Survival in the fetal rabbit exposed to intermittent hypoxia GORDON TOM
To investigate the time course and mechanisms of‘fetal O2 deprizlation, UP exposed 93 near-term, pegnant rabbits and their fetwes in utero to intermittent hypoxia. We udminktered a ga.s mixture containing 3, 4, 5 or 6 prr rent Oz in N2 to a rabbit for 0.5, I. 2, or jr minutes, alternately with a recozlery period oj air bwathing of 0.5, 1, 2, or 3 minutes. A given cyclical pattern was continued for 2 hours and fetal survival recorded I hour thc,wafter. Fetal survival decreased sign#icantly at louw OS IevrLs, longer duration., of exposure. and shorter durations of recovery intervals betwen ~~xposur~~s. Chancc~~ for survival were significantly greater among fetuses located at either end of the uterine horn. (AM. J. OBSTET. GYNECOL. 127: 428, 1977.)
these results occur because during each uterine contraction, pressure increases inside the uterus, decreasing maternal blood flow, and briefly interrupting the fetal oxygen supply. This report presents findings of survival of fetuses exposed intermittently to oxygen deprivation. We chose the duration and frequency of exposure to approximate normal human labor. We varied the intensity of the hypoxic challenge to include both levels that would be expected during normal labor, and more se\‘ere and damaging degrees of‘ hypoxia. We also cmphasize the imperfect parallel between labor and hypoxia. Labor obviously includes many changes from otherwise normal physiology (PH. hormones. and others) not related to oxygenation. We designed this study to investigate the following aspects of intermittent fetal hypoxia: (1) the relative importance of hypoxic intensity, duration, and duration of the recovery period upon fetal survival; (2) the inspired O2 concentrations resulting in 50 per cent fetal death with different patterns of hypoxic duration and recovery periods; (3) whether fetal death occurs before maternal death, and if so, under what circumstances: (4) to what extent O2 deficiency is cumulative or whether it is restored during the recover); period between hypoxic bouts; and (5) to determine what fetal characteristics correlate with survival.
ADEQUATE OXYGENATION is essential for normal growth and function of the central nervous, cardiovascular, and other systems in the developing fetus. The effects of prolonged oxygen deprivation on fetal well being have been studied intensively. Chronic hypoxic exposure impairs fetal survival. Intermittent oxygen deprivation, on the other hand, has been less well studied. Acute intermittent fetal O2 deprivation may occur during labor, particularly in “high-risk” pregnancy. The scalp blood Paz of pregnant animals and human beings during labor has been shown to fall about 5 mm. Hg at the contraction peak, but to return toward normal during the interval between contractions (see below). While not large, the consequences of these changes for the fetus are unknown. Presumably,
From the Division of Prrinatal Biology, Departments of Physiology and Gynecology, School qf Medicine, Loma Linda University. Recrwed
Reprint requests: Gordon G. Power, M.D., Depart. of Physiology and Perinatal Research, Loma Linda University School of Medicine, Loma Linda, Caltyornia 92354. *Dr. G. G. Power is recipient of United States Public Health Service Career Development Award No. 1 K4 HD 20, 253.
**Dr. L. D. Long0 is recipient of United States Public Health Sentice Career DeveloFent Award No. 2 K4 HD 23, 676.
The essence of the study was to expose near-term rabbits to low oxygen gas mixtures for varying intervals 428
of time, interspersed with intervals of normal oxygenation (Fig. 1). Following 2 hours of intermittent hypoxia. we recorded fetal survival. Rabbits were studied at 26 to 29 days of gestation (term 31 days). We secured a small face mask (dead spate, 20 ml.) over the rabbit’s nose and mouth and administered various oxygen-nitrogen gas mixtures at flow rates of 10 L. per minute. This rate adequately flushed ai\ ay exhaled gases and allowed a change from one gas to another in a few seconds. We monitored inflowing 02 concentrations (Beckman, Model E-2). maintaining them within 0.1 per cent of the predetermined level with a series of mixing valves, gas cylinders, a timing relay, and solenoids. The animal was allowed to breathe ai; f-or several minutes to become accustomed to the apparatus. Then cycles of air and hypoxic gas mixtures were begun. We maintainetl a given pattern throughout the experimental period in a given animal. Following intermittent hypoxia for 2 hours, we removed the face mask and left the animal undisturbed to breathe air for the next hour. After this time interval, the animal was killed, the uterus immediately opened, and fetuses were removed. We recorded the position of each fetus in the uterine horn and weighed it to the closest one hundredth of a gram. We classified fetuses as: (1) alive (moving in response to stimuli), (2) having recently died (no movement or heart beat) but otherwise appearing normal for gestational age, and (3) dead and appearing abnormal (having died some time prior to the experiment). These latter fetuses were not included in lhe analysis. We studied four levels of O2 concentration: 3, 4, 5, and 6 per cent oxygen in nitrogen: four durations of exposure: 0.5, 1,2. and 3 minutes; and four time intervals of recovery between hypoxic intervals: 0.5, 1, 2, and 3 minutes. Thus, 64 patterns of intermittent hypoxia formed the basis of the formal study protocol. Occasionally, however, we studied additional levels of oxygenation or repeated the study with the same pattern in more than one rabbit. In all, 93 rabbits and their 916 fetuses were investigated. In 14 rabbits, we determined maternal arterial blood gas values after 2 minutes’ exposure to different inspn-ed O2 concentrations. We measured l’oq. PCOB. and pH with microelectrodes (Radiometer, London Company, Model BMSS). Statistical methods. The average litter size was 10, with a range from three to 14 pups. Because of this variance in litter size, we used the percentage of fetuses surviving in each litter as the variable in calculating the statistics. We realized this introduced an error caused by non-Gallssian distribution of percentage survivals.
123456 TIME (min.)
Experimental procedure for a pregnant rabbit exposed 5 per cent inspired O2 for 1 minute intervals alternating with 2 minute recovery periods breathing air. The entire procedure was continued for 2 hours and fetal survival recorded 1 hour thereafter. Fig. 1. to
5 IAl,’ 0” z.5 0 ii RECOVERY=
t = Maternal
of two or more rabbits
2. Percentage of fetuses surviving intermittent hypoxia. Results are shown for 93 does and their 916 fetuses exposed to various hypoxic levels, durations. and recovery periods.
We tested the importance of this error by transforming the data numerically and repeating the analysis. We found the results were not altered appreciably. We analyzed the data with a stepwise linear regression (BMD02R). Initially. a regression was done with 10 variables, including litter size, litter weight, and several other variables, as well as the controlled variables in the pattern of hypoxia. At each step the multiple correlation coefficient of each of the remaining vari-
2 OF EXPOSURE
Fig. 3. Conditions for SO per cent fetal survival. Summary curves show combinations of hypoxic level, duration, and recovery that result in one half of fetuses dying. In some instances, data were insufficient to indicate a mean and only an upper limit is shown. ables was calculated. If this was not significant at the 0.05 level for any variable, then the least significant variable was eliminated and the procedure repeated. The four variables that remained explained 77 per cent of the variance of the experimental results. With multiple correlation the p values obtained are with all other variables held constant. It is these values that Iye report here.
Results Effect of variations in intensity, duration, and recovery time. Fig. 2 shows per cent of fetal survival for the four different oxygen concentrations and four different hypoxic durations. The four grids depict these values for the four different recovery periods. The figure also shows the per cent of fetal survival for other levels of hypoxia or duration studied to complete the fetal survival curves as discussed below. Also shown are instances in which a given value of survival is the mean 01’ two or more studies and those cases in which the maternal rabbit died. In each of the four grids of Fig. 2, fetal survival was highest (essentially 100 per cent) in the lower right corner of the grid; i.e., the region corresponding to the highest O2 level and shortest duration of hypoxia. Fetal survival was lowest (essentially 0 per cent) in the upper left region of a given grid: i.e., that area corresponding to the most severe degree of hypoxia and the longest duration. Between these two groups, lying in a band across the table from the upper right to lower left corners, are those fetuses in which the three factors interacted resulting in only a fraction of the fetuses dy-
ing. With shorter recovery periods, this band ou urrctl progressively nearer the lower right region ofthr grid. Each of these variables COI rclatetl highly \vith tC.c,~l survival. For instance, the intensity of hypoxia corrclated highly with fetal survkal, all else’ being equal (p < 0.01). Most f’etuses died when the doe breathed :\ per cent 02 for intervals of I minute or longer, \\-her~as most fetuses survived 6 per bunt Oz. regardless of the duration and recovery times. The maternal blood gas \,aluc-s indicate the sc\crit\ of the hypoxia exposure. Maternal arttrial Po,! fell tc, 26 mm. Hg (46 per cent satul.ation) after 2 minutca‘ exposure to 6 per c‘ellt 02. After rtlc Sillllt’ exposul’t time, maternal Pal ftll to 23 mm. Hg (24 per ((‘III saturation) while breathing 3 pt’r cent Op. to 19 I~IIII. Hg (1 7 per c.enr saturation) breathing 4 per- cc‘ilt Oz. and to 16 mm. Hg (I 2 per cent saturation) breathing :I per cent 02. ‘The duration of hypoxia also correlared closely \\ ith survival (p < 0.0 1). Thus, almost all (97 per cei~r) fetuses survived exposures f’or 30 seconds, regardless of intensity and recovery. The majority of fetuses died when the exposure equaled 3 minutes over the tull range of intensities and recover) times studied. The importance of recovery time ~\as greater than anticipated and correlated closely \\ith fktal survival (p < 0.01). For example, fetuses could not withstand maternal exposures to 3 per cent Oi for I minute wticli recovery times were 0.5 and 1 minute, but most survived this intensity and duration 1, hen r(~ over\ rimes were 2 and 3 minutes. 50 per cent fetal survival. M’e plotted the data shown in the four grids (Fig. 2) and estimated those combinations of intensity. duration, and recovery it’sulting in 50 per cent survival (based on smoothed turves for the composite rrsults). This permitted the data to be condensed and shown graphically as a familv of curves. Fig. 3 sho\vs this idealized summary. \Ye cstimate the curves shown to be lalicl to about t- 1 per cent Oz and t 1 minute of. duration md rccovc’rv. (They apply only for a total exposure time of 2 hour\) Fifty per cent survival was assoc.iated with relatively lo\\ OS levels for short durations or higher Op levels till longer periods. the recovery period being the same. Or from another standpoint. 50 per cent survival was associated with lower O2 le\Tels when the recover) time, was longer and \\ith higher O2 levels when the recover) time was shorter, the duration being the same. In about 7.5 per cent of all litters either all fetuses survived or all died. This occurred despite our efforts to arrange experimental conditions to study the rangrs critical to fetal survival. While the implications of this
“all or none” phenomenon are not apparent, it seems likely that more subtle tests of CNS, cardiac, and other vital functions would have shown a greater transition zone between normal and abnormal. Fetal death occurred before maternal death during intermittent hypoxia, but only over a surprisingly narrow range. The intermediate range of fetal death with maternal survival typically differed by only 1 per cent O2 or 0.5 to 1 minute of exposure time. Effect of position in uterus and weight of the fetus. Fetal survival also correlated with position within the uterus. In analyzing litters in which some fetuses died while others lived, fetal survival was significantly greater among fetuses located in the distal ends of the uterine horns (85 per cent) than for the average of all fetuses (76 per cent) (p < 0.05). Fetuses located in the midregion of a uterine horn \vere most likely to die, survival averaged 67 per cent. Since fetuses located in the ends of‘ the uterine horns were closer to the origin of the uterine and ovarian arteries, their improved survival might be because their placentas received a greater blood supply. Fetuses at the ends also weighted about 10 per cent more than their littermates. We did not find any correlation between fetal survival and litter size, litter weight, gestational age (from 26 to 29 days), maternal weight. or number of abnormal fetuses in the litter. We could not identify any interaction among the variables when tested with a general linear hypothesis.
Comment Physiologic implications. During labor, the fetus may be subjected to intermittent hypoxia. At the peak of a contraction. maternal blood flow into the placenta may decrease sharply with less oxygen exchange. The exact rate of placental blood flow or oxygen transfer has not been measured experimentally during a contraction in humans. Ahlquist and Woodbury’ first reported a reciprocal relation between the intensity of uterine contractions and uterine blood flow in anesthetized dogs. In unanesthetized dogs and sheep, Assali and associates3 showed that uterine blood flow decreased 30 per cent or more during uterine contractions, the decrease being roughly proportional to the intensity of the contractions. Using angiographic techniques in human subjects, Bore11 and associate?, ’ demonstrated decreased flow through the intervillous space of humans with increase in the nonuniformity distribution of uterine blood flow during contractions. Fetal blood oxygen tensions may fall significantly during a uterine contraction. Apparently, Corbit” first measured this effect. He demonstrated a decrease in
Survival with exposure to intermittent hypoxia
umbilical venous oxyhemoglobin saturation from 48 per cent during a control period to 30 per cent during an oxytocin-induced uterine contraction at the time of elective cesarean section. Several groups’. “. ‘-’ denlonstrated consistent and significant decreases in fetal scalp blood oxygen tensions during uterine contractions in women in labor. Studies in fetal monkeys’. ” and lambsI also have shown decrements of 4 to 6 mm. Hg during uterine contractions, whether spontaneous or induced with oxytocin. Estimates of the fetal blood O2 changes using a mathematical model indicate that placental O2 transfer falls 25 to 50 per cent below control levels during a contraction of moderate intensity.’ Relatively soon thereafter, fetal arterial PoZ falls. the response occurring in a matter of seconds, because fetal blood rapidly circulates between the placenta and peripheral tissues.*’ Quantitatively, fetal arterial PO2 is predicted to fall at most about 5 mm. Hg at the peak of a contraction, from about 24 to 19 mm. Hg. This dccrease is relatively small, and nowhere near the levels estimated to cause death in the present study. The decrease during a uterine contraction probably is not greater because the total O2 deficit accumulated during a contraction (estimated as 5 ml. for human subjects) derives from total fetal oxygen stores of about I6 ml. This line of reasoning points out the advantage of contractions lasting no longer than a minute interspersed with periods of recovery. It also emphasizes the importance of fetal hemoglobin as an 0, reservoir. As a contraction subsides, the placental O2 transfer rate is restored to precontraction rate, and under some circumstances, it may overshoot to greater than precontraction rates, thereby replenishing the 02 deficit accumulated during the contraction.’ Whether this replacement is entirely adequate, however, remains doubtful. Evidence suggests that repeated exposures to hypoxia, especially with frequent contractions separated by only short periods of recovery, cause progrcssive fetal deterioration. The results of the present study in which the recovery period was varied, lvhile the intensity and duration of hypoxia did not change, confirm this. It is also born out by clinical observations. Thus, the experimental evidence and results of the mathematical model suggest that normal labor may result in moderate fetal hypoxia, but with a safety margin more than adequate for survival. The hypoxia of labor, of course, will compound any pre-existing problem of impaired fetal-maternal exchange and under these circumstances, fetal oxygenation may become marginal. Fetal hypoxia is linked to neurophysiologic changes, neurological damage, and fetal death. In many cases of cerebral palsy, for example, periods of prolonged fetal
Bennett, and Longo
OI- labor. hypoxia.
Numerous studies hat-e explored different asp ts of‘ the effects of hypoxia on the developing fetus in uterc,. Methods of producing this hypoxia have included decreasing the availability of O2 to the maternal organism, decreasing blood flow to the uterus and placenta, decl-easing umbilical blood floF\ , decreasing the 02 c-apacity of fetal blood as ~vell as traumatic sellaration of the placenta. killing the mother or dccapitating the fetus. Fazekas. Alesander. and Himlvich”’ demonstrated in rats and a cat that 5 per cent O2 for 10 minutes resulted in maternal death, but that most of the fetuses survived and breathed spontaneously when removed. Other studies suggest that the more immature the fetus, the greater its resistance to the effects of hypoxia.“‘+‘“~ ” XIartin and Becker”’ reported a 30 per cent mortality rate of pregnant rats subjected to O2 concentrations that decreased from 21 to 4 per cent over a period ot 11 minutes. and held at 4 per cent for 5 minute>. The only other study using a specific end-point to study the effects of fetal hypoxia %$-asthat of h4ann’” and Mann. Prichard. and Symmes.‘” These pvorkers administered 0, 5, 7.j. and 15 per cent 0.’ to e~vcs and followed the electroencephalogram (EEC) until it became isoelectric, i.e., fetal death. With 7.5 per cellt OS inhalation, the EEG became isoelectric by 12.5 (? 1.2) minutes at Fvhich time umbilical vein Pea was about 5 mm. Hg. Some limitations of the experimental method. The results of’ this study must he viewed in light oi its limirations. One of these is that the end-point of death gaw no indication of‘ preceding, mow subtle neuropathologic and physiologic alterations occurring
in the fetus. .A\second limitation is that these result\ in rabbits arc’ most assuredly not dire& applicable to human sub,jects. While some parallels mav be possibk,. the differences between the spccics. their physiology oi labor, body size. and the maternal/fetal r\cighr ratio itI1 suggest that these cannot be precise A third limitation is that acute intermittent maternal hvposia constitutes ;111 imperfect model for fetal hypoxia during 1aho1.. Uterine (onlractions are assotiatcd with changes ill maternal and fetal placental l~lood flouts that arc IIOI duplitatrrl bv hypoxia. Thcrc arc’ other differencc~s in acid-base status and carbon dioxide elimination. In conclusion. intermittclit hypoxia \\ ill c.ausc tetal death with durations of exposure as short as 30 seconcl< if \‘ery IOXV oxygen tension levels arc’ rcac hcd. Tll~~ range of hypoxia necessary IO cause death probably is far more severe than the ftitus encounters during nov ma1 labor. Hypoxic death in rabbit pups appears to bc an “all or none” phenomenon. i.e.. under most ccmditions. either all f’etuses in the litter liw or die a\
but again, over which
the range of inspired this occurs is small.
02 deficiency may accumulare h) poxia. As the rc’co\ er) pe~~iotts shortet-, cannot
c~f Oz OI- death
We thank Dr. Jan Kuzma and Pat Brenneman assistance with calculations of statistical significanw.
REFERENCES Adamsons, K., Beard, R. W., and Myers, R. E.: Comparison of the composition of arterial venous and capillary blood of the fetal monkey during labor. AM. J. ORSTET. GYNECOL. 107: 435, 1970. Ahlquist, R. P.. and Woodbury, R. A.: Influence of drugs and uterine activity upon uterine blood flow, Fed. Proc. 6: 305, 1947.
Assali, N. S., Dasgupta, K., Kolin, A.. et al.: Measurement of uterine blood flow and uterine during spontaneous and induced
metabolism. V. Changes labor in unanesthetized
pregnant sheep and dogs, Am. J. Physiol. 195: 614, 1958. Berg, D., Kronenberger, G., and Kubli. F.: Statistische Untersuchungen zur subportalen Acidose des Feten. Arch. Gynaekol. 209: 34, 1970. 5. Berg. F., and Saling, E.: The oxygen partial pressures in the human fetus during labor and delivery, in Longo, L. D., and Bartels, H., editors: Respiratory Gas Exchange and Blood Flow in the Placenta, Bethesda, 1972. U. S.
Department of Health, Education & Welfare, USPHS. NIH, NICHD. DHEW Publication No. (NIH) 73-361. 6. Borell, C., Fernstrom, I., Ohlson, L., et al.: Effect of uterine contractions on the human uteroplacental blood circulation. An arteriographic study, AM. ,J. ORSXT. GYNECOL. 89: 881, 1964. Borrell, LT.. Fernstrom, I., Ohlson, L., et al.: Influence of uterine contractions on the uteroplacental blood How at term, AM. J. OBSTET. GYNECOL. 93: 44, 1965. Butler, L. A., Longo, L. D., and Power, G. G.: Placental blood flows and O2 transfer during uterine contractions: A mathematical model, J. Theor. Viol. 61: 81, 1976. Butler, N. R.. and Alberman. E. D.. Editors: Perinatal problems, The second report of the 1958 British perinatal mortality survey under the auspices of the National Birthday Trust Fund, Edinburgh. ., 1969. E. & S. Livingstone. ’ 10. Corhit. J. D.: Effect of oxytocin on oxygen exchange be-
t\veen maternal and fetal circulation at the time of cesarean section, AM. J. ORSTET. GYNECOL. 84: 1313, 1962. Dawes. G. S.: Foetal and neonatal physiology. A comparative study of the changes at birth, Chicago, 1968, Year Book Medical Publishers, Inc. Dawes, G. S.. Jacobson, H. N., Mott, J. C., et al.: Some observations on foetal and newborn rhesus monkeys, J. Physiol. (Lond.) 152: 271. 1960. Fazekas. J. F., Alexander, F. A. D., and Himwich, H. E.: Tolerance of the nelvborn to anoxia, Am. J. Physiol. 134: 281. 1911. Glass, Ii. G.. Snyder, F. F., and Webster, E.: The rate of decline in resistance to anoxia of rabbits, dogs and guinea-pigs from the onset of viability to adult life, Am. J. Physiol. 140: 609. 1944. (;ood~~n. J. W., and Paul, W. M.: Umbilical arterial oxygen tension of the fetal lamb in utero. Effect of oxytocic drugs, .4~. J. OBSTET. GYNECOL. 96: 362, 1966. Greiss, F. ti.. Jr.: Effect of labor on uterine blood flow. Observations on gravid ewes. AM. J. OBSTET. GYNECOL. 93:
17. Kabat. H.: The greater resistance of very young animals to arrest of the brain circulation, Am. J. Physiol. 130: 588, 1940.
18. Mann, L. I.: Effects of hypoxia on umbilical circulation and fetal metabolism, Am. J. Physiol. 218: 1453. 1970. 19. Mann, L. I., Prirhard, J. W., and Symmes, D.: EEG. ECG, and acidbase observations during acute feral hypoxia, Abf. J. OBSTET. GYNLCOL. 106:39. 1970. 20. Martin, J. C., and Becker, R. F.: The effects of maternal nicotine absorption or hypoxic episodes upon appetitive behavior of rat offspr-ing.. Des. Prychobiol. 4: 133. 197 1. 21. G. G., Power, and Longo. L. D.: Fetal circulation times and their implications for oxygenation, Gynecol. Invest. 6: 342, 1975. 22. Reiss, M., and Haurott-itz. F.: clber das verhalten junger und alter tiere bei erstickung, Klin. Wochenschr. 8: 743, 1929. 23. Renou, P., Newman, W., Lumley, J., t’t al.: Fetal scalp blood changes in relation to uterine contractions. J. Obstet. Gynaecol. Br. Commons. 75: 629, 1968. 24. Wulf, H.. Kunzel. W’., and Lehmann, V.: Vcrgleichende Untersuchgen der oktuellen Blutgase und des SaureBasen-Status im fetalen and maternen Kapillarblut wahrend der Geburt, Z. Geburtshilfe Gynaekol. 167: 1 13, 1967.