- Email: [email protected]
The effect of regular maternal exercise on erythropoietin in cord blood and amniotic fluid
Volume 172, Number 5 Am J Obstet Gynecol
10. 11. 12.
13.
14.
15. 16.
during pregnancy: prevalence and correlates. Pediatrics 1988;82:888-95. Chavez GF, Mulinare J, Cordero JF. Maternal cocaine use during pregnancy as a risk factor for congenital urogenital anomalies. JAMA 1989;262:795-8. Hoyme HE, Jones KL, Dixon SD, et al. Prenatal cocaine exposure and. fetal vascular disruption. Pediatrics 1990; 85:743-7. Chasnoff IJ, Landress HJ, Barrett ME. The prevalence of illicit-drug or alcohol use during pregnancy and discrepancies in mandatory reporting in Pinellas County, Florida. N EnglJ Med 1990;322:1202-6. Hoffman RS, Henry GC, Howland MA, Weisman RS, Weil L, Goldfrank LR. Association between life threatening cocaine toxicity and plasma cholinesterase activity. Ann Emerg Med 1992;21:247-53. Hoffman RS, Henry GC, Wax PM, Weisman RD, Howland MA, Goldfrank LR. Decreased plasma cholinesterase activity enhances cocaine toxicity in mice. J Pharmacol Exp Ther 1992;263:698-702. Roe DA, Little BB, Bawdon RE, Gilstrap LC. Metabolism of cocaine by human placentae: implications for fetal exposure. AM J OBSTETGYNECOL1990;163:713-8. Whittaker M. Cholinesterase. In: Beckman L, ed. Monographs in human genetics, New York: Karger, 1986:45-63.
Clapp et al.
17. Foerester EH, Dempsey J, Gariott JC. A gas chromatographic screening procedure for acid and neutral drugs in blood. J Anal Toxicol 1979;3:87-91. 18. DeVane CL, Simpkins JW, Miller RL, Braun SB. Tissue distribution of cocaine in the pregnant rat. Life Sci 1989; 45:1271-6. 19. Shah NS, May DA, Yates JD. Disposition of levo-[3H] cocaine in pregnant and nonpregnant mice. Toxicol Appl Pharmacol 1980;53:279-84. 20. Simone C, Derewlany LO, Oskamp M, Knie B, Koren G. Transfer of cocaine and benzoylecgonine across the perfused human placental cotyledon. AM J OBSTETGYNECOL 1994;170:1404-10. 21. Cunningham FG, MacDonald PC, Gant NF. Williams' obstetrics. 18th ed. Norwalk, Connecticut: Appleton & Lange, 1989:755. 22. Mule SJ, Casella GA, Misra AL. Intracellular disposition of [3HI-cocaine, [3H]-norcocaine, [3H]-benzoylecgonine and [3H-benzoylnorecogonine in the brain of rats. Life Sci 1976;19:1585-96. 23. Stettler RW, Bohman VR, Little BB, Westfall KL, Sobhi S: Metabolism of cocaine by cholinesterase during pregnancy: maternal and fetal activity. J Mat Fet Med [In press].
The effect of regular maternal exercise on erythropoietin in cord blood and amniotic fluid James F. Clapp III, MD, a Kathleen D. Little, PhD, a Sarah K. Appleby-Wineberg, BS, a and J o h n A. Widness, MD b Cleveland, Ohio, and Iowa City, Iowa OBJECTIVE: Our purpose was to test the hypothesis that continuing regular, high-intensity exercise until the onset of labor produces significant fetal hypoxemia, as evidenced by elevated erythropoietin levels in the fetal compartment. STUDY DESl6N: Erythropoietin levels were measured in samples of amniotic fluid and cord blood obtained from fetuses born to 31 exercising women and 29 matched controls. RESULTS" Erythropoietin levels (mean _+ SEM) in amniotic fluid obtained at the time of membrane rupture (9 -+ 2 vs 11 - 2 mU/ml) and in cord blood (38 -+ 6 vs 53 -+ 16 mU/ml) and amniotic fluid at delivery (9 +- 1 vs 24 -+ 12 mU/ml) were no different in women who exercised regularly until the onset of labor. In both groups the majority of elevated cord blood levels (> 50 mU/ml) could be explained by labor events. Amniotic fluid erythropoietin levels correlated directly (r = 0.52) with cord blood hematocrit and increased slowly during labor. CONCLUSION: We conclude that the initial hypothesis is incorrect and speculate that cord blood erythropoietin reflects fetal oxygenation during labor, whereas amniotic fluid erythropoietin primarily reflects the adequacy of oxygenation before the onset of labor. (AMJ OBSTETGYNECOL1995;172:1445-51 .) K e y w o r d s : Pregnancy, exercise, erythropoietin, fetus, amniotic fluid
From the Departments of Reproductive Biology and Obstetrics and Gynecology, Case Western Reserve University and MetroHealth Medical Center,° and the Department of Pediatrics, University of Iowa College of Medicine.b Supported in part by National Institutes of Health grant No. HD21268, The Iowa Children's Network Telethon, and MetroHealth Medical Center.
Received for publication June 28, 1994; revised November 2, 1994; accepted November 7, 1994. Reprint requests:James F. Clapp HI, MD, Department of Obstetrics and Gynecology, MetroHealth Medical Center, 2500 MetroHealth Dr., Cleveland, OH 44109. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $3.00 + 0 6/1/61856 1445
1446
Clapp otal.
Although studies in animal models suggest that maternal exercise in late pregnancy produces transient fetal hypoxemia, 1' ~ there is no evidence at delivery to suggest that regular, sustained exercise in late pregnancy produces recurrent fetal hypoxemia in the human.2. 3 Nonetheless, when untrained pregnant women perform short-term, self-limited, high-intensity exercise in late pregnancy there is a significant incidence of transient fetal bradycardia, suggesting that even short p e r i o d s of high-intensity exercise may produce a marked decrease in placental perfusion and fetal oxygenation. 46 In contrast, when fit pregnant women regularly engage in sustained aerobic exercise at intensities between 40% and 88% of Vo2max, the fetal heart rate (FHR) consistently increases during exercise. The magnitude of this increase is directly related to the intensity and duration of the exercise and the muscle mass used. 7 Thes e relationships suggest that this F H R response represents a graded fetal sympathetic response to a minor but progressive fall in Po 2 that is the result of a small but progressive decrease in placental perfusion. The critical outcome question is does this F H R response reflect a normal adaptive response with no adverse effect or does it reflect a physiologically significant decrease in tissue oxygen availability that recurs each time a pregnant woman performs her exercise routine? If the latter is the case, then primary biochemical and endocrine markers of decreased tissue oxygen availability, such as erythropoietin, should be increased in the fetal compartment. Erythropoietin production is sensitive to changes in tissue oxygenation, and circulating levels rise and fall rapidly in response to hypoxemia and hyperoxia, s' 9 In fetal life it is synthesized primarily by the liver/° and hypoxemia initiates synthesis, which produces a rise in circulating erythropoietin within 90 to 120 minutes? ~ It does not cross the placenta, and its half-life in the fetal circulation is < 2 hours. It is partially cleared by the kidney into the amniotic fluid c o m p a r t m e n t w h e r e it should have a prolonged half-life, because concentrations of peptides of similar molecular weight remain elevated for a week or more. T M Fetal erythropoietin levels are elevated in both cord blood and amniotic fluid in situations where there is clinical evidence of an intermittent or long-term reduction in fetal oxygen delivery? ~-17 Likewise, cord blood levels are significantly higher after normal labor and increase further in cases in which abnormal F H R patterns are noted. ~7-19 These observations indicate that levels of erythropoietin in the fetal c o m p a r t m e n t rise in response to recurrent mild to moderate decreases in fetal Po 2 in as little as 3 to 4 hours and that elevated amniotic fluid levels reflect recurrent or persistent episodes of fetal hypox-
May 1995 AmJ Obstet Gynecol
emia that antedate labor. Accordingly, the current study was designed to obtain erythropoietin levels in the cord and amniotic fluid of the fetuses of exercising women to test the hypothesis that regular, sustained maternal exercise in late pregnancy produces recurrent mild to moderate fetal hypoxemia.
Material and methods Subjects for this protocol were recruited from the participants in a continuing study of the interaction between pregnancy and exercise. A detailed description of their characteristics may be found in earlier publications?' 7 A prospective case-control design was used. Thirty-eight healthy nonsmoking women who maintained a regular exercise regimen throughout an uncomplicated singleton pregnancy were recruited at the time of the 36 to 38 week evaluation. To achieve the study's objectives, two additional criteria were required for retention. First, an exercise associated increase in FHR ( > 5 beats/min) after regular exercise regimen had to be documented. Second, the women had to continue to exercise three or more times a week for > 20 minutes at an intensity > 5 0 % of Vo2max until the onset of labor, with the last exercise session occurring within 48 hours of delivery. Thirty-one of the 38 subjects met these retention criteria. A similar number of healthy nonsmoking women with uncomplicated singleton pregnancies who discontinued regular sustained exercise by 28 weeks' gestation were recruited to serve as the control group. Demographic data were obtained from a questionnaire that was administered at the time of enrollment in the primary study. Exercise performance was monitored prospectively in all subjects by use of a portable heart rate monitor and exercise log, and the F H R response to exercise was assessed by Doppler imaging at each subject's final evaluation. 7 Exercise intensity was determined by measuring the oxygen c o n s u m p t i o n - h e a r t rate relationship during exercise in the laboratory, and correlating heart rates were obtained in the field with the laboratory data. The course of labor and delivery was monitored by a m e m b e r of the study team who was successful in obtaining one or more samples of amniotic fluid (at membrane rupture or delivery) and a sample of umbilical venous blood immediately after cord clamping in 60 of the 62 subjects (31 continued exercise and 29 control). In 33 of the 60 cases matched amniotic fluid samples were obtained at m e m b r a n e rupture and again at delivery, which occurred between 20 minutes and 31 hours later. All samples were spun, and the supernatant and serum were stored at - 70 ° C until analysis for erythro, poietin concentration. In addition, at 39 of the deliveries (17 continued exercise, 22 control) a sample of umbilical arterial blood was obtained immediately after
Clapp et al.
Volume 172, Number 5 Am J Obstet Gynecol
1447
180
E D 150 E
0 0 0
0
--J 120 i,i _.J
§ ft.. o
90
0
o oo
0
3o 0 CONTROL EXERCISE CONTROL EXERCISE CORD B L O O D
A M N I O T I CFLUID
Fig. 1. Individual erythropoietin levels and their distribution by group in cord blood serum and amniotic fluid. Each case is represented by single open circle in appropriate group for both cord blood and amniotic fluid.
the cord was clamped for determination of pH, hematocrit, plasma protein, and total solids. Arterial cord blood gasses were not obtained because the equipment available in most delivery suites in this area (Corometrics Medical Systems 220 p H machine, Wallingford, Conn.) does not have Po 2 or Pco 2 capability. Erythropoietin levels were batched by subject and run in duplicate by means of a specific double-antibody radioimmunoassay with a sensitivity of 1.2 mU/ml and m a x i m u m intraassay and interassay coefficients of variation over the assay range of 4% and 6%, respectively. 19 p H was measured with one of seven different blood gas analyzers. Each was calibrated with a known standard immediately before use. Hematocrit was determined in duplicate by means of the standard microcapillary technique. Total solids and protein content were determined by refractometry. Between-group differences were detected with an u n p a i r e d t test. Changes in amniotic fluid erythropoietin over time were sought with the X2 statistic. Leastsquares regression was used to detect significant relationships between variables. The distribution of the erythropoietin levels was skewed (see Fig. 1). Therefore they were normalized by natural logarithmic transformation before analysis. Significance was Set at the 0.05 level. On the basis of the control data available in the literature/5-~9 elevated levels of erythropoietin were defined as > 50 mU/ml for cord blood serum and > 30 mU/ml for amniotic fluid. On the basis of our experience with sampling variability at the low levels present in amniotic fluid; a significant change over time in amniotic fluid level was set at -> 3 mU/ml. Data are presented as the m e a n -+ SEM.
Results
Subject characteristics. The women were 31 +- l years old with parity ranging from 0 to 2 before the index pregnancy. Preconceptional weight averaged 59 + 1 kg, and 88% of the subjects worked outside the home throughout pregnancy. All subjects were middle or u p p e r socioeconomic class (on the basis of family income in the u p p e r two quartiles and 14 to 22 years of education). There were no between-group differences in these parameters. Significant between-group differences included an increased preconceptional ~'o2max (50.3 -+ 1.4 vs 45.6 _+ 1.6 ml/kg/min) and a decrease in gestational length (276 + 1 vs 280-+ 1 days), birth weight (3320 +- 67 vs 3570 -+ 94 gm), and pregnancy weight gain (12.9 -+ 0.9 vs 17.1 +_ 0.9 kg) in the continued exercise group. These between-group differences have been consistently present in all earlier reports from this population? T h e fact that the betweengroup difference in maximal capacity was small (< 10%) coupled with the fact that subjects were all near term and with birth weights >2600 gm suggests that these between-group differences should not act as confounders for the major outcome variable? The exercise logs indicated that the women who continued to exercise were performing at intensities that ranged between 50% and 88% of their Vo2max between three and six times a week for between 20 and 60 minutes a session. The types of exercise were diverse and included running, a variety of aerobics regimens, stair stepping, biking, and swimming. Erythropoietin levels. The raw data on erythropoietin levels in cord blood and amniotic fluid are detailed in Fig. 1 by group and by individual. In the 33 cases
May 1995 AmJ Obstet Gynecol
1448 Clapp et al.
55 E
50
,
~0
45 40 35 30 z
25
o 13_ o rv' "1"
Ok'~{~7
0
I I
I
I
I
I
I
I
I
4
8
12
16
20
24
28
32
TIME (hours)
Fig. 2. Changes in amniotic fluid erythropoietin levels over time during labor. Two data points (membrane rupture and delivery) for each of 33 subjects are separated by appropriate time interval and connected to solid line.
with two amniotic fluid levels the average of the two values has been used to avoid statistical weighting. T h e r e were slightly more control cases with elevated levels in both cord blood (8/29 vs 5/31) and amniotic fluid (4/29 vs 0/31). However, overall the erythropoietin levels in cord blood serum (Continued e x e r c i s e Mean = 38 -+ 6 mU/ml, median 27 mU/ml; Control Mean = 53 - 16 mU/ml, median 32 mU/ml) and amniotic fluid (Continued exercise - M e a n = 9 + 1 mU/ml, median 7 mU/ml; Control - Mean = 12 -+ 2 mU/ml, median 9 mU/ml) were not significantly different between groups (p > 0.65 and 0.20, respectively). However, there was a significant correlation (r = 0.42) between individual cord blood and amniotic fluid erythropoietin levels (p < 0.001). In 13 of the 60 cases the erythropoietin level in cord blood was elevated ( > 50 mU/ml). However, this was accompanied by elevated levels in amniotic fluid ( > 30 mU/ml) in only four instances, each of which was clearly related to clinical events that a p p e a r e d (on the basis of findings at delivery) to have persisted for some time. In the first instance the subject had uterine cramping associated with the intermittent passage of dark blood from the vagina for 15 days before labor, and during labor the fetus had baseline tachycardia with deceleration patterns a p p e a r i n g at 8 cm of dilatation. This led to amniotomy, which revealed meconium staining, followed by a vaginal operative delivery of a term female infant with an arterial cord blood p H of 7.18, A p g a r scores of 5 at 1 minute and 8 at 5 minutes, and meconium below the cords. T h e placenta was meconium stained and weighed 380 gm, and 30% of the maternal surface was covered by old clot with one area of infarction.
T h e second subject had a history oi~p o o r fetal growth (confirmed by uhrasonography) with 01igohydramnios and multiple leiomyomas. A nonreactive stress test followed by a positive oxytocin challenge test in the thirty-eighth week led to delivery by cesarean section of a 2109 gm female infant (lst percentile) with evidence of cord entanglement, a cord blood p H of 7.15, and A p g a r scores of 9 at 1 and 5 minutes. T h e placenta was unremarkable in appearance and weighed 380 gm. T h e third subject had a clinically uneventful pregnancy, but periodic severe decelerations of the F H R were noted in early labor. T h e initial scalp blood p H of 7.20 fell to 7.17 30 minutes later when fresh meconium was noted. An immediate cesarean section was performed, with delivery of a 3300 gm female infant with A p g a r scores of 6 at 1 minute and 9 at 5 minutes. The placenta weighed 430 gm and was grossly unremarkable. T h e only potential abnormality noted was a 58 cm umbilical cord, but there was no visual evidence of entanglement at delivery. T h e fourth subject had a positive group B streptococci culture followed by premature rupture of the membranes in the forty-first week and induction of labor. Labor was terminated after a 14-hour latent phase and a 17-hour active phase with 6 hours of arrest at 6 to 7 cm. A 3750 gm male infant with A p g a r scores of 8 and 9 at 1 and 5 minutes was delivered by cesarean section. T h e F H R pattern was normal throughout labor. Cord blood gases were not obtained, but both m o t h e r and child were culture positive for group B streptococci and had a clinically septic course. In the remaining nine cases the course of labor was clinically normal in three and protracted in six with arrest or FHR abnormalities leading to operative deliv-
Volume 172, Number 5 Am J Obstet Gyaaecol
Clapp et al. 1449
d
w
4
J
r = 0.52
Z
0 0
O0
3
5
D_
0
o n," "1-
0
0 0 000 ~
0
2
Z
O ~ (30
0 o
n.W
o " d < n.,
n
O0
1
O0 0
38
n =
39
I
I
I
I
l
I
I
I
I
I
40
42
44
46
48
50
52
54
56
58
60
UMBILICAL ARTERY HEMATOCRIT (%)
Fig. 3. Relationship between amniotic fluid erythropoietin levels and cord blood hematocrit after logarithmic transformation.
ery in four of the six. T h e amniotic fluid erythropoietin concentration was never elevated without a concomitant elevation in cord serum. Fig. 2 illustrates the relationship between time and an increase in amniotic fluid erythropoietin concentration during labor in the 33 cases in which duplicate samples were available. Note that the only instance in which the concentration rose to abnormal levels during labor was in the case in which labor was allowed to continue for 31 hours. Also note that there is a general tendency for levels to increase when the time between samples exceeds 6 hours. T h e level of amniotic fluid rose significantly ( > 3 mU/ml) in only one of 13 cases when the time between samples was < 6 hours, but it rose significantly in 12 and fell significantly in three of 20 cases in which there were >-6 hours between samples (p < 0.001).
FHR response, exercise performance, and erythropoietin levels. T h e magnitude of the increase in F H R with maternal exercise ranged between 8 and 34 beats/min after 20 to 60 minutes of continuous exercise that was p e r f o r m e d at intensities ranging between 52% and 80% of Vo2max. No significant relationship could be detected between the FHR response to maternal exercise and erythropoietin levels in either amniotic fluid (r = 0.0163) or cord blood (r = 0.0233). Likewise, no significant relationship could be detected between weekly exercise volume (the product of average exercise intensity and the weekly duration of exercise) and erythropoietin levels in the fetal c o m p a r t m e n t (r < 0.020).
Hematocrit, pH, and protein concentrations in cord blood. Significant between-group differences were present in the hematocrit and p H in the 39 cases (17
continued exercise, 22 control) in which arterial samples were obtained. The hematocrit was significantly lower (46% +-2% vs 51%-+ 1%) and the p H significantly higher (7.32 -+ 0.02 vs 7.23 + 0.02) in the continued exercise group. There were no significant betweengroup differences in protein content (5.8 +-0.1 vs 5.7 + 0.1 gm/dl) or total solids (129 -+ 2 vs 127 -+ 2). There were no significant relationships detected between cord blood erythropoietin levels and p H or hematocrit (p > 0.1). However, as illustrated in Fig. 3, amniotic fluid levels correlated directly with cord blood hematocrit (r = 0.52, p < 0.006). In addition, there was a weak inverse correlation between amniotic fluid erythropoietin and cord blood p H (r = 0.36,p < 0.05). Comment
The principal finding of this study is that continuing regular, sustained, strenuous maternal exercise throughout late pregnancy is not associated with an increase in fetal erythropoietin levels at labor and delivery in the human. Because the levels of erythropoietin in fetal blood and amniotic fluid are sensitive markers of recurrent or persistent mild to moderate decreases in fetal Po2,l~-19it is likely that the exercise-associated changes in FHR observed in fit women represents a normal reflex response to a physiologically insignificant decrease in fetal Po2, rather than tissue hypoxemia. This interpretation is also supported by the lack of a correlation between erythropoietin levels and either the magnitude of the F H R response or the overall volume of exercise. Thus these data do not support the initial hypothesis that continuing sustained exercise produces fetal hypoxemia in late pregnancy. Indeed, the trends noted in the raw data, coupled with the lower hematocrit and the correla-
May 1995 Am J Obstet Gynecol
1450 Clapp etal.
tion between cord blood hematocrit and amniotic fluid erythropoietin levels, suggest that the fetuses in the continued exercise group may actually have had slightly better tissue oxygenation than those in the control group. This should be reassuring to physically active p r e g n a n t women and those who care for them. Likewise, the lack of significant increases in erythropoietin levels in the fetal c o m p a r t m e n t in the continued exercise group indicates that regular exercise during pregnancy probably induces a variety of compensatory mechanisms that blunt the magnitude of any exercise-induced decrease in placental perfusion. T h r e e aspects of the current data suggest that determining erythropoietin levels in both cord blood and amniotic fluid may be of clinical value in determining the presence or absence, and the timing of, significant fetal hypoxemia before or during labor. First, in the four cases where both were elevated there was clear clinical evidence to suggest that a risk of fetal hypoxemia had been present for some time, confirming the described association between elevated levels in the two compartments that is present before labor. 15 Second, the presence of isolated elevations in cord blood under clinical circumstances where additional abnormalities a p p e a r e d during labor probably reflects the significant lag time between elevation in the circulation and a subsequent elevation in the amniotic fluid. Finally, the absence of isolated elevations of erythropoietin in amniotic fluid in clinically normal pregnancies in this and in one other study,15 coupled with its slow response time during labor and relatively weak correlation with cord blood levels after labor, suggest that it may be a very good marker of isolated or recurrent episodes of hypoxia that antedate labor by a significant period of time. T h e relationship between amniotic fluid erythropoietin levels and cord blood hematocrit, a reasonable long-term marker of earlier erythropoiesis, clearly supports this interpretation. Thus normal levels in both compartments at delivery would be consistent with the absence of fetal hypoxemia in late pregnancy and during labor. Isolated elevation of levels in the amniotic fluid would indicate significant hypoxemic stress a day or more before the onset of labor, whereas elevations in cord blood only would reflect the events during labor and elevation in both compartments would signify that fetal hypoxemia antedated labor and persisted during labor. Finally, it should be recognized that this study and the interpretation of its data are limited by the fact that sampling is restricted to one or, in the case of amniotic fluid, two time points during labor and delivery. It also is limited by the fact that detailed pharmacokinetic data for the rate of appearance and turnover of erythropoietin in amniotic fluid are not available. These issues must be addressed either in animal models or in the
h u m a n by amniocentesis coupled with cordocentesis before labor and r e p e a t e d scalp blood sampling in the course of labor. TM 20-22 This project would not have been possible without the technical expertise of Robert L. Schmidt and the help and cooperation of numerous obstetricians and midwives throughout northeast Ohio. REFERENCES
1. Lotgering FK, Gilbert RD, Longo LD. Maternal and fetal responses to exercise during pregnancy. Physiol Rev 1985; 65:1-36. 2. Clapp, JF, Rokey R, Treadway JI, Carpenter MW, Artal RM, Warrnes C. Exercise in pregnancy. Med Sci Sports Exerc 1992;24(suppl):S294-300. 3. Clapp JF. A clinical approach to exercise in pregnancy. Clin Sports Med 1994;13:443-58. 4. Carpenter MW, Sady SP, Hoegsberg B, Sady MA, Haydon B, Coustan DR. Fetal heart rate response to maternal exertion. JAMA 1988;259:3006-9. 5. Artal RM, Posner MD. Fetal responses to maternal exercise. In: Mittlemark RA, Wistwell RA, Drinkwater BL, eds. Exercise in pregnancy, 2nd ed. Baltimore: Williams and Wilkins, 1991:213-24. 6. Clapp JF. The effects of exercise on uterine blood flow. In: Rosenfeld CR, ed. The uterine circulation. Ithaca: Perinatology Press, 1989:289-310. 7. Clapp JF, Little KD, Capeless EL. The fetal heart rate response to various intensities of recreational exercise in mid and late pregnancy. AMJ OBsrEx GYNECOL1993;168: 198-206. 8. Krantz SB. Erythropoietin. Blood 1991;77:419-34. 9. Finne PH, Halvorsen S. Regulation of erythropoiesis in the fetus and newborn. Arch Dis Child 1972;47:683-8. 1O. Zanjani ED, Poster J, Burlington H. Liver as the primary site of erythropoietin formation in the fetus. J Lab Clin Med 1977;89:640-4. 11. Widness JA, Clemons GK, Teramo KA, Coustan DR, Schwartz R. Temporal relationship of fetal plasma erythropoietin to acute hypoxemia in the sheep and human. In: Stern L, Oh W, eds. Physiological foundations of perinatal care. New York: Elsevier, 1987:30-43. 12. Widness JA, Veng-Pederson P, Modi N, Schmidt RL, Chestnut DH. Developmental changes in erythropoietin kinetics in fetal and neonatal sheep. Pediatr Res 1990;28: 284. 13. Widness JA, Sawyer ST, Schmidt RL, Chestnut DH. Lack of maternal to fetal transfer of 125-1abelled-erythropoietin in sheep. J Dev Physiol 1991;15:139-43. 14. GitlinJD, Gitlin D. Protein binding by cell membranes and selective transfer of proteins from mother to young across tissue barriers. In: Hemmings WA, ed. Maternal foetal transmission of immunoglobulins. London: Cambridge University Press, 1974:113-21. 15. Teramo KA, Widness JA, Clemons GK, Voutilainen PEJ, McKinlay S, Schwartz R. Amniotic fluid erythropoietin correlates with umbilical plasma erythropoietin in normal and abnormal pregnancy. Obstet Gynecol 1987;69:710-6. 16. Rollins MD, Maxwell AP, Afrasiabi M, Halliday HL, Lappin TRJ. Cord blood erythropoietin, pH, PaO 2 and haematocrit following cesarian section before labour. Biol Neonate 1993;63:147-52. 17. Maier RF, Bohme K, Dudenhausen JW, Obladen M. Cord blood erythropoietin in relation to different markers of fetal hypoxia. Obstet Gynecol 1993;81:575-80. 18. Widness JA, Clemons GK, Garcia JF, Oh W, Schwartz R. Increased immunoreactive erythropoietin in cord serum after labor. Au J OBSTETGYNECOL1984;148:194-7. 19. Widness JA, Teramo KA, Clemons GK, et al. Correlation of the interpretation of fetal heart rate records with cord
Volume 172, Number 5 Am J Obstet Gynecol
plasma erythropoietin levels. Br J Obstet Gynaecol 1985; 92:326-32. 20. Georgieff MK, Landon MB, Mills MM, et al. Abnormal iron distribution in infants of diabetic mothers: spectrum and maternal antecedents. J Pediatr 1990;117:455-61. 21. Moya FR, Grannum PAT, Widness JA, Clemons GK, Copel JA, Hobbins JC. Erythropoietin in human fetuses with
Gagnon et al.
immune hemolytic anemia and hydrops fetalis. Obstet Gynecol 1993;82:353-8. 22. Salvesen DR, Brudenell JM, Snijders RJM, Ireland RM, Nicolaides KH. Fetal plasma erythropoietin in pregnancies complicated by maternal diabetes mellitus. AMJ OBSTET GYNECOL1993 ; 168:88-94.
Alterations in fetal and placental deoxyribonucleic acid synthesis rates after chronic fetal placental embolization R o b e r t G a g n o n , M D , a' b H o w a r d R u n d l e , B S c , c' d, ~ L a u r a J o h n s t o n , A C T , " b a n d V i c t o r K . M . H a r t , M D c' d, e
London, Ontario, Canada OBJECTIVE: Fetal growth and development are closely related to normal placental growth and function. We performed a study to determine the effect of a 10-day period of fetal hypoxemia induced by umbilical-placental hypoperfusion on tissue deoxyribonucleic acid synthesis rates in the 0.84 to 0.91 of gestation ovine fetus and placenta. STUDY DESIGN: Daily fetal placental embolization was performed in four chronically catheterized sheep fetuses until fetal arterial oxygen content decreased by - 3 0 % compared with preembolization values. Five control fetuses received vehicle only. On experimental day 10, the deoxyribonucleic acid synthesis rate was determined by injecting tritiated thymidine (1 mCi/kg) intravenously approximately 8 hours before the end of the study. RESULTS: Fetal arterial oxygen decreased from 3.2 - 0.1 (SEM) mmol/L preembolization to 2.2 -+ 0.2 mmol/L on day 10 (p < 0.001) and remained unchanged in controls. On day 10 deoxyribonucleic acid synthesis rates were significantly reduced in embolized fetuses compared with controls, by 38% in cotyledons (83.0 -+ 15.1 vs t33.7 + 9.9 disintegrations/min/~g deoxyribonucleic acid, p < 0.05), 28% in the left ventricular wall (36.8 - 3.7 vs 51.0 -+ 4.7 disintegrations/min/#g deoxyribonucleic acid, p < 0.05), and 45% in the quadriceps muscle (15.4 _+ 4.0 vs 28.1 _+ 3.0 disintegrations/min/l~g deoxyribonucleic acid, p < 0.05). Tritiated thymidine autoradiography demonstrated that cotyledonary deoxyribonucleic acid synthesis occurred exclusively in the fetal trophoblasts cells. CONCLUSION: We concluded that a reduction in cotyledonary, quadriceps muscle, and left ventricular myocardium deoxyribonucleic acid synthesis rates are the earliest adaptive mechanisms of fetal growth associated with development of umbilical-placental insufficiency. We speculate that alteration in the myocardial deoxyribonucleic acid synthesis rate could be a major contributing factor in the deterioration of fetal myocardial function associated with increased placental vascular resistance. (AM J OBSTETGYNECOL 1995;172:1451-8.)
Key words: Placental insufficiency, fetal hypoxia, fetal growth
From the Departments of Obstetrics and Gynaecolog'y,° Physiology,b Paediatrics,c Biochemistr3d and Anatomy/St. Joseph's Health Centre, Lawson Research Institute, Medical Research Council Group in Fetal and Neonatal Health and Development, University of Western Ontario. Supported by Medical Research Council of Canada and Physicians' Services Incorporated Foundation. Received for publication July 12, 1994; revised October 28, 1994; accepted November 7, 1994. Reprint requests: Robert Gagnon, MD, Department of Obstetrics and Gynaecology, St.Joseph's Health Centre, 268 Grosvenor St., London, Ontario, Canada N6A 4V2. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $3.00 + 0 6/1/61851
Fetal growth and development are closely related to normal placental growth and function. It has been shown that fetal placental embolization by latex microspheres is associated with progressive fetal hypoxemia and changes in umbilical artery Doppler flow velocity waveforms suggestive of an increase in placental vascular resistance similar to that observed in pregnancies complicated with intrauterine growth retardation (IUGR). 1.2 It is not clear yet how and to what extent fetal and placental growth would be altered during the development of umbilicalplacental hypoperfusion.l' 3 Changes in cellular growth 1451
10. 11. 12.
13.
14.
15. 16.
during pregnancy: prevalence and correlates. Pediatrics 1988;82:888-95. Chavez GF, Mulinare J, Cordero JF. Maternal cocaine use during pregnancy as a risk factor for congenital urogenital anomalies. JAMA 1989;262:795-8. Hoyme HE, Jones KL, Dixon SD, et al. Prenatal cocaine exposure and. fetal vascular disruption. Pediatrics 1990; 85:743-7. Chasnoff IJ, Landress HJ, Barrett ME. The prevalence of illicit-drug or alcohol use during pregnancy and discrepancies in mandatory reporting in Pinellas County, Florida. N EnglJ Med 1990;322:1202-6. Hoffman RS, Henry GC, Howland MA, Weisman RS, Weil L, Goldfrank LR. Association between life threatening cocaine toxicity and plasma cholinesterase activity. Ann Emerg Med 1992;21:247-53. Hoffman RS, Henry GC, Wax PM, Weisman RD, Howland MA, Goldfrank LR. Decreased plasma cholinesterase activity enhances cocaine toxicity in mice. J Pharmacol Exp Ther 1992;263:698-702. Roe DA, Little BB, Bawdon RE, Gilstrap LC. Metabolism of cocaine by human placentae: implications for fetal exposure. AM J OBSTETGYNECOL1990;163:713-8. Whittaker M. Cholinesterase. In: Beckman L, ed. Monographs in human genetics, New York: Karger, 1986:45-63.
Clapp et al.
17. Foerester EH, Dempsey J, Gariott JC. A gas chromatographic screening procedure for acid and neutral drugs in blood. J Anal Toxicol 1979;3:87-91. 18. DeVane CL, Simpkins JW, Miller RL, Braun SB. Tissue distribution of cocaine in the pregnant rat. Life Sci 1989; 45:1271-6. 19. Shah NS, May DA, Yates JD. Disposition of levo-[3H] cocaine in pregnant and nonpregnant mice. Toxicol Appl Pharmacol 1980;53:279-84. 20. Simone C, Derewlany LO, Oskamp M, Knie B, Koren G. Transfer of cocaine and benzoylecgonine across the perfused human placental cotyledon. AM J OBSTETGYNECOL 1994;170:1404-10. 21. Cunningham FG, MacDonald PC, Gant NF. Williams' obstetrics. 18th ed. Norwalk, Connecticut: Appleton & Lange, 1989:755. 22. Mule SJ, Casella GA, Misra AL. Intracellular disposition of [3HI-cocaine, [3H]-norcocaine, [3H]-benzoylecgonine and [3H-benzoylnorecogonine in the brain of rats. Life Sci 1976;19:1585-96. 23. Stettler RW, Bohman VR, Little BB, Westfall KL, Sobhi S: Metabolism of cocaine by cholinesterase during pregnancy: maternal and fetal activity. J Mat Fet Med [In press].
The effect of regular maternal exercise on erythropoietin in cord blood and amniotic fluid James F. Clapp III, MD, a Kathleen D. Little, PhD, a Sarah K. Appleby-Wineberg, BS, a and J o h n A. Widness, MD b Cleveland, Ohio, and Iowa City, Iowa OBJECTIVE: Our purpose was to test the hypothesis that continuing regular, high-intensity exercise until the onset of labor produces significant fetal hypoxemia, as evidenced by elevated erythropoietin levels in the fetal compartment. STUDY DESl6N: Erythropoietin levels were measured in samples of amniotic fluid and cord blood obtained from fetuses born to 31 exercising women and 29 matched controls. RESULTS" Erythropoietin levels (mean _+ SEM) in amniotic fluid obtained at the time of membrane rupture (9 -+ 2 vs 11 - 2 mU/ml) and in cord blood (38 -+ 6 vs 53 -+ 16 mU/ml) and amniotic fluid at delivery (9 +- 1 vs 24 -+ 12 mU/ml) were no different in women who exercised regularly until the onset of labor. In both groups the majority of elevated cord blood levels (> 50 mU/ml) could be explained by labor events. Amniotic fluid erythropoietin levels correlated directly (r = 0.52) with cord blood hematocrit and increased slowly during labor. CONCLUSION: We conclude that the initial hypothesis is incorrect and speculate that cord blood erythropoietin reflects fetal oxygenation during labor, whereas amniotic fluid erythropoietin primarily reflects the adequacy of oxygenation before the onset of labor. (AMJ OBSTETGYNECOL1995;172:1445-51 .) K e y w o r d s : Pregnancy, exercise, erythropoietin, fetus, amniotic fluid
From the Departments of Reproductive Biology and Obstetrics and Gynecology, Case Western Reserve University and MetroHealth Medical Center,° and the Department of Pediatrics, University of Iowa College of Medicine.b Supported in part by National Institutes of Health grant No. HD21268, The Iowa Children's Network Telethon, and MetroHealth Medical Center.
Received for publication June 28, 1994; revised November 2, 1994; accepted November 7, 1994. Reprint requests:James F. Clapp HI, MD, Department of Obstetrics and Gynecology, MetroHealth Medical Center, 2500 MetroHealth Dr., Cleveland, OH 44109. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $3.00 + 0 6/1/61856 1445
1446
Clapp otal.
Although studies in animal models suggest that maternal exercise in late pregnancy produces transient fetal hypoxemia, 1' ~ there is no evidence at delivery to suggest that regular, sustained exercise in late pregnancy produces recurrent fetal hypoxemia in the human.2. 3 Nonetheless, when untrained pregnant women perform short-term, self-limited, high-intensity exercise in late pregnancy there is a significant incidence of transient fetal bradycardia, suggesting that even short p e r i o d s of high-intensity exercise may produce a marked decrease in placental perfusion and fetal oxygenation. 46 In contrast, when fit pregnant women regularly engage in sustained aerobic exercise at intensities between 40% and 88% of Vo2max, the fetal heart rate (FHR) consistently increases during exercise. The magnitude of this increase is directly related to the intensity and duration of the exercise and the muscle mass used. 7 Thes e relationships suggest that this F H R response represents a graded fetal sympathetic response to a minor but progressive fall in Po 2 that is the result of a small but progressive decrease in placental perfusion. The critical outcome question is does this F H R response reflect a normal adaptive response with no adverse effect or does it reflect a physiologically significant decrease in tissue oxygen availability that recurs each time a pregnant woman performs her exercise routine? If the latter is the case, then primary biochemical and endocrine markers of decreased tissue oxygen availability, such as erythropoietin, should be increased in the fetal compartment. Erythropoietin production is sensitive to changes in tissue oxygenation, and circulating levels rise and fall rapidly in response to hypoxemia and hyperoxia, s' 9 In fetal life it is synthesized primarily by the liver/° and hypoxemia initiates synthesis, which produces a rise in circulating erythropoietin within 90 to 120 minutes? ~ It does not cross the placenta, and its half-life in the fetal circulation is < 2 hours. It is partially cleared by the kidney into the amniotic fluid c o m p a r t m e n t w h e r e it should have a prolonged half-life, because concentrations of peptides of similar molecular weight remain elevated for a week or more. T M Fetal erythropoietin levels are elevated in both cord blood and amniotic fluid in situations where there is clinical evidence of an intermittent or long-term reduction in fetal oxygen delivery? ~-17 Likewise, cord blood levels are significantly higher after normal labor and increase further in cases in which abnormal F H R patterns are noted. ~7-19 These observations indicate that levels of erythropoietin in the fetal c o m p a r t m e n t rise in response to recurrent mild to moderate decreases in fetal Po 2 in as little as 3 to 4 hours and that elevated amniotic fluid levels reflect recurrent or persistent episodes of fetal hypox-
May 1995 AmJ Obstet Gynecol
emia that antedate labor. Accordingly, the current study was designed to obtain erythropoietin levels in the cord and amniotic fluid of the fetuses of exercising women to test the hypothesis that regular, sustained maternal exercise in late pregnancy produces recurrent mild to moderate fetal hypoxemia.
Material and methods Subjects for this protocol were recruited from the participants in a continuing study of the interaction between pregnancy and exercise. A detailed description of their characteristics may be found in earlier publications?' 7 A prospective case-control design was used. Thirty-eight healthy nonsmoking women who maintained a regular exercise regimen throughout an uncomplicated singleton pregnancy were recruited at the time of the 36 to 38 week evaluation. To achieve the study's objectives, two additional criteria were required for retention. First, an exercise associated increase in FHR ( > 5 beats/min) after regular exercise regimen had to be documented. Second, the women had to continue to exercise three or more times a week for > 20 minutes at an intensity > 5 0 % of Vo2max until the onset of labor, with the last exercise session occurring within 48 hours of delivery. Thirty-one of the 38 subjects met these retention criteria. A similar number of healthy nonsmoking women with uncomplicated singleton pregnancies who discontinued regular sustained exercise by 28 weeks' gestation were recruited to serve as the control group. Demographic data were obtained from a questionnaire that was administered at the time of enrollment in the primary study. Exercise performance was monitored prospectively in all subjects by use of a portable heart rate monitor and exercise log, and the F H R response to exercise was assessed by Doppler imaging at each subject's final evaluation. 7 Exercise intensity was determined by measuring the oxygen c o n s u m p t i o n - h e a r t rate relationship during exercise in the laboratory, and correlating heart rates were obtained in the field with the laboratory data. The course of labor and delivery was monitored by a m e m b e r of the study team who was successful in obtaining one or more samples of amniotic fluid (at membrane rupture or delivery) and a sample of umbilical venous blood immediately after cord clamping in 60 of the 62 subjects (31 continued exercise and 29 control). In 33 of the 60 cases matched amniotic fluid samples were obtained at m e m b r a n e rupture and again at delivery, which occurred between 20 minutes and 31 hours later. All samples were spun, and the supernatant and serum were stored at - 70 ° C until analysis for erythro, poietin concentration. In addition, at 39 of the deliveries (17 continued exercise, 22 control) a sample of umbilical arterial blood was obtained immediately after
Clapp et al.
Volume 172, Number 5 Am J Obstet Gynecol
1447
180
E D 150 E
0 0 0
0
--J 120 i,i _.J
§ ft.. o
90
0
o oo
0
3o 0 CONTROL EXERCISE CONTROL EXERCISE CORD B L O O D
A M N I O T I CFLUID
Fig. 1. Individual erythropoietin levels and their distribution by group in cord blood serum and amniotic fluid. Each case is represented by single open circle in appropriate group for both cord blood and amniotic fluid.
the cord was clamped for determination of pH, hematocrit, plasma protein, and total solids. Arterial cord blood gasses were not obtained because the equipment available in most delivery suites in this area (Corometrics Medical Systems 220 p H machine, Wallingford, Conn.) does not have Po 2 or Pco 2 capability. Erythropoietin levels were batched by subject and run in duplicate by means of a specific double-antibody radioimmunoassay with a sensitivity of 1.2 mU/ml and m a x i m u m intraassay and interassay coefficients of variation over the assay range of 4% and 6%, respectively. 19 p H was measured with one of seven different blood gas analyzers. Each was calibrated with a known standard immediately before use. Hematocrit was determined in duplicate by means of the standard microcapillary technique. Total solids and protein content were determined by refractometry. Between-group differences were detected with an u n p a i r e d t test. Changes in amniotic fluid erythropoietin over time were sought with the X2 statistic. Leastsquares regression was used to detect significant relationships between variables. The distribution of the erythropoietin levels was skewed (see Fig. 1). Therefore they were normalized by natural logarithmic transformation before analysis. Significance was Set at the 0.05 level. On the basis of the control data available in the literature/5-~9 elevated levels of erythropoietin were defined as > 50 mU/ml for cord blood serum and > 30 mU/ml for amniotic fluid. On the basis of our experience with sampling variability at the low levels present in amniotic fluid; a significant change over time in amniotic fluid level was set at -> 3 mU/ml. Data are presented as the m e a n -+ SEM.
Results
Subject characteristics. The women were 31 +- l years old with parity ranging from 0 to 2 before the index pregnancy. Preconceptional weight averaged 59 + 1 kg, and 88% of the subjects worked outside the home throughout pregnancy. All subjects were middle or u p p e r socioeconomic class (on the basis of family income in the u p p e r two quartiles and 14 to 22 years of education). There were no between-group differences in these parameters. Significant between-group differences included an increased preconceptional ~'o2max (50.3 -+ 1.4 vs 45.6 _+ 1.6 ml/kg/min) and a decrease in gestational length (276 + 1 vs 280-+ 1 days), birth weight (3320 +- 67 vs 3570 -+ 94 gm), and pregnancy weight gain (12.9 -+ 0.9 vs 17.1 +_ 0.9 kg) in the continued exercise group. These between-group differences have been consistently present in all earlier reports from this population? T h e fact that the betweengroup difference in maximal capacity was small (< 10%) coupled with the fact that subjects were all near term and with birth weights >2600 gm suggests that these between-group differences should not act as confounders for the major outcome variable? The exercise logs indicated that the women who continued to exercise were performing at intensities that ranged between 50% and 88% of their Vo2max between three and six times a week for between 20 and 60 minutes a session. The types of exercise were diverse and included running, a variety of aerobics regimens, stair stepping, biking, and swimming. Erythropoietin levels. The raw data on erythropoietin levels in cord blood and amniotic fluid are detailed in Fig. 1 by group and by individual. In the 33 cases
May 1995 AmJ Obstet Gynecol
1448 Clapp et al.
55 E
50
,
~0
45 40 35 30 z
25
o 13_ o rv' "1"
Ok'~{~7
0
I I
I
I
I
I
I
I
I
4
8
12
16
20
24
28
32
TIME (hours)
Fig. 2. Changes in amniotic fluid erythropoietin levels over time during labor. Two data points (membrane rupture and delivery) for each of 33 subjects are separated by appropriate time interval and connected to solid line.
with two amniotic fluid levels the average of the two values has been used to avoid statistical weighting. T h e r e were slightly more control cases with elevated levels in both cord blood (8/29 vs 5/31) and amniotic fluid (4/29 vs 0/31). However, overall the erythropoietin levels in cord blood serum (Continued e x e r c i s e Mean = 38 -+ 6 mU/ml, median 27 mU/ml; Control Mean = 53 - 16 mU/ml, median 32 mU/ml) and amniotic fluid (Continued exercise - M e a n = 9 + 1 mU/ml, median 7 mU/ml; Control - Mean = 12 -+ 2 mU/ml, median 9 mU/ml) were not significantly different between groups (p > 0.65 and 0.20, respectively). However, there was a significant correlation (r = 0.42) between individual cord blood and amniotic fluid erythropoietin levels (p < 0.001). In 13 of the 60 cases the erythropoietin level in cord blood was elevated ( > 50 mU/ml). However, this was accompanied by elevated levels in amniotic fluid ( > 30 mU/ml) in only four instances, each of which was clearly related to clinical events that a p p e a r e d (on the basis of findings at delivery) to have persisted for some time. In the first instance the subject had uterine cramping associated with the intermittent passage of dark blood from the vagina for 15 days before labor, and during labor the fetus had baseline tachycardia with deceleration patterns a p p e a r i n g at 8 cm of dilatation. This led to amniotomy, which revealed meconium staining, followed by a vaginal operative delivery of a term female infant with an arterial cord blood p H of 7.18, A p g a r scores of 5 at 1 minute and 8 at 5 minutes, and meconium below the cords. T h e placenta was meconium stained and weighed 380 gm, and 30% of the maternal surface was covered by old clot with one area of infarction.
T h e second subject had a history oi~p o o r fetal growth (confirmed by uhrasonography) with 01igohydramnios and multiple leiomyomas. A nonreactive stress test followed by a positive oxytocin challenge test in the thirty-eighth week led to delivery by cesarean section of a 2109 gm female infant (lst percentile) with evidence of cord entanglement, a cord blood p H of 7.15, and A p g a r scores of 9 at 1 and 5 minutes. T h e placenta was unremarkable in appearance and weighed 380 gm. T h e third subject had a clinically uneventful pregnancy, but periodic severe decelerations of the F H R were noted in early labor. T h e initial scalp blood p H of 7.20 fell to 7.17 30 minutes later when fresh meconium was noted. An immediate cesarean section was performed, with delivery of a 3300 gm female infant with A p g a r scores of 6 at 1 minute and 9 at 5 minutes. The placenta weighed 430 gm and was grossly unremarkable. T h e only potential abnormality noted was a 58 cm umbilical cord, but there was no visual evidence of entanglement at delivery. T h e fourth subject had a positive group B streptococci culture followed by premature rupture of the membranes in the forty-first week and induction of labor. Labor was terminated after a 14-hour latent phase and a 17-hour active phase with 6 hours of arrest at 6 to 7 cm. A 3750 gm male infant with A p g a r scores of 8 and 9 at 1 and 5 minutes was delivered by cesarean section. T h e F H R pattern was normal throughout labor. Cord blood gases were not obtained, but both m o t h e r and child were culture positive for group B streptococci and had a clinically septic course. In the remaining nine cases the course of labor was clinically normal in three and protracted in six with arrest or FHR abnormalities leading to operative deliv-
Volume 172, Number 5 Am J Obstet Gyaaecol
Clapp et al. 1449
d
w
4
J
r = 0.52
Z
0 0
O0
3
5
D_
0
o n," "1-
0
0 0 000 ~
0
2
Z
O ~ (30
0 o
n.W
o " d < n.,
n
O0
1
O0 0
38
n =
39
I
I
I
I
l
I
I
I
I
I
40
42
44
46
48
50
52
54
56
58
60
UMBILICAL ARTERY HEMATOCRIT (%)
Fig. 3. Relationship between amniotic fluid erythropoietin levels and cord blood hematocrit after logarithmic transformation.
ery in four of the six. T h e amniotic fluid erythropoietin concentration was never elevated without a concomitant elevation in cord serum. Fig. 2 illustrates the relationship between time and an increase in amniotic fluid erythropoietin concentration during labor in the 33 cases in which duplicate samples were available. Note that the only instance in which the concentration rose to abnormal levels during labor was in the case in which labor was allowed to continue for 31 hours. Also note that there is a general tendency for levels to increase when the time between samples exceeds 6 hours. T h e level of amniotic fluid rose significantly ( > 3 mU/ml) in only one of 13 cases when the time between samples was < 6 hours, but it rose significantly in 12 and fell significantly in three of 20 cases in which there were >-6 hours between samples (p < 0.001).
FHR response, exercise performance, and erythropoietin levels. T h e magnitude of the increase in F H R with maternal exercise ranged between 8 and 34 beats/min after 20 to 60 minutes of continuous exercise that was p e r f o r m e d at intensities ranging between 52% and 80% of Vo2max. No significant relationship could be detected between the FHR response to maternal exercise and erythropoietin levels in either amniotic fluid (r = 0.0163) or cord blood (r = 0.0233). Likewise, no significant relationship could be detected between weekly exercise volume (the product of average exercise intensity and the weekly duration of exercise) and erythropoietin levels in the fetal c o m p a r t m e n t (r < 0.020).
Hematocrit, pH, and protein concentrations in cord blood. Significant between-group differences were present in the hematocrit and p H in the 39 cases (17
continued exercise, 22 control) in which arterial samples were obtained. The hematocrit was significantly lower (46% +-2% vs 51%-+ 1%) and the p H significantly higher (7.32 -+ 0.02 vs 7.23 + 0.02) in the continued exercise group. There were no significant betweengroup differences in protein content (5.8 +-0.1 vs 5.7 + 0.1 gm/dl) or total solids (129 -+ 2 vs 127 -+ 2). There were no significant relationships detected between cord blood erythropoietin levels and p H or hematocrit (p > 0.1). However, as illustrated in Fig. 3, amniotic fluid levels correlated directly with cord blood hematocrit (r = 0.52, p < 0.006). In addition, there was a weak inverse correlation between amniotic fluid erythropoietin and cord blood p H (r = 0.36,p < 0.05). Comment
The principal finding of this study is that continuing regular, sustained, strenuous maternal exercise throughout late pregnancy is not associated with an increase in fetal erythropoietin levels at labor and delivery in the human. Because the levels of erythropoietin in fetal blood and amniotic fluid are sensitive markers of recurrent or persistent mild to moderate decreases in fetal Po2,l~-19it is likely that the exercise-associated changes in FHR observed in fit women represents a normal reflex response to a physiologically insignificant decrease in fetal Po2, rather than tissue hypoxemia. This interpretation is also supported by the lack of a correlation between erythropoietin levels and either the magnitude of the F H R response or the overall volume of exercise. Thus these data do not support the initial hypothesis that continuing sustained exercise produces fetal hypoxemia in late pregnancy. Indeed, the trends noted in the raw data, coupled with the lower hematocrit and the correla-
May 1995 Am J Obstet Gynecol
1450 Clapp etal.
tion between cord blood hematocrit and amniotic fluid erythropoietin levels, suggest that the fetuses in the continued exercise group may actually have had slightly better tissue oxygenation than those in the control group. This should be reassuring to physically active p r e g n a n t women and those who care for them. Likewise, the lack of significant increases in erythropoietin levels in the fetal c o m p a r t m e n t in the continued exercise group indicates that regular exercise during pregnancy probably induces a variety of compensatory mechanisms that blunt the magnitude of any exercise-induced decrease in placental perfusion. T h r e e aspects of the current data suggest that determining erythropoietin levels in both cord blood and amniotic fluid may be of clinical value in determining the presence or absence, and the timing of, significant fetal hypoxemia before or during labor. First, in the four cases where both were elevated there was clear clinical evidence to suggest that a risk of fetal hypoxemia had been present for some time, confirming the described association between elevated levels in the two compartments that is present before labor. 15 Second, the presence of isolated elevations in cord blood under clinical circumstances where additional abnormalities a p p e a r e d during labor probably reflects the significant lag time between elevation in the circulation and a subsequent elevation in the amniotic fluid. Finally, the absence of isolated elevations of erythropoietin in amniotic fluid in clinically normal pregnancies in this and in one other study,15 coupled with its slow response time during labor and relatively weak correlation with cord blood levels after labor, suggest that it may be a very good marker of isolated or recurrent episodes of hypoxia that antedate labor by a significant period of time. T h e relationship between amniotic fluid erythropoietin levels and cord blood hematocrit, a reasonable long-term marker of earlier erythropoiesis, clearly supports this interpretation. Thus normal levels in both compartments at delivery would be consistent with the absence of fetal hypoxemia in late pregnancy and during labor. Isolated elevation of levels in the amniotic fluid would indicate significant hypoxemic stress a day or more before the onset of labor, whereas elevations in cord blood only would reflect the events during labor and elevation in both compartments would signify that fetal hypoxemia antedated labor and persisted during labor. Finally, it should be recognized that this study and the interpretation of its data are limited by the fact that sampling is restricted to one or, in the case of amniotic fluid, two time points during labor and delivery. It also is limited by the fact that detailed pharmacokinetic data for the rate of appearance and turnover of erythropoietin in amniotic fluid are not available. These issues must be addressed either in animal models or in the
h u m a n by amniocentesis coupled with cordocentesis before labor and r e p e a t e d scalp blood sampling in the course of labor. TM 20-22 This project would not have been possible without the technical expertise of Robert L. Schmidt and the help and cooperation of numerous obstetricians and midwives throughout northeast Ohio. REFERENCES
1. Lotgering FK, Gilbert RD, Longo LD. Maternal and fetal responses to exercise during pregnancy. Physiol Rev 1985; 65:1-36. 2. Clapp, JF, Rokey R, Treadway JI, Carpenter MW, Artal RM, Warrnes C. Exercise in pregnancy. Med Sci Sports Exerc 1992;24(suppl):S294-300. 3. Clapp JF. A clinical approach to exercise in pregnancy. Clin Sports Med 1994;13:443-58. 4. Carpenter MW, Sady SP, Hoegsberg B, Sady MA, Haydon B, Coustan DR. Fetal heart rate response to maternal exertion. JAMA 1988;259:3006-9. 5. Artal RM, Posner MD. Fetal responses to maternal exercise. In: Mittlemark RA, Wistwell RA, Drinkwater BL, eds. Exercise in pregnancy, 2nd ed. Baltimore: Williams and Wilkins, 1991:213-24. 6. Clapp JF. The effects of exercise on uterine blood flow. In: Rosenfeld CR, ed. The uterine circulation. Ithaca: Perinatology Press, 1989:289-310. 7. Clapp JF, Little KD, Capeless EL. The fetal heart rate response to various intensities of recreational exercise in mid and late pregnancy. AMJ OBsrEx GYNECOL1993;168: 198-206. 8. Krantz SB. Erythropoietin. Blood 1991;77:419-34. 9. Finne PH, Halvorsen S. Regulation of erythropoiesis in the fetus and newborn. Arch Dis Child 1972;47:683-8. 1O. Zanjani ED, Poster J, Burlington H. Liver as the primary site of erythropoietin formation in the fetus. J Lab Clin Med 1977;89:640-4. 11. Widness JA, Clemons GK, Teramo KA, Coustan DR, Schwartz R. Temporal relationship of fetal plasma erythropoietin to acute hypoxemia in the sheep and human. In: Stern L, Oh W, eds. Physiological foundations of perinatal care. New York: Elsevier, 1987:30-43. 12. Widness JA, Veng-Pederson P, Modi N, Schmidt RL, Chestnut DH. Developmental changes in erythropoietin kinetics in fetal and neonatal sheep. Pediatr Res 1990;28: 284. 13. Widness JA, Sawyer ST, Schmidt RL, Chestnut DH. Lack of maternal to fetal transfer of 125-1abelled-erythropoietin in sheep. J Dev Physiol 1991;15:139-43. 14. GitlinJD, Gitlin D. Protein binding by cell membranes and selective transfer of proteins from mother to young across tissue barriers. In: Hemmings WA, ed. Maternal foetal transmission of immunoglobulins. London: Cambridge University Press, 1974:113-21. 15. Teramo KA, Widness JA, Clemons GK, Voutilainen PEJ, McKinlay S, Schwartz R. Amniotic fluid erythropoietin correlates with umbilical plasma erythropoietin in normal and abnormal pregnancy. Obstet Gynecol 1987;69:710-6. 16. Rollins MD, Maxwell AP, Afrasiabi M, Halliday HL, Lappin TRJ. Cord blood erythropoietin, pH, PaO 2 and haematocrit following cesarian section before labour. Biol Neonate 1993;63:147-52. 17. Maier RF, Bohme K, Dudenhausen JW, Obladen M. Cord blood erythropoietin in relation to different markers of fetal hypoxia. Obstet Gynecol 1993;81:575-80. 18. Widness JA, Clemons GK, Garcia JF, Oh W, Schwartz R. Increased immunoreactive erythropoietin in cord serum after labor. Au J OBSTETGYNECOL1984;148:194-7. 19. Widness JA, Teramo KA, Clemons GK, et al. Correlation of the interpretation of fetal heart rate records with cord
Volume 172, Number 5 Am J Obstet Gynecol
plasma erythropoietin levels. Br J Obstet Gynaecol 1985; 92:326-32. 20. Georgieff MK, Landon MB, Mills MM, et al. Abnormal iron distribution in infants of diabetic mothers: spectrum and maternal antecedents. J Pediatr 1990;117:455-61. 21. Moya FR, Grannum PAT, Widness JA, Clemons GK, Copel JA, Hobbins JC. Erythropoietin in human fetuses with
Gagnon et al.
immune hemolytic anemia and hydrops fetalis. Obstet Gynecol 1993;82:353-8. 22. Salvesen DR, Brudenell JM, Snijders RJM, Ireland RM, Nicolaides KH. Fetal plasma erythropoietin in pregnancies complicated by maternal diabetes mellitus. AMJ OBSTET GYNECOL1993 ; 168:88-94.
Alterations in fetal and placental deoxyribonucleic acid synthesis rates after chronic fetal placental embolization R o b e r t G a g n o n , M D , a' b H o w a r d R u n d l e , B S c , c' d, ~ L a u r a J o h n s t o n , A C T , " b a n d V i c t o r K . M . H a r t , M D c' d, e
London, Ontario, Canada OBJECTIVE: Fetal growth and development are closely related to normal placental growth and function. We performed a study to determine the effect of a 10-day period of fetal hypoxemia induced by umbilical-placental hypoperfusion on tissue deoxyribonucleic acid synthesis rates in the 0.84 to 0.91 of gestation ovine fetus and placenta. STUDY DESIGN: Daily fetal placental embolization was performed in four chronically catheterized sheep fetuses until fetal arterial oxygen content decreased by - 3 0 % compared with preembolization values. Five control fetuses received vehicle only. On experimental day 10, the deoxyribonucleic acid synthesis rate was determined by injecting tritiated thymidine (1 mCi/kg) intravenously approximately 8 hours before the end of the study. RESULTS: Fetal arterial oxygen decreased from 3.2 - 0.1 (SEM) mmol/L preembolization to 2.2 -+ 0.2 mmol/L on day 10 (p < 0.001) and remained unchanged in controls. On day 10 deoxyribonucleic acid synthesis rates were significantly reduced in embolized fetuses compared with controls, by 38% in cotyledons (83.0 -+ 15.1 vs t33.7 + 9.9 disintegrations/min/~g deoxyribonucleic acid, p < 0.05), 28% in the left ventricular wall (36.8 - 3.7 vs 51.0 -+ 4.7 disintegrations/min/#g deoxyribonucleic acid, p < 0.05), and 45% in the quadriceps muscle (15.4 _+ 4.0 vs 28.1 _+ 3.0 disintegrations/min/l~g deoxyribonucleic acid, p < 0.05). Tritiated thymidine autoradiography demonstrated that cotyledonary deoxyribonucleic acid synthesis occurred exclusively in the fetal trophoblasts cells. CONCLUSION: We concluded that a reduction in cotyledonary, quadriceps muscle, and left ventricular myocardium deoxyribonucleic acid synthesis rates are the earliest adaptive mechanisms of fetal growth associated with development of umbilical-placental insufficiency. We speculate that alteration in the myocardial deoxyribonucleic acid synthesis rate could be a major contributing factor in the deterioration of fetal myocardial function associated with increased placental vascular resistance. (AM J OBSTETGYNECOL 1995;172:1451-8.)
Key words: Placental insufficiency, fetal hypoxia, fetal growth
From the Departments of Obstetrics and Gynaecolog'y,° Physiology,b Paediatrics,c Biochemistr3d and Anatomy/St. Joseph's Health Centre, Lawson Research Institute, Medical Research Council Group in Fetal and Neonatal Health and Development, University of Western Ontario. Supported by Medical Research Council of Canada and Physicians' Services Incorporated Foundation. Received for publication July 12, 1994; revised October 28, 1994; accepted November 7, 1994. Reprint requests: Robert Gagnon, MD, Department of Obstetrics and Gynaecology, St.Joseph's Health Centre, 268 Grosvenor St., London, Ontario, Canada N6A 4V2. Copyright © 1995 by Mosby-Year Book, Inc. 0002-9378/95 $3.00 + 0 6/1/61851
Fetal growth and development are closely related to normal placental growth and function. It has been shown that fetal placental embolization by latex microspheres is associated with progressive fetal hypoxemia and changes in umbilical artery Doppler flow velocity waveforms suggestive of an increase in placental vascular resistance similar to that observed in pregnancies complicated with intrauterine growth retardation (IUGR). 1.2 It is not clear yet how and to what extent fetal and placental growth would be altered during the development of umbilicalplacental hypoperfusion.l' 3 Changes in cellular growth 1451