- Email: [email protected]
Relationship of fetal oxygen consumption and acid-base balance to fetal hematocrit
Relationship of fetal oxygen consumption and acid-base balance to fetal hernatocrit Daniel I. Edelstone, M.D., Mark E. Caine, M.D., and Fred D. Furnia, M.D. Pittsburgh, Pennsylvania We evaluated the effects of alterations in fetal hematocrit on fetal oxygenation in 1O chronically catheterized fetal lambs. Hematocrit was varied from 10% to 55% by slow isovolemic exchange transfusions with plasma or packed red blood cells obtained freshly from donor fetuses. At each hematocrit studied, we measured umbilical blood flow (Oumb) and the oxygen concentrations in umbilical venous blood (Cuv02 ) and arterial blood (CAo2 ) and calculated fetal oxygen delivery (Oumb · Cuva2), oxygen extraction [(Cuv02 - CAo2)/Cuv0 J, and oxygen consumption [Oumb (Cuva2 - CAo:!)]. Fetal oxygen delivery was maximal at a fetal hematocrit of 33% (mean oxygen delivery = 23 ml of oxygen per minute per kilogram of fetus) and decreased as hematocrit was raised or lowered from that value. Despite these reductions in oxygen delivery, fetal oxygen consumption was relatively stable (at about 7 ml of oxygen per minute per kilogram) at hematocrits ranging from about 16% to 48% because of compensatory increases in fetal oxygen extraction. Regardless of whether oxygen delivery decreased because of anemia or polycythemia, fetal oxygen consumption was maintained as long as oxygen delivery was greater than about 14 ml of oxygen per minute per kilogram of fetus. When oxygen delivery was < 14 ml of oxygen per minute per kilogram, fetal oxygen consumption fell while arterial blood base deficit increased, indicating that oxygen supply was inadequate for fetal oxygen demands. These results indicate that fetal aerobic metabolism can be sustained over a wide range of fetal hematocrits. Furthermore, our data support the concept that the level of fetal oxygen delivery is an important determinant of the adequacy of fetal oxygenation. (AM J 0BSTET GYNECOL 1985;151:844-51.)
During pregnancy, several maternal and fetal complications are known to be associated with alterations in fetal hematocrit. 1-3 These complications include conditions as varied as diabetes mellitus, fetal growth retardation, rhesus isoimmunization, and the twin-totwin transfusion syndrome. The resultant changes in fetal hematocrit and thus hemoglobin concentration produce alterations in both the oxygen concentration and the viscosity of fetal blood. 3 · ' In adult animals, each of these two variables has been shown to affect perfusion and oxygen supply to individual organs as well as to the organism as a whole. 5 · 6 Although these effects must theoretically occur in the fetus as well, there are very few experimental data describing the effects of hematocrit variations on fetal oxygenation. Tenenbaum et al.' found that oxygen delivery to the fetal lamb (umbilical blood flow x umbilical venous blood oxygen concentration) decreased when fetal he-
matocrit was increased 40% above normal. Even though fetal oxygen delivery decreased, the quantity of oxygen consumed by the fetus was maintained by compensatory increases in fetal oxygen extraction. Their studies did not identify the limits of the fetal compensatory response to polycythemia because only one level of polycythemia was studied.' It is likely that if hematocrit had been increased further the limits of compensation would have been exceeded and the rate of fetal aerobic metabolism would have decreased. Furthermore, no studies have been done to define the extent of the fetal adaptive response to anemia. In view of the many clinical conditions associated with fetal anemia and polycythemia, 1-3 it would be important to determine the range of hematocrits over which fetal aerobic metabolism can be sustained. The purpose of our studies was to identify the fetal compensatory responses to varying degrees of anemia and polycythemia and to define the limits of those responses.
From the Department of Obstetrics and Gynecology, University of Pittsburgh School of Medicine, Magee-Womens Hospital. This work was supported by a grant from the National Institutes of Health (HD-16368). Mark E. Caine and Fred D. Fumia were Fellows in Maternal-Fetal Medicine. The Charles A. Hunter, Jr., Award Thesis, presented at the Third Annual Meeting of the American Gynecological and Obstetrical Society, Hot Springs, Virginia, September 5-8, 1984. Reprint requests: Daniel I. Ede/stone, M.D., Department of Obstetrics and Gynecology, Magee-Womens Hospital, Forbes and Halket St., Pittsburgh, PA 15213.
Methods Preparation of animals. We prepared 10 fetal lambs ranging in gestational age from 100 to 130 days (term = 147 days). The surgical techniques used have been described in detaiF· 8 and will be summarized only. After an epidural or spinal anesthestic (tetracaine hydrochloride, 1%) was administered to the mother, we opened the maternal abdomen and uterus and placed
844
Volume 151 Number 7
catheters into a fetal hind limb artery and a vein. These catheters were advanced into the descending aorta and inferior vena cava. Anesthesia for catheter placements in the fetus was provided by lidocaine hydrochloride, 0.5%, administered locally. We then incised the uterus over the area of the fetal abdomen and inserted a catheter into the common umbilical vein. A catheter was also placed in the amnionic sac. In eight of the I 0 fetuses, twins were present; in these cases, we placed catheters into a hind limb artery and a vein in the second twin. We used this second twin as a plasma or red blood cell donor. The maternal abdomen was closed and all catheters were brought out subcutaneously to the ewe's flank where they were stored in a pouch. Two to 5 days after operation, we studied the fetuses of the unanesthetized sheep as the sheep stood unrestrained in a transport cage. Experimental protocol. In each fetus, we measured umbilical blood flow, umbilical venous and arterial blood oxygen concentrations and acid-base balance and calculated fetal oxygen delivery, oxygen extraction, and oxygen consumption at various randomly selected fetal hematocrits ranging from 10% to 55%. To change the fetal hematocrit, we performed isovolemic exchange transfusions by infusing plasma or packed fetal red blood cells through one fetal catheter while withdrawing fetal blood at the same rate from a second catheter. 8 Hematocrits were changed slowly over 1 to 4 hours, during which time fetal arterial blood pressure and heart rate were continuously monitored and recorded. Both variables were stable throughout the study. Plasma for exchange transfusion was obtained freshly either from the donor twin or from an adult sheep, whereas fetal red blood cells were obtained from a donor twin only. When hematocrit was shown to be stable, we withdrew 0.8 ml blood samples simultaneously from the descending aorta and umbilical vein for the analyses of blood gases, pH, base deficit, hematocrit, and oxygen concentration. Blood gases, pH, and base deficit were measured or calculated at 38° C with a blood gas meter (Instrument Laboratories or Corning), hematocrit was determined in duplicate with the microcapillary technique, and blood oxygen concentration was quantified with a Lex-OrCon oxygen analyzer. 7 Immediately after obtaining the blood samples, we injected 1 to 2 x 106 radionuclide-labeled microspheres ( 15 µ.m diameter, labeled with either iodine 125, cerium 141, chromium 51, strontium 85, niobium 95, or scandium 46, New England Nuclear or 3M Co.) into the fetal inferior vena cava while withdrawing a reference blood sample from the descending aorta at 3.9 or 7.9 ml/min. 7 9 We used the higher withdrawal rate in the anemic fetuses to ensure that blood flow measurements were accurate.' 0 Microsphere injections and blood withdrawals did not
Relationship of fetal hematocrit and oxygenation
845
affect fetal arterial blood pressure or heart rate. 9 The volume of blood withdrawn was replaced with a comparable volume of fetal blood so that a constant intravascular volume could be maintained. After obtaining the oxygen and blood flow measurements at this hematocrit, we changed fetal hematocrit to a new level and repeated the flow and oxygen measurements. Each fetus was studied at an average of three (range, two to five) randomly chosen hematocrits. In preliminary experiments on two fetuses, we increased hematocrit first and decreased it second; in two additional fetuses we reversed this order. We found no differences in the responses of fetal oxygenation relating to the order in which fetal hematocrits were changed. Therefore, in the remaining fetuses, we either raised or lowered hematocrit; none of these fetuses were studied at both high and low hematocrits. Using this approach, we could considerably reduce the time involved in varying the hematocrit with exchange transfusions. Preparation of tissues and calculation and analyses of data. When the study was completed, sheep and fetuses were killed with an overdose of sodium pentobarbital. The placenta was removed and the amounts of each radionuclide present in it and in the reference blood samples were measured with standard methods. 9 We computed umbilical (fetal-placental) blood flow (Qumb) as follows: Qumb = (cpmp1aJcpm,.,)Q,.1 where cpmP'" = radioactive counts per minute in the placenta, cpm,., = radioactive counts per minute in the reference arterial blood sample, and Qcer = reference arterial blood flow determined by calibrating the withdrawal pump. Qumb was expressed as milliliters per minute per kilogram of fetal body weight. We calculated three variables of fetal oxygenation, with the use of modifications of the Fick principle"· 12 : Fetal oxygen delivery (fetal Do2) = Qumb (CUVo2) Fetal oxygen consumption (fetal Vo 2) = Qumb(CUV02 - CA02) Fetal oxygen extraction = fetal Vo2 _ Qumb (Cuv 02 - CA 02 ) _ CUVo 2 - CA 02 fetal Do2 Qumb (CUVo 2) CUVo2 where Cuvo2 = umbilical venous blood oxygen concentration and CA02 = arterial (fetal descending aortic) blood oxygen concentration, in milliliters of oxygen per deciliter of blood. Fetal Do2 and Vo 2 were expressed as milliliters of oxygen per minute per kilogram of fetus. We plotted the following dependent variables against fetal hematocrit: umbilical venous and arterial blood oxygen concentrations and oxygen tensions, umbilical
846
Edelstone, Caine, and Furnia
April 1, 1985 Am J Obstet Gynecol
40
16
y•36.7-0.30x
•
•
••
y•2.0+0.23x
30 12
Puv02 (mm Hg)
20
Cuv02 (;ml02J di blood
10
•• 4
0
•
8
•
••
••
•• •
..••• • •
• ••
•
30 y•9.6+0.74x-0.013x'
20
Po 02 (mm Hg) 10
o~~~~~~~~~~~~~~~~~~
0
10
20
30
40
50
60
Fetal hematocrit (%)
Fig. I. Umbilical venous blood Po, (PuvO,, top) and fetal arterial (descending aortic) blood Po 2 (Pa0 2 , bottom) as functions of fetal hematocrit. In this and subsequent figures, data were obtained from IO fetal lambs and were analyzed with analysis of covariance. The regression lines that were drawn take into account responses observed in individual fetuses.
blood flow, and fetal oxygen delivery, oxygen extraction, and oxygen consumption. Data analyses were based on 31 measurements made in the 10 fetuses studied. We analyzed our data with a least-squares method based on analysis of covariance, with which we could generate regression lines that took into account responses seen in individual animals."' Using this method, we could define mathematically the relationships that existed between the various dependent variables and fetal hematocrit. In most cases, equations of the general form, y = a + bx + ex' or y = a + blnx, had the greatest correlation coefficients with the least residuals and thus best described these relations. When quadratic or logarithmic functions did not provide the best fit, linear correlation analyses were done on the raw data. 11
Results On the day of study, fetal lambs were 118 ± 7 days of gestation (mean ± SD), and fetal body weight was 2.1 ± 0.6 kg. Hematocrit before the start of the study was 32% ± 3%. Fetal heart rate and arterial blood pressure (mean and systolic or diastolic) varied< 15% from baseline values over the entire range of hematocrits studied (10% to 55%); neither heart rate nor blood pressure was related to fetal hematocrit. Umbilical venous blood Po, decreased linearly (p < 0.005) as he-
Fetal hemotocrit (%)
Fig. 2. Umbilical venous blood oxygen concentration (Cuv0 2 , top) and fetal arterial (descending aortic) blood oxygen concentration (Ca0 2, bottom) as functions of fetal hematocrit. Cuv 02 and CA 02 are expressed as milliliters of oxygen per deciliter of blood.
matocrit was increased from 10% to 55% (Fig. 1). In contrast, fetal arterial blood Po, did not change significantly until hematocrit was raised above about 40%, at which point arterial Po, fell (p < 0.005). Umbilical venous blood oxygen concentration increased linearly (p < 0.005) with hematocrit (Fig. 2), whereas fetal arterial blood oxygen concentration was greatest at a fetal hematocrit of approximately 34%. At higher and lower hematocrits, arterial blood oxygen concentration decreased. Fig. 3 shows that umbilical blood flow was relatively unaffected by variations in hematocrit from 10% to approximately 40% but fell at greater hematocrits (p < 0.0 l ). Because of the relationships of umbilical venous blood oxygen concentration and umbilical blood flow to fetal hematocrit, fetal oxygen delivery was maximal at a hematocrit of 33% (optimal hematocrit5· 15). We calculated the optimal hematocrit by determining the first-order differential equation (dy/dx) relating fetal oxygen delivery (y) to hematocrit (x) and by solving for x when dy/dx = 0. When hematocrit was increased above or decreased below the optimal hematocrit, oxygen delivery to the fetus declined (Fig. 4,
Relationship of fetal hematocrit and oxygenation
Volume 151 Number 7
400
40 y•l90•2.47x-0.077x2
300
Oumb
r~~in)
200
847
• ••• •
• • • •
• • • • •
30
• •
Y'-19.7•2.59x-0.039x 2
•
D02 fetus ( mlOz) min/Kg
100
20
• • 10
• 0 0~~~,o~~-2~0~~-3~0~~4~0~~-5~0~~~60
•
•• • •
•
•
•
0 Y'l24- 5.9x •0.09x 2
Fig. 3. Relationship of umbilical blood flow (Qumb• in milliliters per minute per kilogram of fetal body weight) to fetal hematocrit.
Comment The data presented in this report establish that fetal oxygen consumption can be maintained over a wide range of fetal hematocrits. In our studies, aerobic metabolism by the fetal lamb was sustained by compensatory increases in oxygen extraction as oxygen delivery decreased. Our data indicate that regardless whether fetal oxygen delivery dropped as a consequence of anemia or as a result of polycythemia, fetal oxidative metabolism was unaffected as long as oxygen delivery remained above a certain minimum, approximately 14
•
••••
•
100
Fetal hemotocrit (%)
top). As compensation for reductions in oxygen delivery that occurred as hematocrits were raised or lowered from the optimum, fetal oxygen extraction (Fig. 4, middle) increased sufficiently to enable fetal oxygen consumption to be relatively constant over the range of hematocrits from about 16% to 48% (Fig. 4, bottom). We were also interested in identifying which of the two variables, oxygen delivery or hematocrit, was more important in predicting the adequacy of fetal oxygenation. Using multiple regression analysis, we found that fetal oxygen delivery, rather than fetal hematocrit, was the better predictor of the level of fetal oxygen consumption (the sum of squares of oxygen consumption versus oxygen delivery was greater than the sum of squares of oxygen consumption versus hematocrit 1'). At a fetal oxygen deliyery of greater than approximately 14 ml of oxygen per minute per kilogram of fetal body weight, fetal oxygen consumption was unaffected by changes in oxygen delivery (Fig. 5, top). When oxygen delivery was reduced below about 14 ml of oxygen per minute per kilogram, regardless whether this response was secondary to anemia or to polycythemia, oxygen consumption by the fetus decreased. Oxygen consumption fell in association with increases in the base deficit in arterial blood (Fig. 5, bottom), indicating that fetal oxygen supply was inadequate for fetal oxygen demands.
.., •
,
• •
•
80
Fetal
60
Oz
Extraction (%)
40
• 20
0 10
V0 2 fetus ( mI02J min/Kg
5
r3.23•0.23x-0.0035x 2
• • • • • • •• • • • •• • • • • • • • • •• • • • •
...
Fetal hemotocrit (%)
Fig. 4. Fetal oxygen delivery (DO.fetus, top), fetal oxygen extraction (middle), and fetal oxygen consumption (VO.fetus, bottom) as functions of fetal hematocrit. Fetal Do2 and Vo2 are expressed as milliliters of oxygen per minute per kilogram of . fetal body weight.
ml of oxygen per minute per kilogram of fetus. This value is about the same as has been reported in studies in which other methods were used to reduce fetal oxygen delivery."· 12 In those studies, when either uterine blood ftow 11 or umbilical blood ftow 12 was reduced, fetal oxygen consumption did not change significantly until oxygen delivery fell to about 14 ml of oxygen per minute per kilogram of fetus (0.6 mmol/L of oxygen per minute per kilogram 11 ). When oxygen delivery fell below that level, fetal oxygen consumption decreased, while the base deficit in fetal arterial blood increased, indicating the development of anaerobic metabolism due to tissue hypoxia. When our results are considered with those just cited, they support the concept that the level of fetal oxygen delivery, rather than the specific
848
Edelstone, Caine, and Furnia
April I, 1985 Am J Obstet Gynecol
10
• • V02 fetus (ml~)
•
5
mm/Kg
...
yo241nx-40.8
0 -15
•
Bose deficit (meg/Ll
• .,
0
• • • • • •• • •
• •
•
+15
O'--~~-"-~~--_J__~~~.J__~~__J
0
10
20
30
40
002 fetus (ml Cilmin/Kg) Fig. 5. Relationships of fetal oxygen consumption (V0 2 fetus, and arterial blood base deficit (bottom) to fetal oxygen delivery (DO, fetus). Fetal Vo, and Do2 are expressed as milliliters of oxygen per minute per kilogram of fetal body weight; base deficit is in milliequivalents per liter of blood.
top)
factor used to affect oxygen delivery adversely, is more important in predicting the adequacy of fetal oxygenation. Oxygen delivery to the fetus is a function of two variables, umbilical blood flow and the oxygen concentration of umbilical venous blood. At the hematocrit associated with maximal fetal oxygen delivery (optimal hematocrit' 15 ), we found that fetal oxygen delivery averaged 23 ml of oxygen per ~inute per kilogram of fetus, a value considerably in excess of the minimum of about 14 ml of oxygen per minute per kilogram necessary for aerobic metabolism. It is of particular interest that the optimal hematocrit we calculated (33%) was about the same as the hematocrit normally found in the fetal lamb (32%). 4 7 In adult animals, previous investigators5 · 6 have observed similar close relationships of optimal hematocrit to the hematocrit normally present in those animals. These observations suggest that, in the fetus as well as in the adult, maintenance of a hematocrit at or near the optimum for oxygen delivery is an important means of ensuring that oxygen supplies to all tissues are adequate for oxidative metabolism. In our studies, when hematocrit was decreased from the optimum, fetal oxygen delivery fell, because the oxygen concentration in umbilical venous blood de-
creased as fetal hemoglobin concentration and blood oxygen-carrying capacity declined, while umbilical blood flow did not change. When we increased fetal hematocrit above the optimal hematocrit, fetal oxygen delivery also decreased, but this decrease was due to a reduction in umbilical blood flow that was proportionally greater than the increase in umbilical venous oxygen concentration due to the rise in fetal hemoglobin concentration. The reduced umbilical flow most likely resulted from increases in blood viscosity that have been shown to occur during fetal polycythemia. Tenenbaum et al. 1 showed that umbilical blood flow fell when fetal hematocrit and viscosity were increased secondary to isovolemic exchange transfusion with adult red blood cells . Although we noted that umbilical venous blood oxygen concentration increased with increasing fetal hematocrit, the rise in blood oxygen concentration was less than expected, because umbilical venous blood Po2 (and thus oxygen saturation 11 ) fell linearly as hematocrit was raised. The reduction in umbilical venous Po2 with increasing hematocrit may have resulted from a change in the relationship of maternal to fetal blood flows within the placenta. Such a change would represent the placental equivalent of an alteration in the matching of ventilation to perfusion in the adult lung. In accord with this idea, variations of hematocrit in the adult dog have been shown to affect arterial blood Po, inversely, an effect that is due to alterations in the relationship of ventilation to perfusion. 16 In contrast to our observation that umbilical venous blood oxygen concentration continued to increase with increasing hematocrit, fetal arterial blood oxygen concentration was maximal at a hematocrit of about 34% and decreased at hematocrits above and below that value. The decreases in fetal arterial blood oxygen concentration reflected the increases in oxygen extraction that occurred as compensations for the reduced oxygen delivery of the anemic and polycythemic states. This relationship can be seen if we examine the equations describing fetal oxygen extraction: fetal Vo, . Feta 1 oxygen extraction = fetal Do 2 · where fetal Vo2 = fetal oxygen consumption and fetal Do 2 = fetal oxygen delivery. Since Feta! Vo, Fetal Do 2
_ Cuv -
CA 02 02 Cuv 02
by rearranging these latter equations, we find that C Ao2 -_ C UVo2
(i _
fetal Vo,) fetal Do2
Volume 151 Number 7
where Cuv 02 = umbilical venous blood oxygen concentration and CA02 = fetal arterial blood oxygen concentration. During anemia, CAo2 decreases because fetal oxygen extraction increases ([l - (fetal Vo2 /fetal D0 2 )] decreases) while Cuv02 decreases as fetal hemoglobin concentration falls. During polycythemia, CA02 also falls, because the increase in fetal oxygen extraction (the decrease in [ 1 - (fetal Vo/fetal D0 2)]) is proportionally greater than the associated increase in Cuvo 2 due to the rise in fetal hemoglobin concentration. Although fetal arterial blood oxygen concentration decreased, there was no change in the arterial blood base deficit over the range of fetal hematocrits from abut 16% to 48%, indicating the absence of tissue hypoxia, anaerobic metabolism, and acidemia. At hematocrits <16% or >48%, which corresponded roughly to a fetal oxygen delivery less than about 14 ml of oxygen per minute per kilogram of fetus, arterial blood base deficit increased in association with reductions in fetal oxygen consumption, confirming that oxygen delivery to some fetal tissues was inadequate for tissue oxygen demands. II In view of our findings, we might speculate on the effects of anemia and polycythemia on the oxygen requirements of the human fetus. Numerous clinical conditions are associated with the development of fetal anemia or polycythemia. I- 3 If the responses of the human fetus to changes in hematocrit are qualitatively similar to those of the lamb fetus, then anemia or polycythemia would cause a reduction in fetal oxygen delivery. During severe anemia or polycythemia, this reduction in oxygen delivery could result in tissue hypoxia and a decrease in fetal oxidative metabolism. Such responses could help to explain why perinatal morbidity and mortality rates are greater in anemic and polycythemic infantsI· 3 than in infants with normal hematocrits. In summary, the results of our experiments show that aerobic metabolism in the fetal lamb can be maintained over a wide range of fetal hematocrits. The principal compensatory response observed was an increase in oxygen extraction when fetal oxygen delivery decreased as a result of alterations in hematocrit. Our results also indicate that, regardless whether fetal oxygen delivery fell as a result of anemia or of polycythemia, fetal oxygenation was normal as long as oxygen delivery remained above a certain level. These data therefore support the concept that the level of fetal oxygen delivery, rather than the specific condition that may be adversely affecting oxygen supply, is the important determinant of the adequacy of oxygenation in the fetal lamb. We thank Anthony Battelli, Patrick Moran, and Betty Steranka for their skillful technical assistance.
Relationship of fetal hematocrit and oxygenation
849
REFERENCES I. Bowman JM. The management of Rh-isoimmunization. Obstet Gynecol 1978;52:1-16. 2. Benirschke K, Kim CK. Multiple pregnancy. N EnglJ Med 1973;288: I 276-84. 3. Ramamurthy RJ, Brans YW. Neonatal polycythemia. I. Criteria for diagnosis and treatment. Pediatrics 1981;68:I68-74. 4. Tenenbaum DG, Piasecki GJ, Oh W, Rosenkrantz TS, Jackson BT. Fetal polycythemia and hyperviscosity: effect on umbilical blood flow and fetal oxygen consumption. AMJ 0BSTETGYNECOL 1983;147:48-51. 5. Shepherd AP, Riedel GL. Optimal hematocrit for oxygenation of canine intestine. Circ Res I 982;5 I :233-40. 6. 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 I 980;238:H54552. 7. Edelstone DI, Holzman IR. Fetal intestinal oxygen consumption at various levels of oxygenation. Am J Physiol l 982;242:H50-4. 8. Furnia FD, Edelstone DI, Holzman IR. Blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit. AMJ 0BSTETGYNECOL 1984;150:274-82. 9. Heymann MA, Payne BD, Hoffman JIE, Rudolph AM. Blood flow measurements wit11 radionuclide-labeled particles. Prog Cardiovasc Dis 1977;20:55-79. 10. Rosenberg AA, Jones MD, Koehler RC, Traystman RJ, Lister G. Precautions for measuring blood flow during anemia with the microsphere technique. Am J Physiol 1983;244:H308-l l. 11. Wilkening RB, Meschia G. Fetal oxygen uptake, oxygenation, and acid-base balance as a function of uterine blood flow. Am J Physiol I 983;244:H749-55. 12. Itskovitz J, LaGamma EF, Rudolph AM. The effect of reducing umbilical blood flow on fetal oxygenation. AMJ 0BSTET GYNECOL I 983; I 45:813-8. 13. Snedecor GW, Cochran WG. Statistical methods. 6th ed. Ames, Iowa: Iowa State Universitv Press, 1967. 14. Brace RA. Fitting straight lines to ~xperimental data. Am J Physiol 1977;233:R94-9. 15. Crowell JW, Smith EE. Determinant of the optimal hematocrit. J Appl Physiol 1967;22:501-4. 16. Michalski AH, Lowenstein E, Austen WG, Buckley MJ, Laver MB. Patterns of oxygenation and cardiovascular adjustment to acute, transient normovolemic anemia. Ann Surg I 968; 168:946-56.
Discussion DR. JOHN PATRICK, London, Ontario, Canada. Was your experimental design the same for fetuses with normal hematocrits as it was for fetuses with low or high hematocrits? This concerns me because the experiment itself could alter many fetal controls, particularly cardiovascular and hormonal controls. Such alterations could considerably alter overall oxygen consumption. Second, I think your data might be relevant to fetal anemia during human pregnancy but I am worried that the injection of blood with a high hematocrit to a normoxic fetus does not really reflect the kind of situation we see in growth-retarded fetal lambs or humans. Could you comment on how your observations might relate to a probable change in the distribution of the fetal circulation under conditions of fetal growth retardation. We have certainly all seen fetal lambs that were small and growth retarded, that had high he-
850
Edelstone, Caine, and Furnia
matocrits with normal oxygen content and low Po 2 , and that exhibited normal cardiovascular and brain patterns. DR. ROBERT C. CEFALO, Chapel Hill, North Carolina. Recently we performed an isovolemic exchange transfusion of the fetus with the erythrocyte substitute Fluosol-DA and brought down the hematocrit to levels similar to those you saw. The respiratory/cardiovascular system parameters measured in the fetus all remained within the normal range even with the hematocrit as low as 10% or 12%. Interestingly, whether the maternal ewe was hyperoxygenated or in ambient air, in the fetus the Po 2 did not increase above that which we would see in a normovolemic fetus with a hematocrit of 34% even though the hematocrit was in the range of 10 vol/ 100 ml. My question would be what was the state of the mother? Did she receive ambient air or was she oxygenated? DR. EDGAR L. MAKOWSKI, Denver, Colorado. Dr. Edelstone, there is one question that I have, and I noted the data were conspicuously absent from your organ flows. Have you had the opportunity to quantitate hepatic flow under these circumstances? I realize how difficult that is because of the complexities of the hepatic vascular bed. DR. SEYMOUR L. ROMNEY, New York, New York. Are you saying basically that the fetus in utero has a totally oxidated metabolic commitment comparable to that of the adult and that there are no differences between adult oxygenated metabolism and that of the fetus, which I thought was anaerobic metabolism? If this is so, what is the range of change in pH? DR. THOMAS H. KIRSCHBAUM, Bethesda, Maryland. Dr. Edelstone, in the illustration in which you plotted fetal V0 2 as a function of hematocrit, it was your contention that oxygen consumption declined at hematocrit levels >50%. It seemed to me that conclusion was based on a single data point which may have stemmed from the decision to use a parabola to fit the aggregate of data more than anything else. Would you comment on the certainty that underlies that? DR. JACK N. BLECHNER, Farmington, Connecticut. Dr. Edelstone, I think you have clearly elucidated the aspect of the fetal ability to compensate for physiologic deficiency. My question is what do you think is the mechanism of the compensatory increase in fetal oxygen extraction for the fetus as a whole above critical oxygen delivery levels? Do you have any thoughts on that? DR. EDELSTONE (Closing). The fetuses we used in our experiments recovered well from operation as indicated by normal respiratory gases, pH, and hematocrits in both arterial and umbilical venous blood. There were, however, two growth-retarded fetuses (which are not included among the results I presented today) that did not recover well from operation. After operation they had high hematocrits and relatively low umbilical
April I, 1985 Am J Obstet Gynecol
arterial and venous blood Po 2 values. The blood pH, Pco 2 , and base deficit in these animals, however, were normal. Since we were interested in the relationship of hematocrit to fetal oxygenation in normally grown as well as in undergrown fetuses, we studied these two animals, as follows. We measured umbilical blood flow and fetal oxygen delivery, oxygen extraction, and oxygen consumption at the initial high hematocrits and at several other hematocrits ranging down to 20%. Fetal hematocrit was reduced by isovolemic exchange transfusions with plasma obtained from a donor fetus (see Methods). As hematocrit was decreased toward normal levels in these growth-retarded animals, arterial and umbilical venous blood Po2 values, umbilical blood flow, fetal oxygen delivery, and fetal oxygen consumption all increased. The observations from these two growthretarded fetuses, when combined with the observations we made in normal fetuses, strongly suggest that there is an optimal hematocrit for oxygenation in the fetal lamb, irrespective whether the fetus is normally grown or undergrown in utero. In response to the question concerning whether the metabolic or hormonal environment of the fetus was affected by our changing the fetal hematocrit, it is likely that the in utero environment of the fetus was altered during our experiments. Nonetheless, the purpose of our studies was to assess the net effect of changes in fetal hematocrit on the intact fetus. To eliminate a potential systematic error due to the direction in which the hematocrit was altered, we increased hematocrit first in half of the fetuses and decreased hematocrit first in the remaining fetuses after determining that there were no differences in the response of fetal oxygenation to alterations in hematocrit relating to the order in which hematocrits were changed (see Methods). Studies were begun only after the fetus was found to be in a stable condition as exemplified by a constant arterial blood pressure, heart rate, and arterial blood gases and pH. In these studies, the mother was unanesthetized and breathing room air. Concerning the response of hepatic blood flow to changes in fetal hematocrit, we have data relating hepatic blood flow and oxygen delivery to fetal hematocrit. In fetal life, hepatic blood flow is derived from three sources: hepatic arterial blood flow (5% of the total flow), portal venous blood flow (20% of the total), and umbilical blood flow (the remaining 75%). When we increased hematocrit from I 0% to about 40% to 45%, blood flow to the liver varied little. At hematocrits >40% to 45%, however, hepatic blood flow decreased substantially. At these high hematocrits, hepatic blood flow decreased for two reasons: (1) total umbilical blood flow decreased at hematocrits above 40% to 45% and (2) the fraction of umbilical blood flow shunted past the liver through the ductus venosus increased. Because of this response of hepatic perfusion to hematocrit, oxygen delivery to the fetal liver was maximal at a hematocrit of 32% and decreased when hematocrits were raised or lowered from that value. This optimal hematocrit for the fetal liver is similar to the optimal
Volume 151 Number 7
hematocrits that we have found for the following organs: gastrointestinal tract organs, spleen, kidneys, placenta, and carcass tissues (skeletal muscle, skin, and bones). It is generally accepted that the fetus functions in an aerobic environment, even though fetal arterial blood Po2 is normally low relative to that of the adult. Fetal oxygen consumption in several species, including the human, ranges from 7 to 10 ml of oxygen per minute per kilogram of body weight. This rate is approximately double that of the adult non pregnant or pregnant animal or human. Despite the greater oxygen consumption rate in the fetus, there is a considerable oxygen reserve, since oxygen delivery to the fetus exceeds oxygen demand by the fetus by about threefold. This oxygen reserve enables the fetus to compensate for variations in its oxygen delivery that may result, for example, from changes in fetal hematocrit or alterations in umbilical or uterine blood flow. Although several fetal responses to stress are similar to those observed in the human, it is quite risky to try to extrapolate results of studies obtained on the lamb fetus to those on the human fetus. In a few situations, extrapolation from experimental data to clinical situations may be warranted, but in many instances this extrapolation may not be valid because of species differences. For example, the basal hematocrit in the human fetus averages 45%, whereas the lamb fetus has a mean hematocrit of about 32%. In our experiments, oxygen consumption by the fetal lamb was stable over the range of hematocrits from about 16% to 48% but decreased when hematocrits were increased above approximately 48% or decreased below 16% (comparable data for the human fetus are not available). We could not identify the precise he-
Relationship of fetal hematocrit and oxygenation
851
matocrit associated with decreases in oxygen consumption, because curved lines described the relationship of oxygen consumption to fetal hematocrit. We generated these curves with analysis of covariance with which we could take into account individual animal responses. The equations for the curves selected were those having the highest regression coefficients associated with the least residual sums of squares. Tested statistically, these equations best described our data. The observation that oxygen consumption fell at hematocrits above about 48% is not based on data from a single animal. Rather, it is based on the observation that oxygen consumption in several fetuses decreased at relatively high hematocrits. Fetal oxygen consumption was maintained at a value of about 7 ml of oxygen per minute per kilogram at hematocrits between about 16% and 48% because of alterations in oxygen extraction. Oxygen extraction is that quantity of oxygen consumed by the fetus, expressed as a fraction of the total oxygen delivered to the fetus. Changes in oxygen extraction result from alterations in capillary diffusion variables such as the capillary-to-cell Po2 gradient or the capillary surface area available for oxygen exchange (capillary density). In fetal life, all organs, with the exception of the heart and the brain, compensate for reductions in oxygen delivery by altering oxygen extraction (in the brain and heart, perfusion increases as compensation, while oxygen extraction does not change substantially). It is not known whether these changes in oxygen extraction in the fetus are primarily due to adjustments in capillary density or to modifications of the capillary-to-cell Po 2 gradient. In the adult animal, adjustments in oxygen extraction occur mainly as a result of increases in capillary density.
During pregnancy, several maternal and fetal complications are known to be associated with alterations in fetal hematocrit. 1-3 These complications include conditions as varied as diabetes mellitus, fetal growth retardation, rhesus isoimmunization, and the twin-totwin transfusion syndrome. The resultant changes in fetal hematocrit and thus hemoglobin concentration produce alterations in both the oxygen concentration and the viscosity of fetal blood. 3 · ' In adult animals, each of these two variables has been shown to affect perfusion and oxygen supply to individual organs as well as to the organism as a whole. 5 · 6 Although these effects must theoretically occur in the fetus as well, there are very few experimental data describing the effects of hematocrit variations on fetal oxygenation. Tenenbaum et al.' found that oxygen delivery to the fetal lamb (umbilical blood flow x umbilical venous blood oxygen concentration) decreased when fetal he-
matocrit was increased 40% above normal. Even though fetal oxygen delivery decreased, the quantity of oxygen consumed by the fetus was maintained by compensatory increases in fetal oxygen extraction. Their studies did not identify the limits of the fetal compensatory response to polycythemia because only one level of polycythemia was studied.' It is likely that if hematocrit had been increased further the limits of compensation would have been exceeded and the rate of fetal aerobic metabolism would have decreased. Furthermore, no studies have been done to define the extent of the fetal adaptive response to anemia. In view of the many clinical conditions associated with fetal anemia and polycythemia, 1-3 it would be important to determine the range of hematocrits over which fetal aerobic metabolism can be sustained. The purpose of our studies was to identify the fetal compensatory responses to varying degrees of anemia and polycythemia and to define the limits of those responses.
From the Department of Obstetrics and Gynecology, University of Pittsburgh School of Medicine, Magee-Womens Hospital. This work was supported by a grant from the National Institutes of Health (HD-16368). Mark E. Caine and Fred D. Fumia were Fellows in Maternal-Fetal Medicine. The Charles A. Hunter, Jr., Award Thesis, presented at the Third Annual Meeting of the American Gynecological and Obstetrical Society, Hot Springs, Virginia, September 5-8, 1984. Reprint requests: Daniel I. Ede/stone, M.D., Department of Obstetrics and Gynecology, Magee-Womens Hospital, Forbes and Halket St., Pittsburgh, PA 15213.
Methods Preparation of animals. We prepared 10 fetal lambs ranging in gestational age from 100 to 130 days (term = 147 days). The surgical techniques used have been described in detaiF· 8 and will be summarized only. After an epidural or spinal anesthestic (tetracaine hydrochloride, 1%) was administered to the mother, we opened the maternal abdomen and uterus and placed
844
Volume 151 Number 7
catheters into a fetal hind limb artery and a vein. These catheters were advanced into the descending aorta and inferior vena cava. Anesthesia for catheter placements in the fetus was provided by lidocaine hydrochloride, 0.5%, administered locally. We then incised the uterus over the area of the fetal abdomen and inserted a catheter into the common umbilical vein. A catheter was also placed in the amnionic sac. In eight of the I 0 fetuses, twins were present; in these cases, we placed catheters into a hind limb artery and a vein in the second twin. We used this second twin as a plasma or red blood cell donor. The maternal abdomen was closed and all catheters were brought out subcutaneously to the ewe's flank where they were stored in a pouch. Two to 5 days after operation, we studied the fetuses of the unanesthetized sheep as the sheep stood unrestrained in a transport cage. Experimental protocol. In each fetus, we measured umbilical blood flow, umbilical venous and arterial blood oxygen concentrations and acid-base balance and calculated fetal oxygen delivery, oxygen extraction, and oxygen consumption at various randomly selected fetal hematocrits ranging from 10% to 55%. To change the fetal hematocrit, we performed isovolemic exchange transfusions by infusing plasma or packed fetal red blood cells through one fetal catheter while withdrawing fetal blood at the same rate from a second catheter. 8 Hematocrits were changed slowly over 1 to 4 hours, during which time fetal arterial blood pressure and heart rate were continuously monitored and recorded. Both variables were stable throughout the study. Plasma for exchange transfusion was obtained freshly either from the donor twin or from an adult sheep, whereas fetal red blood cells were obtained from a donor twin only. When hematocrit was shown to be stable, we withdrew 0.8 ml blood samples simultaneously from the descending aorta and umbilical vein for the analyses of blood gases, pH, base deficit, hematocrit, and oxygen concentration. Blood gases, pH, and base deficit were measured or calculated at 38° C with a blood gas meter (Instrument Laboratories or Corning), hematocrit was determined in duplicate with the microcapillary technique, and blood oxygen concentration was quantified with a Lex-OrCon oxygen analyzer. 7 Immediately after obtaining the blood samples, we injected 1 to 2 x 106 radionuclide-labeled microspheres ( 15 µ.m diameter, labeled with either iodine 125, cerium 141, chromium 51, strontium 85, niobium 95, or scandium 46, New England Nuclear or 3M Co.) into the fetal inferior vena cava while withdrawing a reference blood sample from the descending aorta at 3.9 or 7.9 ml/min. 7 9 We used the higher withdrawal rate in the anemic fetuses to ensure that blood flow measurements were accurate.' 0 Microsphere injections and blood withdrawals did not
Relationship of fetal hematocrit and oxygenation
845
affect fetal arterial blood pressure or heart rate. 9 The volume of blood withdrawn was replaced with a comparable volume of fetal blood so that a constant intravascular volume could be maintained. After obtaining the oxygen and blood flow measurements at this hematocrit, we changed fetal hematocrit to a new level and repeated the flow and oxygen measurements. Each fetus was studied at an average of three (range, two to five) randomly chosen hematocrits. In preliminary experiments on two fetuses, we increased hematocrit first and decreased it second; in two additional fetuses we reversed this order. We found no differences in the responses of fetal oxygenation relating to the order in which fetal hematocrits were changed. Therefore, in the remaining fetuses, we either raised or lowered hematocrit; none of these fetuses were studied at both high and low hematocrits. Using this approach, we could considerably reduce the time involved in varying the hematocrit with exchange transfusions. Preparation of tissues and calculation and analyses of data. When the study was completed, sheep and fetuses were killed with an overdose of sodium pentobarbital. The placenta was removed and the amounts of each radionuclide present in it and in the reference blood samples were measured with standard methods. 9 We computed umbilical (fetal-placental) blood flow (Qumb) as follows: Qumb = (cpmp1aJcpm,.,)Q,.1 where cpmP'" = radioactive counts per minute in the placenta, cpm,., = radioactive counts per minute in the reference arterial blood sample, and Qcer = reference arterial blood flow determined by calibrating the withdrawal pump. Qumb was expressed as milliliters per minute per kilogram of fetal body weight. We calculated three variables of fetal oxygenation, with the use of modifications of the Fick principle"· 12 : Fetal oxygen delivery (fetal Do2) = Qumb (CUVo2) Fetal oxygen consumption (fetal Vo 2) = Qumb(CUV02 - CA02) Fetal oxygen extraction = fetal Vo2 _ Qumb (Cuv 02 - CA 02 ) _ CUVo 2 - CA 02 fetal Do2 Qumb (CUVo 2) CUVo2 where Cuvo2 = umbilical venous blood oxygen concentration and CA02 = arterial (fetal descending aortic) blood oxygen concentration, in milliliters of oxygen per deciliter of blood. Fetal Do2 and Vo 2 were expressed as milliliters of oxygen per minute per kilogram of fetus. We plotted the following dependent variables against fetal hematocrit: umbilical venous and arterial blood oxygen concentrations and oxygen tensions, umbilical
846
Edelstone, Caine, and Furnia
April 1, 1985 Am J Obstet Gynecol
40
16
y•36.7-0.30x
•
•
••
y•2.0+0.23x
30 12
Puv02 (mm Hg)
20
Cuv02 (;ml02J di blood
10
•• 4
0
•
8
•
••
••
•• •
..••• • •
• ••
•
30 y•9.6+0.74x-0.013x'
20
Po 02 (mm Hg) 10
o~~~~~~~~~~~~~~~~~~
0
10
20
30
40
50
60
Fetal hematocrit (%)
Fig. I. Umbilical venous blood Po, (PuvO,, top) and fetal arterial (descending aortic) blood Po 2 (Pa0 2 , bottom) as functions of fetal hematocrit. In this and subsequent figures, data were obtained from IO fetal lambs and were analyzed with analysis of covariance. The regression lines that were drawn take into account responses observed in individual fetuses.
blood flow, and fetal oxygen delivery, oxygen extraction, and oxygen consumption. Data analyses were based on 31 measurements made in the 10 fetuses studied. We analyzed our data with a least-squares method based on analysis of covariance, with which we could generate regression lines that took into account responses seen in individual animals."' Using this method, we could define mathematically the relationships that existed between the various dependent variables and fetal hematocrit. In most cases, equations of the general form, y = a + bx + ex' or y = a + blnx, had the greatest correlation coefficients with the least residuals and thus best described these relations. When quadratic or logarithmic functions did not provide the best fit, linear correlation analyses were done on the raw data. 11
Results On the day of study, fetal lambs were 118 ± 7 days of gestation (mean ± SD), and fetal body weight was 2.1 ± 0.6 kg. Hematocrit before the start of the study was 32% ± 3%. Fetal heart rate and arterial blood pressure (mean and systolic or diastolic) varied< 15% from baseline values over the entire range of hematocrits studied (10% to 55%); neither heart rate nor blood pressure was related to fetal hematocrit. Umbilical venous blood Po, decreased linearly (p < 0.005) as he-
Fetal hemotocrit (%)
Fig. 2. Umbilical venous blood oxygen concentration (Cuv0 2 , top) and fetal arterial (descending aortic) blood oxygen concentration (Ca0 2, bottom) as functions of fetal hematocrit. Cuv 02 and CA 02 are expressed as milliliters of oxygen per deciliter of blood.
matocrit was increased from 10% to 55% (Fig. 1). In contrast, fetal arterial blood Po, did not change significantly until hematocrit was raised above about 40%, at which point arterial Po, fell (p < 0.005). Umbilical venous blood oxygen concentration increased linearly (p < 0.005) with hematocrit (Fig. 2), whereas fetal arterial blood oxygen concentration was greatest at a fetal hematocrit of approximately 34%. At higher and lower hematocrits, arterial blood oxygen concentration decreased. Fig. 3 shows that umbilical blood flow was relatively unaffected by variations in hematocrit from 10% to approximately 40% but fell at greater hematocrits (p < 0.0 l ). Because of the relationships of umbilical venous blood oxygen concentration and umbilical blood flow to fetal hematocrit, fetal oxygen delivery was maximal at a hematocrit of 33% (optimal hematocrit5· 15). We calculated the optimal hematocrit by determining the first-order differential equation (dy/dx) relating fetal oxygen delivery (y) to hematocrit (x) and by solving for x when dy/dx = 0. When hematocrit was increased above or decreased below the optimal hematocrit, oxygen delivery to the fetus declined (Fig. 4,
Relationship of fetal hematocrit and oxygenation
Volume 151 Number 7
400
40 y•l90•2.47x-0.077x2
300
Oumb
r~~in)
200
847
• ••• •
• • • •
• • • • •
30
• •
Y'-19.7•2.59x-0.039x 2
•
D02 fetus ( mlOz) min/Kg
100
20
• • 10
• 0 0~~~,o~~-2~0~~-3~0~~4~0~~-5~0~~~60
•
•• • •
•
•
•
0 Y'l24- 5.9x •0.09x 2
Fig. 3. Relationship of umbilical blood flow (Qumb• in milliliters per minute per kilogram of fetal body weight) to fetal hematocrit.
Comment The data presented in this report establish that fetal oxygen consumption can be maintained over a wide range of fetal hematocrits. In our studies, aerobic metabolism by the fetal lamb was sustained by compensatory increases in oxygen extraction as oxygen delivery decreased. Our data indicate that regardless whether fetal oxygen delivery dropped as a consequence of anemia or as a result of polycythemia, fetal oxidative metabolism was unaffected as long as oxygen delivery remained above a certain minimum, approximately 14
•
••••
•
100
Fetal hemotocrit (%)
top). As compensation for reductions in oxygen delivery that occurred as hematocrits were raised or lowered from the optimum, fetal oxygen extraction (Fig. 4, middle) increased sufficiently to enable fetal oxygen consumption to be relatively constant over the range of hematocrits from about 16% to 48% (Fig. 4, bottom). We were also interested in identifying which of the two variables, oxygen delivery or hematocrit, was more important in predicting the adequacy of fetal oxygenation. Using multiple regression analysis, we found that fetal oxygen delivery, rather than fetal hematocrit, was the better predictor of the level of fetal oxygen consumption (the sum of squares of oxygen consumption versus oxygen delivery was greater than the sum of squares of oxygen consumption versus hematocrit 1'). At a fetal oxygen deliyery of greater than approximately 14 ml of oxygen per minute per kilogram of fetal body weight, fetal oxygen consumption was unaffected by changes in oxygen delivery (Fig. 5, top). When oxygen delivery was reduced below about 14 ml of oxygen per minute per kilogram, regardless whether this response was secondary to anemia or to polycythemia, oxygen consumption by the fetus decreased. Oxygen consumption fell in association with increases in the base deficit in arterial blood (Fig. 5, bottom), indicating that fetal oxygen supply was inadequate for fetal oxygen demands.
.., •
,
• •
•
80
Fetal
60
Oz
Extraction (%)
40
• 20
0 10
V0 2 fetus ( mI02J min/Kg
5
r3.23•0.23x-0.0035x 2
• • • • • • •• • • • •• • • • • • • • • •• • • • •
...
Fetal hemotocrit (%)
Fig. 4. Fetal oxygen delivery (DO.fetus, top), fetal oxygen extraction (middle), and fetal oxygen consumption (VO.fetus, bottom) as functions of fetal hematocrit. Fetal Do2 and Vo2 are expressed as milliliters of oxygen per minute per kilogram of . fetal body weight.
ml of oxygen per minute per kilogram of fetus. This value is about the same as has been reported in studies in which other methods were used to reduce fetal oxygen delivery."· 12 In those studies, when either uterine blood ftow 11 or umbilical blood ftow 12 was reduced, fetal oxygen consumption did not change significantly until oxygen delivery fell to about 14 ml of oxygen per minute per kilogram of fetus (0.6 mmol/L of oxygen per minute per kilogram 11 ). When oxygen delivery fell below that level, fetal oxygen consumption decreased, while the base deficit in fetal arterial blood increased, indicating the development of anaerobic metabolism due to tissue hypoxia. When our results are considered with those just cited, they support the concept that the level of fetal oxygen delivery, rather than the specific
848
Edelstone, Caine, and Furnia
April I, 1985 Am J Obstet Gynecol
10
• • V02 fetus (ml~)
•
5
mm/Kg
...
yo241nx-40.8
0 -15
•
Bose deficit (meg/Ll
• .,
0
• • • • • •• • •
• •
•
+15
O'--~~-"-~~--_J__~~~.J__~~__J
0
10
20
30
40
002 fetus (ml Cilmin/Kg) Fig. 5. Relationships of fetal oxygen consumption (V0 2 fetus, and arterial blood base deficit (bottom) to fetal oxygen delivery (DO, fetus). Fetal Vo, and Do2 are expressed as milliliters of oxygen per minute per kilogram of fetal body weight; base deficit is in milliequivalents per liter of blood.
top)
factor used to affect oxygen delivery adversely, is more important in predicting the adequacy of fetal oxygenation. Oxygen delivery to the fetus is a function of two variables, umbilical blood flow and the oxygen concentration of umbilical venous blood. At the hematocrit associated with maximal fetal oxygen delivery (optimal hematocrit' 15 ), we found that fetal oxygen delivery averaged 23 ml of oxygen per ~inute per kilogram of fetus, a value considerably in excess of the minimum of about 14 ml of oxygen per minute per kilogram necessary for aerobic metabolism. It is of particular interest that the optimal hematocrit we calculated (33%) was about the same as the hematocrit normally found in the fetal lamb (32%). 4 7 In adult animals, previous investigators5 · 6 have observed similar close relationships of optimal hematocrit to the hematocrit normally present in those animals. These observations suggest that, in the fetus as well as in the adult, maintenance of a hematocrit at or near the optimum for oxygen delivery is an important means of ensuring that oxygen supplies to all tissues are adequate for oxidative metabolism. In our studies, when hematocrit was decreased from the optimum, fetal oxygen delivery fell, because the oxygen concentration in umbilical venous blood de-
creased as fetal hemoglobin concentration and blood oxygen-carrying capacity declined, while umbilical blood flow did not change. When we increased fetal hematocrit above the optimal hematocrit, fetal oxygen delivery also decreased, but this decrease was due to a reduction in umbilical blood flow that was proportionally greater than the increase in umbilical venous oxygen concentration due to the rise in fetal hemoglobin concentration. The reduced umbilical flow most likely resulted from increases in blood viscosity that have been shown to occur during fetal polycythemia. Tenenbaum et al. 1 showed that umbilical blood flow fell when fetal hematocrit and viscosity were increased secondary to isovolemic exchange transfusion with adult red blood cells . Although we noted that umbilical venous blood oxygen concentration increased with increasing fetal hematocrit, the rise in blood oxygen concentration was less than expected, because umbilical venous blood Po2 (and thus oxygen saturation 11 ) fell linearly as hematocrit was raised. The reduction in umbilical venous Po2 with increasing hematocrit may have resulted from a change in the relationship of maternal to fetal blood flows within the placenta. Such a change would represent the placental equivalent of an alteration in the matching of ventilation to perfusion in the adult lung. In accord with this idea, variations of hematocrit in the adult dog have been shown to affect arterial blood Po, inversely, an effect that is due to alterations in the relationship of ventilation to perfusion. 16 In contrast to our observation that umbilical venous blood oxygen concentration continued to increase with increasing hematocrit, fetal arterial blood oxygen concentration was maximal at a hematocrit of about 34% and decreased at hematocrits above and below that value. The decreases in fetal arterial blood oxygen concentration reflected the increases in oxygen extraction that occurred as compensations for the reduced oxygen delivery of the anemic and polycythemic states. This relationship can be seen if we examine the equations describing fetal oxygen extraction: fetal Vo, . Feta 1 oxygen extraction = fetal Do 2 · where fetal Vo2 = fetal oxygen consumption and fetal Do 2 = fetal oxygen delivery. Since Feta! Vo, Fetal Do 2
_ Cuv -
CA 02 02 Cuv 02
by rearranging these latter equations, we find that C Ao2 -_ C UVo2
(i _
fetal Vo,) fetal Do2
Volume 151 Number 7
where Cuv 02 = umbilical venous blood oxygen concentration and CA02 = fetal arterial blood oxygen concentration. During anemia, CAo2 decreases because fetal oxygen extraction increases ([l - (fetal Vo2 /fetal D0 2 )] decreases) while Cuv02 decreases as fetal hemoglobin concentration falls. During polycythemia, CA02 also falls, because the increase in fetal oxygen extraction (the decrease in [ 1 - (fetal Vo/fetal D0 2)]) is proportionally greater than the associated increase in Cuvo 2 due to the rise in fetal hemoglobin concentration. Although fetal arterial blood oxygen concentration decreased, there was no change in the arterial blood base deficit over the range of fetal hematocrits from abut 16% to 48%, indicating the absence of tissue hypoxia, anaerobic metabolism, and acidemia. At hematocrits <16% or >48%, which corresponded roughly to a fetal oxygen delivery less than about 14 ml of oxygen per minute per kilogram of fetus, arterial blood base deficit increased in association with reductions in fetal oxygen consumption, confirming that oxygen delivery to some fetal tissues was inadequate for tissue oxygen demands. II In view of our findings, we might speculate on the effects of anemia and polycythemia on the oxygen requirements of the human fetus. Numerous clinical conditions are associated with the development of fetal anemia or polycythemia. I- 3 If the responses of the human fetus to changes in hematocrit are qualitatively similar to those of the lamb fetus, then anemia or polycythemia would cause a reduction in fetal oxygen delivery. During severe anemia or polycythemia, this reduction in oxygen delivery could result in tissue hypoxia and a decrease in fetal oxidative metabolism. Such responses could help to explain why perinatal morbidity and mortality rates are greater in anemic and polycythemic infantsI· 3 than in infants with normal hematocrits. In summary, the results of our experiments show that aerobic metabolism in the fetal lamb can be maintained over a wide range of fetal hematocrits. The principal compensatory response observed was an increase in oxygen extraction when fetal oxygen delivery decreased as a result of alterations in hematocrit. Our results also indicate that, regardless whether fetal oxygen delivery fell as a result of anemia or of polycythemia, fetal oxygenation was normal as long as oxygen delivery remained above a certain level. These data therefore support the concept that the level of fetal oxygen delivery, rather than the specific condition that may be adversely affecting oxygen supply, is the important determinant of the adequacy of oxygenation in the fetal lamb. We thank Anthony Battelli, Patrick Moran, and Betty Steranka for their skillful technical assistance.
Relationship of fetal hematocrit and oxygenation
849
REFERENCES I. Bowman JM. The management of Rh-isoimmunization. Obstet Gynecol 1978;52:1-16. 2. Benirschke K, Kim CK. Multiple pregnancy. N EnglJ Med 1973;288: I 276-84. 3. Ramamurthy RJ, Brans YW. Neonatal polycythemia. I. Criteria for diagnosis and treatment. Pediatrics 1981;68:I68-74. 4. Tenenbaum DG, Piasecki GJ, Oh W, Rosenkrantz TS, Jackson BT. Fetal polycythemia and hyperviscosity: effect on umbilical blood flow and fetal oxygen consumption. AMJ 0BSTETGYNECOL 1983;147:48-51. 5. Shepherd AP, Riedel GL. Optimal hematocrit for oxygenation of canine intestine. Circ Res I 982;5 I :233-40. 6. 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 I 980;238:H54552. 7. Edelstone DI, Holzman IR. Fetal intestinal oxygen consumption at various levels of oxygenation. Am J Physiol l 982;242:H50-4. 8. Furnia FD, Edelstone DI, Holzman IR. Blood flow and oxygen delivery to fetal organs as functions of fetal hematocrit. AMJ 0BSTETGYNECOL 1984;150:274-82. 9. Heymann MA, Payne BD, Hoffman JIE, Rudolph AM. Blood flow measurements wit11 radionuclide-labeled particles. Prog Cardiovasc Dis 1977;20:55-79. 10. Rosenberg AA, Jones MD, Koehler RC, Traystman RJ, Lister G. Precautions for measuring blood flow during anemia with the microsphere technique. Am J Physiol 1983;244:H308-l l. 11. Wilkening RB, Meschia G. Fetal oxygen uptake, oxygenation, and acid-base balance as a function of uterine blood flow. Am J Physiol I 983;244:H749-55. 12. Itskovitz J, LaGamma EF, Rudolph AM. The effect of reducing umbilical blood flow on fetal oxygenation. AMJ 0BSTET GYNECOL I 983; I 45:813-8. 13. Snedecor GW, Cochran WG. Statistical methods. 6th ed. Ames, Iowa: Iowa State Universitv Press, 1967. 14. Brace RA. Fitting straight lines to ~xperimental data. Am J Physiol 1977;233:R94-9. 15. Crowell JW, Smith EE. Determinant of the optimal hematocrit. J Appl Physiol 1967;22:501-4. 16. Michalski AH, Lowenstein E, Austen WG, Buckley MJ, Laver MB. Patterns of oxygenation and cardiovascular adjustment to acute, transient normovolemic anemia. Ann Surg I 968; 168:946-56.
Discussion DR. JOHN PATRICK, London, Ontario, Canada. Was your experimental design the same for fetuses with normal hematocrits as it was for fetuses with low or high hematocrits? This concerns me because the experiment itself could alter many fetal controls, particularly cardiovascular and hormonal controls. Such alterations could considerably alter overall oxygen consumption. Second, I think your data might be relevant to fetal anemia during human pregnancy but I am worried that the injection of blood with a high hematocrit to a normoxic fetus does not really reflect the kind of situation we see in growth-retarded fetal lambs or humans. Could you comment on how your observations might relate to a probable change in the distribution of the fetal circulation under conditions of fetal growth retardation. We have certainly all seen fetal lambs that were small and growth retarded, that had high he-
850
Edelstone, Caine, and Furnia
matocrits with normal oxygen content and low Po 2 , and that exhibited normal cardiovascular and brain patterns. DR. ROBERT C. CEFALO, Chapel Hill, North Carolina. Recently we performed an isovolemic exchange transfusion of the fetus with the erythrocyte substitute Fluosol-DA and brought down the hematocrit to levels similar to those you saw. The respiratory/cardiovascular system parameters measured in the fetus all remained within the normal range even with the hematocrit as low as 10% or 12%. Interestingly, whether the maternal ewe was hyperoxygenated or in ambient air, in the fetus the Po 2 did not increase above that which we would see in a normovolemic fetus with a hematocrit of 34% even though the hematocrit was in the range of 10 vol/ 100 ml. My question would be what was the state of the mother? Did she receive ambient air or was she oxygenated? DR. EDGAR L. MAKOWSKI, Denver, Colorado. Dr. Edelstone, there is one question that I have, and I noted the data were conspicuously absent from your organ flows. Have you had the opportunity to quantitate hepatic flow under these circumstances? I realize how difficult that is because of the complexities of the hepatic vascular bed. DR. SEYMOUR L. ROMNEY, New York, New York. Are you saying basically that the fetus in utero has a totally oxidated metabolic commitment comparable to that of the adult and that there are no differences between adult oxygenated metabolism and that of the fetus, which I thought was anaerobic metabolism? If this is so, what is the range of change in pH? DR. THOMAS H. KIRSCHBAUM, Bethesda, Maryland. Dr. Edelstone, in the illustration in which you plotted fetal V0 2 as a function of hematocrit, it was your contention that oxygen consumption declined at hematocrit levels >50%. It seemed to me that conclusion was based on a single data point which may have stemmed from the decision to use a parabola to fit the aggregate of data more than anything else. Would you comment on the certainty that underlies that? DR. JACK N. BLECHNER, Farmington, Connecticut. Dr. Edelstone, I think you have clearly elucidated the aspect of the fetal ability to compensate for physiologic deficiency. My question is what do you think is the mechanism of the compensatory increase in fetal oxygen extraction for the fetus as a whole above critical oxygen delivery levels? Do you have any thoughts on that? DR. EDELSTONE (Closing). The fetuses we used in our experiments recovered well from operation as indicated by normal respiratory gases, pH, and hematocrits in both arterial and umbilical venous blood. There were, however, two growth-retarded fetuses (which are not included among the results I presented today) that did not recover well from operation. After operation they had high hematocrits and relatively low umbilical
April I, 1985 Am J Obstet Gynecol
arterial and venous blood Po 2 values. The blood pH, Pco 2 , and base deficit in these animals, however, were normal. Since we were interested in the relationship of hematocrit to fetal oxygenation in normally grown as well as in undergrown fetuses, we studied these two animals, as follows. We measured umbilical blood flow and fetal oxygen delivery, oxygen extraction, and oxygen consumption at the initial high hematocrits and at several other hematocrits ranging down to 20%. Fetal hematocrit was reduced by isovolemic exchange transfusions with plasma obtained from a donor fetus (see Methods). As hematocrit was decreased toward normal levels in these growth-retarded animals, arterial and umbilical venous blood Po2 values, umbilical blood flow, fetal oxygen delivery, and fetal oxygen consumption all increased. The observations from these two growthretarded fetuses, when combined with the observations we made in normal fetuses, strongly suggest that there is an optimal hematocrit for oxygenation in the fetal lamb, irrespective whether the fetus is normally grown or undergrown in utero. In response to the question concerning whether the metabolic or hormonal environment of the fetus was affected by our changing the fetal hematocrit, it is likely that the in utero environment of the fetus was altered during our experiments. Nonetheless, the purpose of our studies was to assess the net effect of changes in fetal hematocrit on the intact fetus. To eliminate a potential systematic error due to the direction in which the hematocrit was altered, we increased hematocrit first in half of the fetuses and decreased hematocrit first in the remaining fetuses after determining that there were no differences in the response of fetal oxygenation to alterations in hematocrit relating to the order in which hematocrits were changed (see Methods). Studies were begun only after the fetus was found to be in a stable condition as exemplified by a constant arterial blood pressure, heart rate, and arterial blood gases and pH. In these studies, the mother was unanesthetized and breathing room air. Concerning the response of hepatic blood flow to changes in fetal hematocrit, we have data relating hepatic blood flow and oxygen delivery to fetal hematocrit. In fetal life, hepatic blood flow is derived from three sources: hepatic arterial blood flow (5% of the total flow), portal venous blood flow (20% of the total), and umbilical blood flow (the remaining 75%). When we increased hematocrit from I 0% to about 40% to 45%, blood flow to the liver varied little. At hematocrits >40% to 45%, however, hepatic blood flow decreased substantially. At these high hematocrits, hepatic blood flow decreased for two reasons: (1) total umbilical blood flow decreased at hematocrits above 40% to 45% and (2) the fraction of umbilical blood flow shunted past the liver through the ductus venosus increased. Because of this response of hepatic perfusion to hematocrit, oxygen delivery to the fetal liver was maximal at a hematocrit of 32% and decreased when hematocrits were raised or lowered from that value. This optimal hematocrit for the fetal liver is similar to the optimal
Volume 151 Number 7
hematocrits that we have found for the following organs: gastrointestinal tract organs, spleen, kidneys, placenta, and carcass tissues (skeletal muscle, skin, and bones). It is generally accepted that the fetus functions in an aerobic environment, even though fetal arterial blood Po2 is normally low relative to that of the adult. Fetal oxygen consumption in several species, including the human, ranges from 7 to 10 ml of oxygen per minute per kilogram of body weight. This rate is approximately double that of the adult non pregnant or pregnant animal or human. Despite the greater oxygen consumption rate in the fetus, there is a considerable oxygen reserve, since oxygen delivery to the fetus exceeds oxygen demand by the fetus by about threefold. This oxygen reserve enables the fetus to compensate for variations in its oxygen delivery that may result, for example, from changes in fetal hematocrit or alterations in umbilical or uterine blood flow. Although several fetal responses to stress are similar to those observed in the human, it is quite risky to try to extrapolate results of studies obtained on the lamb fetus to those on the human fetus. In a few situations, extrapolation from experimental data to clinical situations may be warranted, but in many instances this extrapolation may not be valid because of species differences. For example, the basal hematocrit in the human fetus averages 45%, whereas the lamb fetus has a mean hematocrit of about 32%. In our experiments, oxygen consumption by the fetal lamb was stable over the range of hematocrits from about 16% to 48% but decreased when hematocrits were increased above approximately 48% or decreased below 16% (comparable data for the human fetus are not available). We could not identify the precise he-
Relationship of fetal hematocrit and oxygenation
851
matocrit associated with decreases in oxygen consumption, because curved lines described the relationship of oxygen consumption to fetal hematocrit. We generated these curves with analysis of covariance with which we could take into account individual animal responses. The equations for the curves selected were those having the highest regression coefficients associated with the least residual sums of squares. Tested statistically, these equations best described our data. The observation that oxygen consumption fell at hematocrits above about 48% is not based on data from a single animal. Rather, it is based on the observation that oxygen consumption in several fetuses decreased at relatively high hematocrits. Fetal oxygen consumption was maintained at a value of about 7 ml of oxygen per minute per kilogram at hematocrits between about 16% and 48% because of alterations in oxygen extraction. Oxygen extraction is that quantity of oxygen consumed by the fetus, expressed as a fraction of the total oxygen delivered to the fetus. Changes in oxygen extraction result from alterations in capillary diffusion variables such as the capillary-to-cell Po2 gradient or the capillary surface area available for oxygen exchange (capillary density). In fetal life, all organs, with the exception of the heart and the brain, compensate for reductions in oxygen delivery by altering oxygen extraction (in the brain and heart, perfusion increases as compensation, while oxygen extraction does not change substantially). It is not known whether these changes in oxygen extraction in the fetus are primarily due to adjustments in capillary density or to modifications of the capillary-to-cell Po 2 gradient. In the adult animal, adjustments in oxygen extraction occur mainly as a result of increases in capillary density.