-27, 1982. Received for publication December 27, 1983; revised March 13, 1984; accepted April13, 1984. Reprint requests: Daniel I. Edelstone, M.D., Magee-Womens Hospital, Forbes and Halket St., Pittsburgh, PA 15213. *Fellow in Maternal-Fetal Medicine.
274
In adult animals, anemia is accompanied by increases in tissue perfusion, whereas polycythemia is accompanied by decreases in perfusion. 1 - 4 These circulatory changes are primarily due to two factors: (I) the ability of tissues to regulate perfusion in response to alterations in blood oxygen concentration or oxygen supply resulting from anemia or polycythemia and (2) the effects of altered blood viscosity on blood flow. Whereas alterations in blood viscosity uniformly affect all organs and tissues, vasomotor adjustments to changes in oxy-
Volume 150 Number 3
gen concentration or oxygen supply vary considerably among tissues. 2 Because of these vasomotor differences, the effects of hematocrit variations on blood flow and oxygen supply in adult animals show regional variations. Blood flows to the brain and heart vary inversely with hematocrit in such a way that oxygen delivery (blood flow X arterial blood oxygen concentration) is kept relatively constant. 2 • 4 In other organs, such as the intestines, liver, and kidneys, blood flows vary only at the extremes of hematocrit alterations. Thus, in these latter organs, oxygen delivery is maximal at certain hematocrits (that is, optimal hematocrits 1 ) and decreases when hematocrits are raised or lowered from the optimum. 2 • 3 Although the relationships of organ blood flows and oxygen deliveries to hematocrit have been extensively studied in the adult animal, there are very few data on the effects of anemia or polycythemia on blood flows and oxygen deliveries to fetal organs. Tenenbaum et al. 5 found that polycythemia caused reductions in umbilical blood flow and fetal oxygen delivery; effects of polycythemia on individual fetal organs were not assessed. Numerous fetal conditions are known to be associated with alterations in fetal hematocrit. These conditions are as diverse as rhesus isoimmunization, intrauterine growth retardation, twin-to-twin transfusion, or diabetes mellitus. The resultant anemia and polycythemia may produce serious, sometimes lifethreatening, insults to the brain, heart, kidneys, and intestines. 6 The purpose of our experiments was to study the relationships of blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit. We were also interested in determining whether optimal hematocrits existed for any or all of the organs of the intact fetus in utero. Methods
Preparation of animals. We studied 12 fetal lambs of gestational ages ranging from 112 to 129 days (term = 147 days). Seven of the lambs were one of a set of twins and the other five were singletons. After an epidural or a spinal anesthetic (tetracaine hydrochloride, 1%) was administered to the mother, we opened the maternal abdomen and pregnant uterine horn and inserted catheters into two fetal hind limb arteries and one hind limb vein. 7 These catheters were advanced into the fetal descending aorta and abdominal inferior vena cava, respectively. Anesthesia for catheter placements in the fetus was provided by local infiltration with lidocaine hydrochloride, 0.5%. A second uterine incision was made over the area of the fetal neck, and a catheter was placed into one carotid artery and advanced into the brachiocephalic artery. We placed a catheter in the amniotic sac and closed the second uterine incision. When twins were present, the second twin had a catheter inserted into a fetal hind limb artery or
Blood flow and oxygen delivery to fetal organs
275
vem. The second twin was used as a plasma or red blood cell donor. The maternal abdominal incision was closed and all catheters were brought out subcutaneously to the ewe's flank, where they were stored in a canvas pouch. Two to four days after operation, we studied the fetuses of the unanesthetized ewes as the ewes stood quietly in an open transport cage. Experimental protocol. In each fetus, we measured blood flows and oxygen deliveries to all organs at various randomly determined levels of fetal hematocrit. To change the fetal hematocrit, we performed isovolemic exchange transfusions by slowly infusing plasma or packed red blood cells through one fetal catheter while withdrawing fetal blood at the same rate from a second fetal catheter. Hematocrits were changed slowly over 1 to 2 hours during which time fetal aortic blood pressure (obtained with P23Db Statham pressure transducers) and heart rate (with a cardiotachometer) were continously monitored and recorded. Both variables were stable throughout the study. We corrected mean aortic blood pressure to a common reference point, intrauterine pressure, by subtracting the amniotic fluid pressure that was measured simultaneously. Plasma for exchange transfusion was freshly obtained either from the second, or donor, twin or from a nonpregnant adult sheep. Packed red blood cells were obtained either from the donor twin or from newborn lambs 1 to 3 days old. In these neonatal lambs, blood oxygen affinity was the same as that of fetal blood. 8 Whole blood drawn from donor animals was placed in transfer bags containing citrate-phosphate-dextrose as anticoagulant. Centrifugation of the bags yielded plasma and packed red blood cells for exchange transfusion. As the exchange transfusions were progressing, we periodically obtained blood samples for hematocrit determination. After changing hematocrit to a previously selected level, we discontinued the exchange transfusion and allowed the fetus 30 to 60 minutes for stabilization. When the hematocrit and circulatory variables were stable for at least 30 minutes, fetal blood flows and oxygen deliveries were measured as described in the next paragraph. Each fetus had an average of four such measurements (range, two to six) of organ flows and oxygen deliveries made at different fetal hematocrits. At each hematocrit studied, we withdrew 0.6 ml blood samples simultaneously from the brachiocephalic artery and the descending aorta for the analyses of blood gases and pH, hematocrit, and oxygen concentration. Blood samples were kept on ice and were analyzed within 15 minutes. Blood gases and pH were determined at 38° C with standard electrodes and a blood gas meter (Instruments Laboratories); hematocrit was measured in duplicate with the microcapillary technique; and blood oxygen concentration was measured with an oxygen analyzer (Lex-02 -Con, Lexington In-
276
Fumia, Edelstone, and Holzman
October I, 1984 Am J Obstet Gynecol
Table I. Fetal blood gases, pH, heart rates, and mean blood pressures at various fetal hematocrits Feta! hematocrits
n Descending aortic pH Po 2 (mm Hg) Pco 2 (mm Hg) Brachiocephalic arterial P0 2 (mm Hg) Heart rate (bpm) Mean arterial blood pressure (mm Hg)
<20%
20%-29%
30%-39%
?.40%
10
12
II
10
7.36 17 46 20 192 50
± ± ± ± ± ±
0.05 I 4 I 10 3
7.42 18 45 21 181 49
± ± ± ± ± ±
0.06 I 6 I 17 2
7.44 19 43 22 185 48
± ± ± ± ± ±
0.03 2 7 2 10 3
7.34 15* 50 17* 175 55
± ± ± ± ± ±
0.09 3 9 3 10 6
Values are means ± SD. *p < 0.05, analysis of variance, Neuman-Keuls test. 12
struments). 7 Immediately after the blood samples were obtained, we injected approximately 1 to 2 X 106 microspheres (15 p,m diameter, labeled with iodine 125, cerium 141, chromium 51, strontium 85, niobium 95, or scandium 46, 3M Co. or New England Nuclear) into the fetal inferior vena cava over 20 seconds. 9 Blood reference samples were withdrawn from the brachiocephalic artery and descending aorta with withdrawal pumps set at 7.9 mllmin, beginning 10 seconds before the microspheres were injected and ending 60 seconds after the injection catheter was flushed with saline. We used high withdrawal rates to ensure that blood flow measurements made in the anemic animals were accurate.10 The volume of fetal blood withdrawn for the reference samples and for the oxygen analyses was replaced with fetal plasma or red blood cells to maintain a constant intravascular blood volume. In all cases, blood and amniotic fluid pressures and fetal heart rates were stable before, during, and after the microsphere injections. When the oxygen and blood flow measurements were completed at this hematocrit, we changed the fetal hematocrit to a new level and repeated the oxygen and flow measurements. Of the first four fetuses studied, two fetuses had increased hematocrits first and decreased ones second; this order was reversed in the other two fetuses. There were no differences in blood flow responses referable to the order or direction in which hematocrits were altered. Because of these observations and because of the length of time involved in performing the exchange transfusions, we either raised or lowered the hematocrit in the remaining fetuses; none of these fetuses was studied at both high and low hematocrits. Preparation of tissues and calculation and analyses of data. At the end of the study, the sheep and fetus were killed with an overdose of sodium pentobarbital. All catheter locations were verified visually. The fetus was dissected into its component organs and each organ was weighed. Fetal organs and carcass tissues were incinerated at 280° C for 48 to 72 hours, after
which the ashed tissues were packed into counting vials. The amounts of each radionuclide in the tissues and reference arterial blood samples were measured with a well-type gamma scintillation counter (Model 1185R, Searle Analytical) and a 1024-channel pulseheight analyzer (Ultima II, Norland). 7 • 9 Nuclide isotope separation was performed by computer (Nova 3, Data General; average error in the calculation of corrected radioactive counts per minute for each isotope <5%). Left and right kidneys and left and right brain hemispheres were evaluated separately; paired radioactive counts per minute per 100 gm of left and right brain hemispheres and of left and right kidneys each differed by <10%, which indicated that microspheres were well mixed with blood in the arterial circulation.9 Reference arterial blood samples and all organs and tissues, except the adrenal glands, contained more than 1500 microspheres to ensure that the error in any blood flow measurement was <5% of the 95% confidence limits. 11 The adrenal glands contained more than 400 microspheres (potential error 10%; 95% confidence limits) in all but seven instances; these seven measurements were not included in the data analyses because the potential error in these flow calculations because of the inadequate numbers of microspheres was too great. 11 We computed blood flow (Q0 ) and oxygen delivery (Do 2.) to fetal organs and tissues as follows: Qo = (cpmo/cpmrer) <:2rer Do2o = Q 0 • Cao, where cpm 0 is the radioactive counts per minute in the organ or tissue, cpmrer is the radioactive counts per minute in the reference arterial blood sample (brachiocephalic arterial reference sample for upper body organ blood flows, and descending aortic reference sample for lower body organ and placental blood flows 9 ), Qrer is the reference arterial blood flow determined by calibrating the withdrawal pump, and Cao, = arterial blood oxygen concentration (brachiocephalic
Blood flow and oxygen delivery to fetal organs
Volume 150 Number 3
277
BOO y-381-17.911.+ 0.26x2
• I
Ocerebrol
( •
••
ml/min)
( IOOg
•
6Coronary ml/min.\ ( IOOg I
•• 400
•
y = 5.2 ... 0.0611
002
(
cerebrol
ml 02 /min)
IOOg
10
. .. • •
•;a,
•
••
•
'•14.6·0.0711.
•
••
002 Coronary (ml 0 2 /min)
IOOg
• • • •• •• • • •• • . • • ••• • • • I • • •• •••• •.: • •• ••• • • •
•
•
•I •••• •• •• •
Hematocrit (%)
Fig. 2. Top: Coronary blood flow (Qcoronaru) versus fetal heFig. I. Top: Cerebral blood flow (Qcerebraz) versus fetal hematocrit. Bottom: Cerebral oxygen delivery (D0 2 c"''"') versus
hematocrit.
arterial blood sample for upper body tissues and descending aortic blood sample for lower body and placental tissues). Blood flows and oxygen deliveries (dependent variables) were plotted against fetal hematocrit (independent variable). Data analyses were based on 46 measurements of blood flow and oxygen delivery for upper body organs and carcass and on 43 measurements of flow and oxygen delivery for lower body organs, carcass, and placenta (except for the adrenal glands where n = 36 measurements). We analyzed the data statistically by means of a weighted least-squares method based on covariance analysis, which allowed us to generate regression lines that took into account the responses seen in individual animals. 12 With this method of analysis, we could define mathematically the relationships that existed between the dependent variables and the independent variable. In most cases, quadratic equations had the greatest correlation coefficients with the least residuals, which meant that they best described these relationsY When quadratic functions did not provide the best fit, linear correlation analyses were done on the data. We also identified the hematocrits associated with maximal oxygen delivery (that is, optimal hematocrit) to the various organs studied. We calculated the optimal hematocrit by determining the first-order differential (dy/ dx) of the equations relating organ oxygen delivery (y) to hematocrit (x) and by solv-
matocrit. Bottom: Coronary oxygen delivery (D0 2 c"'"""'") versus hematocrit. ing these equations for x when dy/dx = 0. Optimal hematocrits could be calculated only for those organs in which oxygen delivery versus hematocrit was best illustrated by quadratic functions. Results
In the 12 fetal lambs, gestational age was 122 ± 6 (mean ± SD) days on the day of study, and fetal body weight was 2.4 ± 0.4 kg. All fetuses had a descending aortic pH 2::7.33, Po 2 2::19 mm Hg, and Pco 2 :s45 mm Hg at the onset of the study. Fetal hematocrit before the start of the studies was 32% ± 1%. Arterial blood pH was not significantly altered by changes in hematocrit (Table I), although at hematocrits 2::40% or <20%, there was a tendency for pH to fall slightly in most of the fetuses studied. Brachiocephalic arterial and descending aortic blood Po 2 did not change significantly as fetal hematocrits were lowered from 40% to 12% (Table I); when hematocrits were increased above 40%, however, arterial blood Po 2 decreased significantly. The oxygen concentrations in brachiocephalic and descending aortic blood were greatest at a fetal hematocrit of about 33%. At higher and lower hematocrits, arterial blood oxygen concentrations decreased. There was no relationship of fetal heart rate or arterial blood pressure (mean, systolic, or diastolic) to fetal hematocrit (Table I). Fetal heart rate and aortic blood pressure varied < 15% over the entire range of hematocrits studied (from 12% to 55%). Analysis of our data showed that fetal organs and
Fumia, Edelstone, and Holzman
278
.. a
Adrenal
ml/min) 400 ( IOOQ
October I, 1984 Am J Obstet Gynecol
1600
y•t426-68•••z
.:· .. ..
Y • 125 + 2LI'l -0.39x2
QSplenic
(ml/minJ
800
.. . . .. .• . •. ... . . •. .• .~
IOOg
y•261-022x
DO
40
•
2Adrenal
(ml O,Jm;nJ IOOg
20
.·
10
30
20
40
50
o__j
60
Y • -101 + 8.5x -0.13x 2
....
. ••, .•
~
0 0
.-._·-~
•
•
•
002
Splen•c
(ml~/min) IOOg
Hematocrit (%)
Fig. 3. Top: Adrenal blood flow (QAdrenal) versus fetal hematocrit. Bottom: Adrenal oxygen delivery (D0 2 """"') versus hematocrit.
• 0
400
p 124 + 4.S7x -0.082x 2
•
... .__j
30
40
50
60
Hematocrit (%)
Fig. 5. Top: Splenic blood flow (Qsp~en;c) versus fetal hematocrit. Bottom: Splenic oxygen delivery (D0 28.,,,) versus hematocrit. (Optimal hematocrit = 33%.)
..
QRenal
0
. •
•• .. ..
ml/min\ ( IOOg/ 200
y=-25.4 T 2.22x-0.03tx 2
(
... .
DOzRenal ml o,Jmin)
IOOg
.• c•...
•
'
Hematocrit (%)
Fig. 4. Top: Renal blood flow (QRenal) versus fetal hematocrit. Bottom: Renal oxygen delivery (D0 2 H,,,.,) versus hematocrit. (Optimal hematocrit= 36%.)
tissues could be classified into two groups depending on the blood flow and oxygen delivery responses observed as functions of fetal hematocrit. Group 1. This group included the brain, heart, and adrenal glands. Figs. 1 to 3 illustrate the relations of cerebral, coronary, and adrenal blood flows and oxy-
gen deliveries to fetal hematocrit. Blood flows to these organs increased as hematocrit was either raised or lowered from normal. Because of the flow responses seen in the brain, heart, and adrenal glands, oxygen delivery to these organs was stable over the whole range of hematocrits studied. For these organs, we could not identify optimal hematocrits because the equations that best described the relations of oxygen delivery to hematocrit for group 1 organs were linear. Group 2. This group included kidneys, spleen, stomach, small and large intestines, placenta, and carcass tissues. Figs. 4 to 10 illustrate the relations of blood flow and oxygen delivery to these organs and tissues as functions of fetal hematocrit. In general, organ blood flows varied little when hematocrit was changed from 12% to about 40% or 45%. At higher hematocrits, blood flows to group 2 organs and tissues decreased. Oxygen delivery to these organs as a function of hematocrit was best described by quadratic functions that had optimal hematocrits ranging from 32% to 38% (the optimal hematocrit for each organ or tissue is given in the legends to Figs. 4 to 10). Although we calculated a specific optimal hematocrit for each organ or tissue, visual inspection of the data in Figs. 4 to 10 showed that oxygen delivery to these tissues was stable over a range of hematocrits. Fig. 11 illustrates the relation of cardiac output (com-
Blood flow and oxygen delivery to fetal organs
Volume 150 Number 3
279
600 y=202+ 2.77x-0,069x2
QPlacental
(ml~~'")
•
.. •
I
•
051......,
·...
400
. .•... . •
ml/min\ ( IOOq
I
•
~· : ·- _. . ~
••
~
200
....· •
40
oo..........
2 y = -22.9 + 2.22x -0.031x
002
Placental
(ml~min) 20
...,
.. •.
.
20
2.5
•• 40
30
Hematocrit
....... '•.•
.. 60
• • • 10
ml
·'
50
60
Fig. 7. Top: Stomach blood flow (Qsromach) versus fetal hematocrit. Bottom: Stomach oxygen delivery (D02s.'m""') versus hematocrit. (Optimal hematocrit= 38%.)
(%)
Fig. 6. Top: Placental blood flow (QPzacental) versus fetal hematocrit. Bottom: Placental oxygen delivery (D0 2 p'"'"'",) versus
hematocrit. (Optimal hematocrit= 36%.)
Qlntestine
ml/min) 200 ( IOOg
bined ventricular output) and whole body oxygen delivery to fetal hematocrit. These relationships were similar to those seen in the group 2 organs, primarily because the great majority of fetal combined ventricular output was distributed to group 2 organs and tissues. Comment
The data presented in this report describe blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit. We observed two general responses of fetal organs to changes in hematocrit. In the group I organs (brain, heart, and adrenal glands), blood flow varied in such a way that oxygen delivery was maintained over a wide range of hematocrits. With anemia, and its consequent reduced blood oxygen concentration, organ blood flows increased. Since perfusion pressure did not change, the calculated vascular resistance (roughly approximated as mean arterial blood pressure divided by blood flow) in the blood vessels of these organs decreased during isovolemic anemia. Since blood viscosity also is known to decrease with anemia, 1 • 2 a portion of the increased flows must have been due to reduced viscosity. With polycythemia, however, blood flows to these organs also increased. The reason for the increase in perfusion when blood viscosity was also increasing relates to the development of fetal hypoxemia
..
.. y= -6.9 + l.llx -0017x2
DOzlntestine
(ml
.....
....
...
Fig. 8. Top: Intestinal blood flow (Contestine) versus fetal hema-
tocrit. Bottom: Intestinal oxygen delivery (D0 21""''"J versus hematocrit. (Optimal hematocrit= 33%.) at these higher hematocrits. Among fetuses studied at high hematocrits, almost all developed arterial hypoxemia as hematocrit was raised above 40%. Even though blood viscosity must have increased, 5 the circulations of these organs dilated sufficiently to enable oxygen delivery to be constant during the polycythemia. Thus, the blood vessels of the brain, heart, and adrenal glands dilated in response to reductions in oxygen concentration regardless of whether blood viscosity in-
280
Fumia, Edelstone, and Holzman
October 1, 1984 Am J Obstet Gynecol
800
.... ....
y • 477+ 4.76x -0.132x 2
y•zo5-3.4a
Oc~.,
••
• •
200
ml/min) ( IOOg
Cardiac Output
••
..
400
(ml~;in)
•
y=-3.9 + 070l-0011aZ
00
....... .. .....
D02Colon
ml04/min\ ( IOOg )
10
20
40
30
50
60
80 y= -70.9 +6.6411- 0.098x 2
• DQ2Totol (ml o::min)
4
.• ....•.. ..
0
Hematocrit(%)
Fig. 9. Top: Colon blood flow (Qcolon) versus fetal hematocrit. Bottom:Colon oxygen delivery (D0 2 c,,,) versus hematocrit. (Optimal hematocrit= 32%.)
' 00
10
20
.. •• • 40
30
50
60
Hematocrit (%) y= 26.2 + 0.085x -0.006x 2
Fig. 11. Top: Cardiac output (combined ventricular output) versus fetal hematocrit. Bottom: Total fetal oxygen delivery (D0 2 r,) versus hematocrit. (Optimal hematocrit 34%.)
.... . . . - .-.. . -. . . . .~
. •
QCorcass
ml/min) ( IOOg
-
~
-
• 40
30
20
10
50
60
4
y = -4.1
002 (
corcoss
+
0.38x -0.00611.2
•• • •
.... .. •• . ••
ml 0 2/min)
IOOg
•
•
.
I
•
10
20
30
40
50
when fetal hematocrits were increased above about 40% is unknown. It is possible that fetal polycythemia adversely affected the relationship of maternal and fetal blood flows within the placenta (that is, the fetal equivalent of an altered ventilation/perfusion relationship of the adult lung). Similar observations have been made in adult animals 15 in which variations of hematocrit produced changes in ventilation/perfusion relationships in the lung, which led to substantial changes in arterial blood Po 2 • A second possibility is that fetal hypoxemia during polycythemia is a reflection of increases in oxygen extraction by the fetus that can occur as compensation for reductions in fetal-placental blood flow and fetal oxygen delivery. 5 By considering the following relationships, we can see how arterial hypoxemia can develop during fetal polycythemia: Fetal oxygen delivery (Doz)
=
Qplacental •
Cuvo 2
Hematocrit ( "'o)
Fig. 10. Top: Carcass blood flow (Qcarcass) versus fetal hematocrit. Bottom: Carcass oxygen delivery (D0 2 c""",.) versus hematocrit. (Optimal hematocrit = 32%.)
creased or decreased. This response is not surprising since these three organs are known to vasodilate as a result of fetal hypoxemia induced by maternal hypoxemia.I3, 14 The cause of the fetal hypoxemia that occurred
Fetal oxygen extraction Fetal oxygen consumption
(Cuv 0 , - Cao,) Cuv 02
=
(fetal Do2 ) (fetal oxygen extraction) (
where
. Qplacemal
Qptacental
•
CuVo2 )
(Cuv 02 - Cao2 ) Cuv Oz
= fetal-placental blood flow and Cuv 02
Volume 150 Number 3
and Cao, = umbilical venous and arterial blood oxygen concentrations. If fetal oxidative needs are to be met during polycythemia (if fetal oxygen consumption is to be maintained) as fetal Do 2 decreases, then fetal oxygen extraction must increase. If the increase in Cuvo, resulting from the greater fetal hemoglobin concentration is proportionally less than the decrease in Qplacentai because of increased viscosity, then Cao, may also decrease. We have preliminary experiments that support these theoretical considerations. 16 In these experiments, increasing the fetal hematocrit to about 50% with fetal red blood cells did not affect oxygen consumption by the fetal lamb. The net result was that Cao, decreased as both Qpiacentai and fetal Do 2 fell. Tenenbaum et al 5 obtained similar results when fetal hematocrit was increased to 46% with adult red blood cells. In contrast to responses observed in group 1 organs, blood flow responses observed in group 2 organs and tissues (gastrointestinal organs, placenta, kidneys, carcass and spleen) indicated the relative inability of these tissues to regulate perfusions in response to reduced blood oxygen concentrations occurring with anemia as well as with polycythemia. Blood flows were relatively unaffected by increases in hematocrit until hematocrits exceeded 40% to 45%, a point at which viscosity has been shown to increase rapidly. 1· 2 • 5 Thus, in group 2 organs and tissues, changes in blood flows observed over the range of hematocrits studied reflected adjustments resulting primarily from altered blood viscosity.1· 2 • 4 Our observations are supported by the experiments of Tenenbaum et al. 5 who found that increases in fetal hematocrit from 33% to 46%, produced by isovolemic exchange transfusion with adult red blood cells, resulted in 58% increases in blood viscosity and 32% reductions in fetal-placental (umbilical) blood flow. The responses of the other fetal organs, however, were not measured. Because of the limited vasodilatory responses of group 2 organs and tissues to changes in hematocrit, tissue oxygen deliveries were not maintained over as wide a range of hematocrits as were those for group 1 organs. This difference in oxygen delivery responses among fetal organs is especially important when we consider that, for the heart, for example, maintenance of aerobic metabolism is critically dependent on a stable oxygen delivery, because oxygen extraction by the heart varies little with changes in blood oxygen concentration.14 In contrast, for organs and tissues having only limited vasodilatory responses to changes in blood oxygen concentration (kidneys, spleen, gastrointestinal organs, placenta, and carcass 13 ), aerobic metabolism can be maintained by adjustments in oxygen extraction.7 These tissues do not require a stable oxygen delivery to meet their oxygen demands. It is of particular
Blood flow and oxygen delivery to fetal organs
281
interest that the hematocrits associated with maximal oxygen delivery (optimal hematocrits) to group 2 organs were the hematocrits generally considered to be normal for the fetal lamb. 5 • 7· 14 The responses of fetal organs to changes in hematocrit were similar to responses that have been observed for the corresponding organs of the adult. 1· 4 In the adult dog, perfusion to the brain and heart varies in such a way that oxygen supply remains relatively constant at hematocrits ranging from 10% to 65%. 2 • 4 In contrast, blood flows to the intestines, liver, kidneys, and spleen vary little as hematocrit is changed from 10% to 60%, although at hematocrits >60% perfusion decreases. 2 Because of these perfusion relationships, oxygen supplies to the intestines, liver, kidneys, and spleen of the adult are stable over narrower ranges than those observed for the brain and heart. Furthermore, optimal hematocrits for these organs are similar to hematocrits normally found in the adult. 2 ~ 4 Thus, for the adult as well as for the fetus, maintaining a normal hematocrit at or near the optimum for oxygen delivery may be an important way of ensuring that oxygen supplies to tissues are sufficient for the metabolic demands of those tissues. We wish to thank Anthony Battelli, Patrick Moran, and Betty Steranka for their skillful technical assistance, and Lynn Heddinger for her assistance in the preparation of this manuscript. REFERENCES 1. Crowell JW, Smith EE. Determinant of the optimal hematocrit. J Appl Physiol 1967;22:501. 2. Fan F-C, Chen RYZ, Schuessler GB, Chien S. Effects of hematocrit variations on regional hemodynamics and oxygen transport in the dog. Am J Physiol 1980;238:H545. 3. Shepherd AP, Riedel GL. Optimal hematocrit for oxygenation of canine intestine. Circ Res 1982;51:233. 4. Jan K-M, Chien S. Effect of hematocrit variations on coronary hemodynamics and oxygen utilization. Am J Physiol 1977;233:H106. 5. Tenenbaum DG, Piasecki GJ, Oh W, Rosenkrantz TS, Jackson BT. Fetal polycythemia and hyperviscosity: effect on umbilical blood flow and fetal oxygen consumption. AMJ 0BSTET GYNECOL 1983;147:48. 6. Ramamurthy RS, Brans YW. Neonatal polycythemia. I. Criteria for diagnosis and treatment. Pediatrics 1981 ;68: 168. 7. Edelstone DI, Holzman IR. Fetal intestinal oxygen consumption at various levels of oxygenation. Am J Physiol 1982;242:H50. 8. ListerG, WalterTK, Versmold HT, Dallman PR, Rudolph AM. Oxygen delivery in lambs: cardiovascular and hematologic development. AmJ Physioi1979;237:H668. 9. Heymann MA, Payne BD, Hoffman JIE, Rudolph AM. Blood flow measurements with radionuclide-labeled particles. Prog Cardiovasc Dis 1977;20:55. 10. Rosenberg AA,Jones MD Jr, Koehler RC, Traystman RJ, Lister G. Precautions for measuring blood flow during anemia with the microsphere technique. Am J Physiol 1983;244: H308. 1 1. Buckberg GD, Luck JC, Payne DB, Hoffman JIE, Archie JP, Fixler DE. Some sources of error in measuring re-
282
Fumia, Edelstone, and Holzman
gional blood flow with radioactive microspheres. J Appl Physiol 1971;31:598. 12. Zar JH. Biostatistical analysis. Englewood Cliffs, New Jersey: Prentice-Hall, 1974. 13. Peeters LLH, Sheldon RE, Jones MD Jr, Makowski EL, Meschia G. Blood flow to fetal organs as a function of arterial oxygen content. AM J 0BSTET GYNECOL 1979; 135:637. 14. Fisher DJ, Heymann MA, Rudolph AM. Fetal myocardial oxygen and carbohydrate consumption during acutely induced hypoxemia. Am J Physiol 1982;242:H657.
October I, 1984 Am J Obstet Gynecol
15. Michalski AH, Lowenstein E, Austen WG, Buckley MJ, Laver MB. Patterns of oxygenation and cardiovascular adjustment to acute, transient normovolemic anemia. Ann Surg 1968;168:946. 16. Edelstone DI, Fumia F. Fetal 0 2 consumption as a function of fetal hematocrit [Abstract]: In: Scientific abstracts of the thirty-first annual meeting of the Society for Gynecologic Investigation, San Francisco: 1984.