The effect of repeated acute hypoxaemia on fetal cardiovascular development in the sheep

Authors


Corresponding author C. Steyn: Department of Obstetrics and Gynaecology, University College London, 86–96 Chenies Mews, London WC1E 6HX, UK. Email: c.steyn@ucl.ac.uk

Abstract

  • 1Hypoxaemia during intrauterine life may be important in the development of cardiovascular diseases in later life. Thus it was the aim of this study to investigate the effect of repeated acute hypoxia on the cardiovascular development and growth of the fetus.
  • 2Fourteen fetal sheep (105–109 days gestational age) were instrumented with amniotic and vascular catheters, an electrocardiogram (ECG) electrode and a flow probe around the femoral artery. Seven animals were given repeated acute isocapnic hypoxaemia (Pa,O2 reduced to ca. 13 mmHg) for 1 h every day for 14 days and they were compared to seven animals which remained normoxic throughout with respect to fetal mean arterial blood pressure (MAP), fetal heart rate (FHR), and fetal baro- and chemoreflexes.
  • 3No differences were found between the two groups of fetuses in FHR, MAP, baro- or chemoreflexes, femoral blood flow, femoral vascular resistance or fetal growth.
  • 4Repeated acute hypoxaemia of a moderate degree over a period of 2 weeks in late gestation does not affect cardiovascular development or growth in the fetal sheep.

At present there is much interest in the idea that events occurring during intrauterine life play a major role in the development of cardiovascular diseases, such as hypertension, in later life (Barker, 1994). Previous work in both the sheep (Robinson et al. 1983) and the rat (Crowe et al. 1995b) suggests that experimental manipulations that are associated with hypoxaemia in utero alter the programming of fetal cardiovascular development. However as Robinson et al. (1983) used restriction of placental size and Crowe et al. (1995b) used severe maternal anaemia, the insults to the fetus are not purely hypoxaemic. We therefore felt that the question of whether hypoxaemia per se perturbs fetal cardiovascular development was worthy of further study.

The fetal reflex responses to a single acute 1 h hypoxaemic challenge in late gestation are now well established (Giussani et al. 1994). There is an initial bradycardia accompanied by a transient increase in mean arterial blood pressure (MAP), and a redistribution of combined ventricular output in favour of the brain, heart and adrenals at the expense of the gastrointestinal tract, kidneys, lungs, skin and skeletal muscle. In response to sustained hypoxaemia (in the absence of acidaemia) for up to 24 h there is a return of heart rate and blood pressure to normoxaemic levels, despite the partial pressure of arterial oxygen (Pa,O2) remaining low (see Bocking et al. 1992), although the redistribution of blood flow to the brain, heart and adrenals is maintained for at least 48 h (Bocking et al. 1988). The mechanisms responsible for the adaptation to hypoxaemia are as yet unknown, but recent work indicates that the peripheral chemoreceptors are involved, as in their absence circulating levels of catecholamines are significantly lower than in intact fetuses (Stein et al. 1997). Also, adrenocorticotrophic hormone (ACTH), cortisol and arginine vasopressin are markedly attenuated in denervated compared with intact fetuses in response to prolonged (6 h) hypoxaemia (Stein et al. 1998). Altered levels of these hormones may have consequences for the cardiovascular development of the fetus (P. Hawkins, C. Crowe, H. G. McGarrigle, T. Saito, T. Ozaki, L. L. Stratford, D. E. Noakes & M. A. Hanson, personal communication) and neonate (Hawkins et al. 1997). Investigations into the development of heart rate and blood pressure in response to longer periods (weeks) of hypoxaemia have so far yielded little further insight into the mechanisms involved, as the results obtained vary from investigator to investigator, depending on the length of exposure, timing, severity and methods used to induce hypoxaemia (see Table 1). In addition to the literature on the effects of chronic hypoxaemia on the fetus, there are a number of studies investigating the acute effects of repeated hypoxaemic/asphyxic insults (lasting seconds or minutes in duration) (Jensen et al. 1985; Akagi et al. 1988; Giussani et al. 1996; Green et al. 1997a, b; Unno et al. 1997). The findings of various workers (Jensen et al. 1987; Akagi et al. 1988; Giussani et al. 1996; Green et al. 1997a, b) suggest adaptation of the chemoreceptors in response to repeated hypoxic insults. There is, however, to our knowledge no published information regarding the effects of repeated acute hypoxaemia (an insult that could perhaps be regarded as lying somewhere between acute and chronic hypoxaemia) on the development of the fetus.

Table 1. Various methods by which chronic hypoxaemia has been produced in fetal sheep, showing the fetal Pa,O2 levels attained, the change in fetal Pa,O2 and the effect on MAP, FHR, haematocrit and fetal growth
MethodDuration of insult (days)GA administered (days) P a,O2 achieved (mmHg)ΔPa,O2 (mmHg)FHRMAPHctBody wt
  1. Data derived from a Davis & Hohimer, 1991; bTrudinger et al. 1987; Block et al. 1989; Gagnon et al. 1994;Carter et al. 1996;cKamitomoet al. 1994; dMurotsuki et al.1997;eMostello et al. 1991;fDanielet al.1989;gKamitomo et al. 1992;hRobinson et al. 1983. GA, gestation age; FI,O2, inspired O2 fraction; MAP, mean arterial blood pressure; Hct, haematocrit; FHR, fetal heart rate;↑, increase;↓, decrease; =, no change.

Fetal anaemia a7124–13114.84.5=
Embolization (maternal)b8–10120–13417–19.94.3–6====/↓
Reduced FI,O2c14124–13817.110.5=
Embolization (fetal)d21109–12916.54.7==
Maternal anaemiae30110–13822.02.0=
Reduced maternal plasma volumef30115–14517.23.9=
High altitudeg9030–12019.34.0==
Carunclectomyh1400–14016.44.9

A naturally occurring event that may be viewed as repeated acute hypoxaemia in the fetus is the occurrence of contractures. There is little pronounced uterine activity during pregnancy, except for the last 6–18 h before parturition. However, it is now well established that there are spontaneously occurring uterine contractions, known as ‘contractures’, that occur throughout pregnancy. These contractures are of low amplitude (causing only 3–10 mmHg changes in amniotic pressure) and low frequency (between 0.5 and 3 per hour) with a duration of 3–15 min. Thus they have a much longer duration and lower amplitude than the high frequency, high amplitude and relatively short duration ‘contractions’ that are associated with labour (Jenkin & Nathanielsz, 1994). Contractures are linked with a modest increase in intra-amniotic pressure, decreased uterine blood flow, decreased fetal Pa,O2 and arterial oxygen content (Ca,O2) and an increase in fetal arterial partial pressure of CO2 (Pa,CO2), haemoglobin concentration and arterial blood pressure (Llanos et al. 1988). More recently, measurements taken over a 6 h period have shown that the changes in fetal blood gases that occur with uterine contractures can be fairly large, with variations from 2.5 to 5.5 % in Pa,O2 and 3.6 to 7.0 % in arterial O2 saturation (Sa,O2) (Woudstra et al. 1995). It is not known how much these physiological changes in blood gases affect the fetus. However, as described above, contractures induce effects other than hypoxaemia alone and our study was not designed to mimic their effect.

Lastly, sleep apnoea in adult humans is a syndrome that also results in repeated reductions in Pa,O2 and is associated with primary hypertension (Fletcher et al. 1995). Sleep apnoea has also been observed in pregnant women (Kowall et al. 1989), although how commonly it occurs during pregnancy is unknown. Schoenfeld et al. (1989) reported acid-base changes in the fetus during maternal apnoeic episodes, which they suggested were due to compromised fetal oxygen delivery. They suggested that sleep apnoea during pregnancy may affect fetal growth and well-being. It is possible therefore that repeated acute hypoxaemia in the fetus may result in an increase in blood pressure, endocrine changes and reduced fetal growth.

Hence, in view of growing interest in the role of intrauterine events in producing disease in later life, and because both acute asphyxia and prolonged sustained hypoxaemia have been reported to produce adaptation in fetal cardiovascular responses (see above), we were interested in investigating the effects of repeated acute hypoxaemia on the fetus. It was our hypothesis that repeated acute hypoxaemia would cause changes in fetal cardiovascular development as a result of resetting of the peripheral chemoreceptor responses. This may then result in a chronic increase in sympathetic tone and thus vascular resistance, which could in turn result in hypertension, baroreceptor resetting, and hence a chronic increase in blood pressure. The aim of this study was, therefore, to investigate the effect of a repeated acute hypoxic insult on the cardiovascular development and growth of the fetus.

METHODS

Surgical methods

All the ewes used in this study were mules (Border Leicester × Cheviot). They were brought into the laboratory at least 2 days prior to surgery and maintained in metabolic crates. They were fed a ration of about 500 g of nuts each per day and hay ad libitum. Fetal gestational age (GA) for both groups was the same and they were evenly distributed in terms of sex and plurality (Table 2).

Table 2. Gestational age, number of male, female, twin and singleton fetuses, crown-rump length and body, heart, lung, liver and kidney weights in C and H fetuses
 ControlHypoxia
  1. GA, gestational age; M, male; F, female; T, twin; S, singleton; CRL, crown-rump length.*P < 0.05, C vs. H.

GA(days)128±1128±1
Sex4M, 3F4M, 3F
Twin/singleton3T,4S3T, 4S
CRL (cm)44.3±1.448.5±1.4
Body weight (g)3345±5363964±251
Heart weight (g)25.4±2.633.2±2.7
Lung weight (g)100.3±14.5120.2±7.5
Liver weight (g)122.2±17.1105.1±8.1
Kidney weight (g)13.8±0.511.8±0.6*

Fourteen fetuses between 105 and 109 days gestation were instrumented under halothane anaesthesia (2 % in O2 after induction with thiopentone (May & Baker Ltd., Dagenham, UK), 1 g, i.v.) with carotid artery, jugular vein and amniotic catheters, an ECG electrode and a Transonic flow probe around the femoral artery (in 12 fetuses). Blood flow data was only obtained in four control group (C) and three hypoxia group (H) fetuses due to the fact that the flow probes did not function after implantation in some animals. A maternal pedal vein was catheterized for drug administration. Antibiotics were given to the ewe (4 ml Streptopen i.m., Pitman and Moore, Crewe Hall, Cheshire, UK; 300 mg Crystapen i.v., Glaxo; and 40 mg Cidomycin i.v., Roussel, Uxbridge, UK) and fetus (150 mg Crystapen i.v., and 150 mg Crystapen and 40 mg Cidomycin intra-amniotically) once daily for 5 days post-operatively.

Experimental protocol

After the 5 day recovery period experiments began. All fetuses were monitored for 3 h each day for 14 days. After 1 h of control recording, seven fetuses were subjected to a daily 1 h episode of acute isocapnic hypoxaemia (hypoxia group, H), by lowering maternal inspired oxygen fraction (FI,O2) to reduce fetal Pa,O2 to 12–14 mmHg, then monitored during an hour of return to normoxia. Seven control (C) fetuses were studied for 3 h in normoxia. Maternal FI,O2 was manipulated by altering the gas mixture delivered into a polythene bag secured over the ewe's head. During the 3 h of experimentation the ewes were unable to eat, but they were free to eat prior to and after removal of the polythene bag.

The chemoreflex response was assessed by the fetal heart rate (FHR) at 5 min and the MAP at 15 min of hypoxia. On day 1, before experimentation, and on day 14, the last experimental day, each fetus was given a single bolus dose (75–100 μg i.v.) of the α1-adrenoceptor agonist phenylephrine, to elevate blood pressure so that the baroreflex could be measured. The baroreflex was assessed by plotting the mean R-R interval of individual animals in each group, measured at 0, 2, 4, 6, 8, 10, 15, 20, 30, 40, 50 and 60 s after the injection of phenylephrine, against the MAP measured at the same times (i.e. while the pressure was rising).

At the end of the study, ewes were killed by an overdose of pentobarbitone (40 ml Euthatal i.v., Rhône Mérieux, Harlow, Essex, UK) and a post-mortem was carried out, where fetal body and organ weights and crown-rump length (CRL) were recorded.

All experiments were carried out in accordance with UK Home Office regulations (Animals Scientific Procedures Act 1986).

Data collection

All basal data values were derived from measurements over the first hour of experimental recording on each day. Pressures were recorded using standard pressure transducers and amplifiers (Gould); heart rate was recorded by passing the ECG signal through an amplifier and tachometer (Gould Biotach); femoral blood flow was measured by flow meter (T206, Transonic Systems, Inc., Ithaca, NY, USA). All signals were recorded onto disk using MacLab/8 hardware and data acquisition software. MAP, FHR and femoral blood flow (FBF) were calculated by averaging blood pressure, heart rate and flow, respectively, over 1 min at 15 min intervals. In addition, measurements were made at 5 min intervals during the first 15 min of hypoxia, as this is the period over which rapid reflex effects are manifested. MAP for each day was calculated as the mean of values at 15, 30 and 45 min for all the animals in each of the two groups. Femoral vascular resistance (FVR) was calculated from MAP divided by FBF.

At 30, 70, 100, 135 and 165 min, 0.5 ml of blood was drawn from the fetal carotid artery for blood gas, ion and haemoglobin analysis. Blood gases, pH and haematocrit were measured on a blood gas analyser (BGE, lnstrumentation Laboratory, Warrington, Cheshire, UK; values corrected to 39.5°C). A haemoximeter was used for measurement of haemoglobin (Instrumentation Laboratory, CO-Oximeter 482), and glucose and lactate were measured in whole blood by a glucose/lactate analyser (2300 Stat Plus, YSI, Farnborough, Hants, UK).

Statistical analysis

Values are expressed as means ±s.e.m. We considered P < 0.05 to be significant.

Basal data. Body weights, organ weights, crown-rump length and gestational age were compared by Student's unpaired t test. The mean values of the cardiovascular measurements, blood gases, blood metabolites and plasma ions for C and H fetuses at 115 and 128 days gestation were compared by paired t test in order to determine whether there was a change between the start and the end of the 2 week period. In order to determine whether the trajectory of developing MAP, systolic blood pressure (SBP), diastolic blood pressure (DBP), FHR, FBF, FVR and rate- pressure product was different between C and H fetuses, values for each variable were reduced to summary measures. In order to do this a line of best fit was derived by linear regression for the data points for each individual animal over the 2 week period, and its slope was used as a summary measure for that parameter in that animal. Summary measures were then compared by unpaired t test. Blood gases were also expressed as a summary measure, by calculating the average of all the measurements over the 14 days, and compared by unpaired t test (see Fig. 6). MAP, SBP, DBP, rate-pressure product, FBF, FVR, haematocrit, haemoglobin, glucose, lactate and blood gases were also analysed using one-way ANOVA to compare the trend over the 2 weeks between C and H groups.

Figure 6.

Fetal arterial blood gases

The graphs on the left-hand side show mean basal values of pH, Pa,CO2 and Pa,O2 over the 14 days of the study in C (□) and H (▪) fetuses. The graphs on the right-hand side show pH, Pa,CO2 and Pa,O2expressed as a mean summary measure for the entire 14 days in C (□) and H (▪) fetuses. Values are means ±s.e.m.***P < 0.001, C vs. H (summary measures, unpaired t test).

Baroreflex data. The entire baroreflex curves were analysed by one-way ANOVA in order to test whether there was a shift in the position of the curve between 115 and 128 days gestation. To measure the gain of the baroreflex, the linear pressor part of the curves for each animal were demarcated visually and a line of best fit was drawn through the points. The slope of the resulting line for each animal was then calculated. These slopes were expressed as a summary measure slope and these were compared by paired and unpaired t test in order to assess changes in the gain of the reflex between 115 and 128 days gestation and between groups, respectively.

Chemoreflex data. Assessment of the chemoreflex on each day was carried out by calculating the fall in FHR and FBF and the rise in FVR at 5 min after the onset of hypoxaemia, and the rise in MAP 15 min after this onset, in H fetuses. It was analysed using summary measures: as above, linear regression through the data points for each individual animal over the 2 week period was used to produce a line of best fit, the slope of which was used as a summary measure and compared by unpaired t test.

RESULTS

Blood pressure and heart rate

MAP was not significantly different between C and H fetuses, but both groups showed a significant increase in MAP between 115 and 128 days GA (P < 0.05) (Fig. 1). There was no significant increase of SBP and DBP in either group over the 14 days. When analysed using ANOVA, both SBP and DBP were significantly higher in H fetuses compared with C fetuses (P < 0.05 in both cases), but no such difference between groups was found when analysed using summary measures, reflecting the inter-animal variation. FHR decreased significantly between 115 and 128 days GA in both C (P < 0.01) and H (P < 0.01) fetuses and was similar in the two groups (Fig. 1). The rate-pressure product (an index of cardiac work), likewise, was not significantly different between C and H fetuses, and it did not change significantly in either group with increasing gestational age (Fig. 1).

Figure 1.

Development of MAP, SBP and DBP, FHR and rate-pressure product

Graphs showing the development of MAP, SBP, DBP, FHR and rate-pressure product between 115 and 128 days GA in C (○/□) and H (•/▪) fetuses. Values are means ±s.e.m.*P < 0.05, **P < 0.01, 115 days GA vs. 128 days GA (Student's paired t test).

Cardiovascular reflexes

The cardiovascular chemoreflex was assessed in H fetuses from the fall in FHR and FBF and the rise in FVR at 5 min after the onset of hypoxaemia, and the rise in MAP 15 min after this onset. During hypoxaemia, Pa,O2 was reduced from ca. 25 mmHg to ca. 14 mmHg. The level of hypoxaemia induced was not significantly different between the beginning and the end of the study. The FHR and MAP response to hypoxaemia did not alter significantly from the beginning to the end of the study (Fig. 2), nor did the change in FBF (−0.4 ± 2 ml min−1 at 115 days gestation, −21.9 ± 16 ml min−1 at 128 days gestation) or FVR (1.1 ± 0.8 mmHg min ml−1 at 115 days gestation, 3.1 ± 0.6 mmHg min ml−1 at 128 days gestation).

Figure 2.

Developmental change in the bradycardic and hypertensive responses to hypoxaemia in H fetuses

Graphs showing the change in the chemoreflex with increasing gestational age and the Pa,O2 levels achieved on each day (H fetuses). Values are means ±s.e.m. HR, heart rate.

There was a significant shift of the baroreflex curve to the right between 115 days and 128 days GA in both C and H fetuses (Fig. 3), suggesting that the range of baroreceptor sensitivity reset in both groups as MAP rose over this period (Fig. 1). The slope of the pressor part of the baroreflex curve was not significantly different between 115 and 128 days GA within either group, or indeed between groups (Fig. 4).

Figure 3.

Baroreflex curves

Graphs showing baroreflex as the change in R-R interval produced by a rise in MAP in C (n= 6) and H (n= 5) fetuses between 115 days GA (▿) and 128 days GA (▴). Values are means ±s.e.m.**P < 0.01, 115 days vs. 128 days GA (one-way ANOVA).

Figure 4.

Gain of the baroreflex response

Graphs showing gain of the baroreflex curve, measured as the change in R-R interval for the change in MAP, in C (n= 6) and H (n= 5) fetuses between 115 days GA (▪) and 128 days GA (□). Values are means ±s.e.m. There was no significant difference between groups (summary measures).

Femoral blood flow and vascular resistance

FBF increased significantly in C fetuses (P < 0.05) but there was no significant change in FVR. In H fetuses there was no significant change in either FBF or FVR. There was no significant difference of either FBF or FVR between groups (Fig. 5).

Figure 5.

Femoral blood flow and femoral vascular resistance

Graphs showing the development of FBF and FVR in C (n= 4; □) and H (n= 3; ▪) fetuses. Values are means ±s.e.m.*P < 0.05 (C fetuses), 115 days GA vs. 128 days GA (paired t test).

Blood gases, haematocrit, haemoglobin, glucose, lactate and ions

Figure 6 shows the change in mean basal values of pH, Pa,CO2 and Pa,O2 over the 14 days of the study. There was also no significant increase or decrease over time in any of these variables and there was no significant difference between groups. However, when expressed as a summary measure for the 14 days, Pa,O2 was found to be significantly higher in H fetuses than in C fetuses (P < 0.001) (Fig. 6).

Haematocrit and haemoglobin values decreased significantly in H fetuses over the 14 days of the study (P < 0.05 for both) (Fig. 7). There was a tendency for haematocrit and haemoglobin to decrease in C fetuses, but the decrease was not significant. Despite the significant decrease of haemoglobin levels in H fetuses, there was no significant difference between groups nor was there a significant difference between them in haematocrit.

Figure 7.

Fetal haematocrit and haemoglobin

Graphs showing haematocrit and haemoglobin levels in C (□) and H (▪) fetuses over the 14 day study period. Values are means ±s.e.m.*P < 0.05 (H fetuses), 114 days GA vs. 128 days GA (paired t test).

Glucose levels were somewhat variable but there was no significant difference between groups, and nor was there any significant change in glucose levels from the beginning to the end of the study in either C or H fetuses. Lactate levels appeared to increase between 115 days and 128 days GA, but the increase was not significant for either group. Lactate values were, however, significantly different over the 14 day study period between C and H fetuses (ANOVA, P < 0.001) (Fig. 8), although when analysed by slope summary measures/t test there was no significant difference between groups.

Figure 8.

Fetal arterial blood glucose and lactate

Graphs showing blood glucose and lactate levels in C (□) or H (▪) fetuses. Values are means ±s.e.m.***P < 0.001, C vs. H (one-way ANOVA).

The concentrations in blood of Na+, K+ and Ca2+ were similar in C and H fetuses and did not change significantly over the 2 week study period (Fig. 9).

Figure 9.

Fetal blood plasma ions

Graphs showing Na+, K+ and Ca2+ concentrations in the blood of C (□) and H (▪) fetuses. Values are means ±s.e.m.

Growth

Fetal growth was unaffected by repeated acute hypoxaemia, as can be seen from body and organ measurements taken post mortem (Table 2). There was no significant difference in CRL, body, heart, lung or liver weights, but kidney weights in H fetuses were found to be significantly smaller than in C fetuses (Table 2).

DISCUSSION

The purpose of this study was to investigate the effects of repeated acute hypoxaemia on fetal cardiovascular development and responses. The effects of a 1 h hypoxic episode, such as that used, have been well documented (Giussani et al. 1994). They are characterized by an initial chemoreflexly mediated bradycardia, occurring within 2–3 min after the onset of hypoxaemia, and a transitory rise in blood pressure, the magnitude of each being dependent on the extent to which the blood gases and pH change. Combined ventricular output (CVO) is redistributed in favour of the cerebral, myocardial and adrenal circulations at the expense of the gastrointestinal, renal, pulmonary, cutaneous and skeletal muscle beds. The peripheral vasoconstriction and rise in blood pressure are partly due to a carotid chemoreflexly mediated increase in sympathetic activity operating at α-adrenergic receptors, and part is non-reflex in nature and is the result of elevated levels of circulating hormones - catecholamines, arginine vasopressin (AVP), atrial natriuretic peptide (ANP), ACTH, cortisol and angiotensin II (AII). Para/autocrine factors are also involved in the response to hypoxaemia, e.g. nitric oxide (NO). We were interested to see whether, if this acute challenge was repeated daily for a period of 2 weeks, it would result in perturbations in fetal cardiovascular development and reflex responses. Our results show that repeated acute hypoxaemia of this degree and for a duration of 14 days does not have a significant effect on MAP, FHR or chemo- and baroreflex responses. Furthermore, fetal growth was not affected.

Blood pressure and heart rate development

There was a progressive increase in MAP and decrease in FHR over the period of the study (Fig. 1), as has previously been observed in the late gestation fetal sheep (Boddy et al. 1974; Kitanaka et al. 1989; Daniel et al. 1989; Mostello et al. 1991; Kamitomo et al. 1994; Gagnon et al. 1994; Murotsuki et al. 1997). However, the gestational changes of MAP, SBP, DBP and FHR were the same in both C and H fetuses (Fig. 1), suggesting that repeated acute hypoxaemia of the degree and duration (14 days) employed in this study does not affect cardiovascular development. Similarly, there was no difference between groups in the development of FBF or FVR (Fig. 5). The increase in FBF was most probably due to the increasing size of the fetuses. As there was a decrease in FVR with increasing gestational age, the increase in MAP must have been due to increased CVO and/or a rise in vascular resistance in beds other than the femoral, for both C and H fetuses.

No previous investigators have examined the effects of repeated acute hypoxaemia on the development of blood pressure in the fetus, but there are such studies in the adult where a pattern of episodic hypoxia was used in rats to mimic that seen during obstructive sleep apnoea in humans. In young male rats exposed to episodic hypoxia every 30 s for 7 h per day over 35 days it was found that MAP increased, and this increase was shown to be chemoreflexly mediated; the peripheral sympathetic nerves also play a role in such a chronic increase of pressure in response to episodic acute hypoxia (see Fletcher et al. 1995). It was recently demonstrated that the renal nerves and a rise in plasma catecholamines are important as efferent components of the chemoreceptor-mediated increase in blood pressure (Bao et al. 1997). However, the precise mechanisms whereby acute increases in blood pressure are translated into a chronic increase are not known. It does appear, however, that the rise in blood pressure is gradual, since it was found that rats exposed to less than 30 days episodic hypoxia showed no significant increase in MAP (Fletcher et al. 1992).

Thus, there is the possibility that repeated acute hypoxaemia for longer than 14 days in our fetuses would have resulted in hypertension. Murotsuki et al. (1997) found that, when they performed repetitive placental embolization for 21 days, fetal MAP was elevated by the end of the study, but when they embolized for only 10 days (Gagnon et al. 1994), MAP did not rise. Similarly, Trudinger et al. (1987), Block et al. (1989) and Carter et al. (1996) found that MAP and FHR in their fetuses were unaffected by repetitive embolization for a period of either 8 days (Carter et al. 1996) or 9 days (Trudinger et al. 1987; Block et al. 1989) in late gestation. The decrease in Pa,O2 for 1 h daily that we imposed on H fetuses is not dissimilar to the changes in Pa,O2 and Ca,O2 which are seen in fetuses where fetal placental embolization is carried out on a daily basis (see Gagnon et al. 1994; Murotsuki et al. 1997). In those studies fetuses were injected daily with non-radioactive microspheres through the descending aorta over a period of 10–21 days, so as to decrease fetal arterial oxygen content by 30–35 % of the pre-embolization value. Each day when the microspheres were injected, Ca,O2 decreased, as expected, but then there was a partial recovery in fetal oxygenation. Repeated embolization is not directly comparable to our repeated acute hypoxaemia challenge, however, because there is a gradual decrease in basal (pre-embolization) oxygenation levels from the beginning of the study to the end. In contrast, basal Pa,O2 values in our H fetuses, although not higher in absolute terms, were higher than those in C fetuses when expressed as a single mean for the whole 2 week period (Fig. 6). In seeking to make comparisons between repetitive placental embolization and our repeated acute hypoxia challenge, we must also remember that embolization results in a decrease in placental blood flow due to the increase in umbilical-placental resistance, seen as an increase in the umbilical artery resistance index (Gagnon et al. 1994; Murotsuki et al. 1997), but that hypoxaemia for up to an hour in late gestation is reported to produce an increase in placental blood flow (Bocking et al. 1988; Rurak et al. 1990). Thus, the immediate haemodynamic effects of the two types of challenge are quite different.

It therefore appears that the method by which hypoxaemia is produced is an important factor in determining cardiovascular adaptation and outcome. A range of methods have been employed previously, and their effects on MAP and FHR are summarized in Table 1. Apart from placental embolization (discussed above) there have been a number of studies investigating the effects of chronic hypoxaemia, using other methods, on aspects of fetal development including MAP and FHR. Davis & Hohimer (1991) found that fetal sheep made anaemic for 7 days developed higher FHR and lower MAP than controls. However, it is important to remember that anaemia may also be accompanied by a decrease in blood viscosity which would affect vascular resistance and hence blood pressure. Kamitomo et al. (1994) reported that fetal sheep made hypoxaemic for 14 days, as a consequence of reduced maternal FI,O2, developed a lower FHR than controls by the end of the study, but MAP was not significantly different. However, they were not able to maintain the fetuses isocapnic, and a drop in Pa,CO2 occurred from day 3 onwards, which complicates interpretation. Daniel et al. (1989) induced fetal hypoxaemia for 30 days (between about 111 and 141 days gestation) in fetal sheep by preventing the normal expansion of maternal blood volume (20 ml day−1) occurring during pregnancy. They found that hypoxaemic fetuses failed to show the expected gestational increase in blood pressure and that the decrease in FHR was smaller than that seen in control fetuses so that, by the end of their study, hypoxaemic fetuses had a lower systolic blood pressure and higher FHR than control animals. Mostello et al. (1991) reported that fetal hypoxaemia for 30 days secondary to chronic maternal anaemia did not affect the developmental decrease of FHR but caused a smaller increase in MAP, so hypoxaemic fetuses had a lower MAP than controls by the end of the study. It is possible that there may be placental adaptations to maternal anaemia, such as a decrease in resistance, that could contribute to the decreased MAP observed in that study. Fetal guinea-pigs developed both a lower MAP and a lower FHR by 60–64 days gestation after hypoxaemia for 30 days as a result of uterine artery ligation (Detmer et al. 1991). High-altitude hypoxaemia for 90 days (30–120 days gestation) has been shown to cause elevated MAP but no change of FHR in late gestation fetal sheep (Kamitomo et al. 1992). Sheep fetuses that were hypoxaemic for an even longer period, as a result of carunclectomy, had a lower MAP and higher FHR than controls (Robinson et al. 1983).

Lastly, gestational age at exposure is probably also an important factor. It is known that the cardiovascular responses of the fetal sheep to hypoxaemia before about 110 days GA are different from those of the fetal sheep at more than 110 days GA (Iwamoto et al. 1989). Our fetuses were older than 110 days GA, thus had relatively more mature cardiovascular reflexes. It is possible that had they been exposed to the repeated hypoxia challenge earlier on, when their reflexes were less mature, there may have been an effect on their development and thus on blood pressure also.

Clearly the mechanisms involved in cardiovascular adaptation and function are complex, with a number of factors acting together to determine fetal cardiovascular development. Those which predominate will depend on the method used to produce the insult, whether it produces hypoxaemia alone, its timing, intensity and duration.

Cardiovascular reflexes

Jensen et al. (1985, 1987) investigated the effects of repeated fetal asphyxia, by occlusion of the maternal abdominal aorta, on cardiovascular variables. They found that in response to each asphyxic episode there was a bradycardia which increased when the duration of the asphyxia was increased. The blood pressure response varied with the number of asphyxial episodes. They also showed that as the duration of asphyxia increased, and as the mean Sa,O2 decreased, there was an increase in the plasma concentrations of adrenaline and noradrenaline (Jensen et al. 1987). This suggests that a longer duration of asphyxia causes increased sympathetic activation, which is mediated in part by increased arterial chemoreceptor activity. Similarly, Mallard et al. (1995) found that episodic asphyxia, by occlusion of the umbilical cord, caused bradycardia during each occlusion which was accompanied by hypertension during the first occlusion, but during subsequent occlusions blood pressure progressively decreased, which they suggested to be indicative of increasing sensitization of the heart to further insults. Asphyxia is a much more severe insult than the hypoxaemia which our fetuses experienced, thus it is difficult to make comparisons. However, there is the suggestion from the studies of Mallard et al. (1995) and Jensen et al. (1987), in the late gestation fetal sheep, that a greater severity of insult causes alterations in reflex control, at least in the acute situation. Interestingly in this regard, human adults suffering from sleep apnoea show a pressor response that is positively correlated with the frequency and severity of apnoeas per hour of sleep (Hedner et al. 1992). Giussani et al. (1996) investigated the cardiovascular responses to repeated, partial umbilical cord occlusions in the fetal sheep, producing only mild reductions in Pa,O2 (from about 21 to about 16 mmHg). In response to each hypoxic episode there was little change in blood pressure, but a rapid and pronounced fall in heart rate and a rapid transient decrease in femoral blood flow and increase in femoral vascular resistance occurred. Similar to the findings of Giussani et al. (1996), Green et al. (1997b) found that the FHR responses to umbilical cord occlusion (caused by complete inflation of an occluder cuff) were blunted by the second day of experimentation, out of a study period of 4 days. The studies of Giussani et al. (1996) and Green et al. (1997b) indicate that a blunting in the magnitude of the chemoreflex response occurred. Whether this was at the level of the arterial chemoreceptors, in the brainstem or in the efferent limb of the response is not known.

In our study there was no significant attenuation or potentiation of the bradycardia, hypertension or vasoconstriction produced in response to hypoxia over the course of the study, i.e. from first to last hypoxic insult (Fig. 2). This shows that there was no alteration in the magnitude of the chemoreflex. Whether this difference from the effects seen as a result of repeated asphyxia is due to the greater intensity, or to the different nature of the stimulus is not known. Unfortunately there are no data available describing the gestational changes in chemoreflex sensitivity in fetuses that have had no previous exposure to hypoxaemic episodes, therefore we are unable to say whether the unaltered chemoreflex response that we observed in H fetuses is ‘normal’ or not.

As there was no difference between H and C fetuses in MAP and FHR, and there was no change in the chemoreflex of H fetuses over time, it was not surprising to find that the baroreflex was also not different between the two groups. The two groups showed a similar shift of the baroreflex curve upwards and to the right (Fig. 3), and there was no difference between groups in the gain of the baroreflex at either 115 or 128 days gestation (Fig. 4). It has previously been shown that there is a decrease in the carotid baroreceptor responses during late gestation (Blanco et al. 1988). We did not see a significant decrease in gain, probably due to the large variability in blood pressure response to phenylephrine in individual animals, but there was the suggestion that the gain of the reflex decreased in both groups from 114 to 128 days gestation.

Growth

In our fetuses exposed to repeated acute hypoxaemia we did not find any effect on body growth (Table 2), similar to the findings of Gagnon et al. (1995). Organ growth was also unaffected, except for the kidneys which were smaller in H fetuses (Table 2), perhaps as a result of a reduction in blood flow, as chronic hypoxaemia is associated with a decrease in blood flow to the kidneys (Kamitomo et al. 1993). It is also possible that there was a change in various of the growth messengers, e.g. the insulin-like growth factors (IGFs). Asymmetrical intra-uterine growth retardation (IUGR) in the human baby has been found to be associated with a decrease in kidney weight and a reduction in nephron number, and babies of normal birthweight who died of sudden infant death syndrome (SIDS) have been found to have a reduced number of nephrons (Hinchliffe et al. 1993). This suggests that renal development may be a more sensitive marker of adverse uterine conditions than altered growth of the other organs or total body weight. Reduced kidney size, if accompanied by a reduction in nephron number, may mean that there is a reduced capacity for the secretion of renin, which would result in perturbation of the renin-angiotensin system and thus cardiovascular responses.

Conclusion

The results of this study show that repeated acute hypoxaemia of a moderate degree over a period of 2 weeks does not affect fetal cardiovascular development in late gestation fetal sheep. Clearly, the fetus is able to compensate for such an insult without sustained effects. Thus, our original hypothesis that repeated acute hypoxaemia would result in perturbations in the development of cardiovascular reflexes and blood pressure is not confirmed. However, whilst there was no effect of repeated acute hypoxaemia on cardiovascular development, individual animals showed different trajectories of blood pressure development (these have been reported in Crowe et al. 1995b). Future studies should examine the role of the placenta and kidney in producing such effects on fetal cardiovascular development.

Acknowledgements

We are grateful to The Wellcome Trust for financial support.

Ancillary