A novel method for controlled and reversible long term compression of the umbilical cord in fetal sheep

Abstract

1. In fetal sheep during late gestation the aims of the present study were to (1) develop a technique for inducing prolonged but reversible periods of controlled compression of the umbilical cord and (2) characterise the cardiovascular, endocrine and metabolic responses to this challenge.

2. Under 1-2 % halothane anaesthesia, 16 Welsh Mountain sheep fetuses were chronically instrumented at 118 ± 2 days of gestation (term is ca 145 days) with an inflatable occluder cuff around the umbilical cord, amniotic and femoral vascular catheters and with transit-time flow probes around the contra-lateral femoral artery and an umbilical artery. At 125 days, umbilical blood flow was reduced by 30 % from a pre-determined 24 h baseline for 3 days by automated servo-controlled inflation of the occluder cuff (n = 8). The occluder was then deflated allowing return of umbilical blood flow to baseline. The remaining eight fetuses were used as sham-operated controls in which the occluder was not inflated throughout the protocol. Fetal cardiovascular variables were recorded at 8 s intervals and arterial blood samples taken for measurement of blood gases, glucose and lactate and plasma adrenaline, noradrenaline and vasopressin concentration throughout the study.

3. Automated servo-controlled inflation of the occluder cuff, programmed to reduce umbilical blood flow by 30 % from baseline, reduced umbilical blood flow by 30.2 ± 1.7 %, with a coefficient of variation during compression of 6.5 ± 1.1 %. Sustained partial compression of the umbilical cord produced falls in fetal arterial pH, P_a,O2, percentage O₂ saturation of haemoglobin, and hindlimb oxygen delivery, and increases in P_a,CO2, haemoglobin concentration, arterial blood oxygen carrying capacity and in blood glucose and lactate concentrations. While the reductions in P_a,O2, percentage saturation of haemoglobin and hindlimb oxygen delivery and the increase in P_a,CO2 were sustained throughout compression, the reduction in arterial pH and the increase in arterial oxygen carrying capacity had returned towards baseline values by 48 h compression. Fetal blood lactate concentrations reached a peak at 8 h of compression and, thereafter, were maintained at an elevated level relative to baseline.

4. Partial compression of the umbilical cord produced fetal hypertension, a reduction in femoral blood flow and, consequently, an increase in calculated fetal femoral vascular resistance for the duration of the challenge. In addition, the fall in heart rate measured in sham control fetuses by the end of the study, did not occur in cord-compressed fetuses. Cosinor analysis on 24 h rhythms of cardiovascular data indicated a significant increase in the amplitude of the 24 h rhythm in heart rate in cord-compressed fetuses relative to sham controls during the period of compression or sham-compression. Furthermore, cord compression led to an increase in fetal plasma noradrenaline, but not adrenaline and vasopressin concentrations relative to sham control fetuses.

5. In conclusion, a novel reversible method for controlled, long-term compression of the umbilical cord in sheep has been developed. The data show that sustained, partial compression of the umbilical cord produced moderate but sustained asphyxia, which resolved after the end of the compression period, and induced changes in fetal cardiovascular, endocrine and metabolic functions.

During gestation, periods of adverse intrauterine conditions may be of varying duration, severity and recovery. Previous studies have shown that the fetus is known to withstand short periods of adverse intrauterine conditions. For example, acute (1 h) episodes of hypoxaemia, of the type which may occur during labour and delivery (Rurak et al. 1997), elicit co-ordinated cardiovascular, endocrine and metabolic responses that facilitate physiological adaptation of the fetus to the period of reduced oxygen availability (Giussani et al. 1994b).

Less is known about the co-ordinated response of the fetus to longer (hours to days) periods of adverse intrauterine conditions. Hence, in sheep, a number of experimental techniques have been used to produce longer periods of adversity in utero. Most common, perhaps, are methods to restrict the placental capacity to transport nutrients and oxygen to the ovine fetus. Thus, surgical removal of endometrial caruncles prior to conception (carunclectomy, Robinson et al. 1979), or repeated injection of microspheres into the uterine (Creasy et al. 1972) or umbilical (Gagnon et al. 1994) circulations (embolisation) reduce total placental size and placental surface area for exchange, lead to fetal hypoglycaemia and hypoxaemia and, thereby, retard fetal growth. These methods are both irreversible and severe as indicated by the extent of brain damage, degree of growth retardation and incidence of fetal mortality (Clapp et al. 1988; Mallard et al. 1998; Rees et al. 1998). Such severe fetal compromise is relatively rare in clinical practice and has been reported to occur in only a small proportion (< 3 %) of human pregnancies in developed societies (Grantham-McGregor, 1998).

Even less information is available about fetal exposure to prolonged, but reversible periods of adverse intrauterine conditions. The few studies that have been performed have investigated the effects of fetal exposure to prolonged hypobaric hypoxaemia (Kitanaka et al. 1989), maternal undernutrition (Hanson, 1999) or compression of the fetal abdominal aorta (Anderson & Faber 1984; Anderson et al. 1986, 1987, 1993). Investigation of reversible adverse intrauterine conditions is important as they are just as likely to affect human pregnancies, and fetal exposure to a reversible challenge earlier in gestation has been associated with a greater susceptibility to fetal cerebral asphyxia during labour and delivery (Mann, 1986) with subsequent postnatal neurodevelopmental handicap (Dijxhoorn et al. 1987; Clapp et al. 1988).

Another common form of reversible adverse intrauterine conditions in human pregnancies is compression of the umbilical cord, which has been reported to occur at an incidence of ≥ 40 % (Rayburn et al. 1981; Clapp et al. 1999, 2000), resulting from nuchal cord (Clapp et al. 2000), torsion of the umbilicus during gestation (Rayburn et al. 1981), oligohydramnios (Leveno et al. 1984) or compression of the cord during the actual processes of labour and delivery (Wheeler & Greene 1975). Despite this, no study to date has investigated the effects on fetal physiology of fetal exposure to prolonged periods of adverse intrauterine conditions produced by reversible compression of the umbilical cord.

Hence, the aims of the present study were to (1) develop an experimental technique for inducing prolonged but reversible periods of mild adverse intrauterine conditions produced by controlled compression of the umbilical cord; and (2) characterise the fetal cardiovascular, endocrine and metabolic responses to the challenge.

Preliminary accounts of some of these data have already been published (Gardner et al. 1999, 2000).

METHODS

Surgical preparation

Sixteen Welsh Mountain ewes carrying singleton pregnancies of known gestational age were used in the study. All procedures were performed under the UK Animals (Scientific Procedures) Act, 1986. All animals were fasted for 24 h prior to surgery.

Surgery was performed under strict aseptic conditions at 118 ± 2 days of gestation (dGA; term being ca 145 dGA). Anaesthesia was induced with sodium thiopentone (20 mg kg⁻¹i.v. Intraval Sodium; Rhone Mérieux, Dublin, Ireland) and maintained with 1-2 % halothane in 50:50 O₂-N₂O. In brief, following mid-line abdominal and uterine incisions, the fetal head was exteriorised for insertion of carotid artery and jugular vein catheters (i.d. 0.86 mm, o.d. 1.52 mm; Critchly Electrical Products, Auburn, NSW, Australia) with the tips of the catheters extended to the ascending aorta and superior vena cava, respectively. The catheters were plugged with sterile brass pins and the uterine incision closed in layers. The fetal hindlimbs were subsequently exteriorised through a second uterine incision for insertion of femoral artery (i.d. 0.86 mm, o.d. 1.52 mm) and femoral vein (i.d. 0.56 mm, o.d. 0.96 mm) catheters, which were extended into the descending aorta and inferior vena cava, respectively. A further catheter was anchored onto the fetal hindlimb for recording of the reference pressure in the amniotic cavity. Transit-time flow transducers (Transonics, Ithaca, NY, USA) were placed around the contralateral femoral artery (2/3RS) and around an umbilical artery (4RS; within the fetal abdominal cavity) for continuous measurement of femoral and umbilical flows, respectively (Gardner et al. 2001). Measurement of femoral blood flow has been shown to provide a good continuous index of blood flow distribution to peripheral circulations (Giussani et al. 1994b) while implantation of a transit-time flow transducer around a single umbilical artery has been validated to provide a reliable measurement of 50 % of the total umbilical blood flow during basal and adverse intrauterine conditions (D. S. Gardner & D. A. Giussani, unpublished observations). In addition, an inflatable occluder cuff (OC20HD, In Vivo Metrics, Healdsburg, CA, USA) was positioned around the proximal end of the umbilical cord and anchored to the fetal abdominal wall so as to avoid contact with the cord when not inflated (Giussani et al. 1997). The second uterine incision and abdomen were closed in layers. A Teflon catheter was placed in the maternal femoral artery and extended to the descending aorta. Antibiotics were administered to the fetus through the femoral vein (300 mg ampicillin; Penbritin, SmithKline Beecham Animal Health, Surrey, UK) and amniotic (300 mg ampicillin) catheters. All catheters were filled with heparinised saline (80 i.u. heparin ml⁻¹ in 0.9 % NaCl) and plugged with brass pins. Then, together with the flow probes and occluder leads, the catheters were exteriorised through an incision in the maternal flank and housed in a pouch sutured to the maternal skin.

Postoperative care

Animals were housed in individual pens with access to hay and water ad libitum. Concentrates were fed twice daily (100 g; Sheep Nuts No. 6; H&C Beart Ltd, Kings Lynn, UK). All ewes received antibiotics (0.20-0.25 mg kg⁻¹i.m. Depocillin; Mycofarm, Cambridge, UK) and analgesia (10-20 mg kg⁻¹ oral phenylbutazone; Equipalozone Paste, Arnolds Veterinary Products Ltd, Shropshire, UK) immediately after surgery and daily for 3 days. Patency of fetal vascular catheters was maintained by a slow continuous infusion of heparinised saline (25 i.u. heparin ml⁻¹ at 0.1 ml h⁻¹ in 0.9 % NaCl) containing antibiotic (1 mg ml⁻¹ benzylpenicillin; Crystapen, Schering-Plough, Animal Health Division, Welwyn Garden City, UK).

Experimental procedure

No experiment was performed until at least 5 days after surgery. In all fetuses at 124 ± 0 days gestation, baseline mean unilateral umbilical blood flow (UBF) was determined over a 24 h period. The animals were then divided randomly into two experimental groups. In eight of these fetuses (umbilical cord compressed, UCC), at 125 ± 0 dGA, the occluder cuff was inflated using an automated servo-controlled compression regulator programmed to reduce UBF from the calculated 24 h baseline UBF by 30 % for 3 days (see below). The occluder cuff was then deflated allowing the return of UBF to baseline. In the remaining eight fetuses the occluder cuff was not inflated throughout the duration of the protocol. These animals were designated as sham-operated animals (control). At the end of the protocol, the ewes and fetuses were humanely killed using a lethal dose of sodium pentobarbitone (200 mg kg⁻¹ Pentoject; Animal Ltd, York, UK) and the positions of implanted catheters and flow probes were confirmed.

Measurements and calculations

Maternal (descending aorta) and fetal (carotid and descending aorta) arterial blood samples (0.5 ml) were drawn into sterile syringes daily for measurement of arterial blood gases, percentage saturation of O₂ in haemoglobin, haemoglobin concentration and acid/base status using an ABL5 blood gas analyser and OSM2 haemoximeter (Radiometer, Copenhagen, Denmark). In addition, maternal and fetal arterial blood samples (2 ml) were taken simultaneously for measurement of blood gases, acid/base status and hormone concentrations before umbilical cord compression at -1 day and -1 h, during umbilical cord compression at +1 h, +8 h, +1 day, +2 day and +3 day and after deflation of the occluder cuff at 1 day. Measurements in maternal and fetal blood were corrected to 38 and 39.5 °C, respectively. Maternal and fetal arterial blood glucose and lactate concentrations were measured using an automated analyser (Yellow Springs 2300 Stat Plus glucose/lactate analyser; YSI, Farnborough, UK).

Fetal arterial blood oxygen content (C_a,O2; mmol l⁻¹), oxygen capacity (O_2cap; mmol l⁻¹), oxygen delivery (O_2del; μmol min⁻¹) and glucose delivery (Glu_del) to the fetal hindlimb were calculated using the following equations

where [Hb] (g dl⁻¹) is the blood concentration of haemoglobin, O₂SatHb (%) is the percentage oxygen saturation of haemoglobin where 1 molecule of Hb (M_r, 64 450) binds 4 molecules of oxygen

where femoral blood flow is in millilitres per minute

The contribution of oxygen dissolved in plasma is regarded as negligible (Owens et al. 1987). Blood samples for hormone analyses were either collected into K⁺/EDTA-treated (0.5 ml, vasopressin) or chilled EGTA (5.0 μmol ml⁻¹ blood) and glutathione (40 μmol ml⁻¹ blood)-treated tubes (1 ml, catecholamines) and centrifuged immediately at 4000 r.p.m. for 4 min at 4 °C. Plasma samples were stored at -70 °C until analyses.

Hormone analyses

The plasma catecholamines, adrenaline and noradrenaline, were analysed by high performance liquid chromatography (HPLC) using electrochemical detection (Fowden et al. 1998). The samples were prepared by absorption of 250 μl of plasma onto acid-washed alumina and 20 μl aliquots of the 100 μl perchloric acid elutes were injected onto the column. Dihydroxy benzylamine was added as the internal standard to each plasma sample before absorption. The limit of sensitivity for the assay was 20 pg ml⁻¹ for adrenaline and noradrenaline. The interassay coefficients of variation for adrenaline and noradrenaline were 7.3 and 6.2 %, respectively.

Plasma arginine vasopressin (AVP) concentrations were measured using a commercially available double-antibody radioimmunoassay kit (Nichols Institute Diagnostics Ltd, Saffron Walden, UK) following separation from plasma proteins by methanol extraction and chromatography as described previously (Giussani et al. 1994a). Briefly, chromatographic separation was achieved using pre-washed columns (25 % methanol; Sep-Pak Plus C18, Waters Corp., MA, USA) onto which 500 μl aliquots of plasma-0.1 m HCl (50:50 v/v) were loaded. Columns were then washed with 10 ml of 4 % acetic acid before elution with 2.5 ml methanol. The elutes were dried under N₂ gas at 37 °C, reconstituted with 1 ml of PBS and separated into two aliquots made up to 300 μl with PBS. Rabbit anti-AVP antiserum (100 μl) was added to each sample, incubated for 24 h at 5 °C and a further 100 μl of ¹²⁵I-labelled AVP was added. The tubes were then vortexed and again incubated for 24 h at 5 °C. Donkey anti-rabbit antiserum precipitant (100 μl) was added to each tube to separate bound and free hormone fractions and incubated at room temperature for 30 min. Deionised water (1 ml) was then added to each sample and all tubes centrifuged at 5 BC for 15 min at 2000 g. The resultant supernatant was subsequently counted for radioactive content on a gamma counter. The lower detection limit of the assay was 1.3 pg ml⁻¹. The intra-assay coefficient of variation for four control plasma pools (mean concentrations: 3.2, 9.9, 12.2 and 28.9 pg ml⁻¹) were 10.0, 6.7, 3.7 and 4.6 %, respectively. The interassay coefficient of variation for two plasma samples (2.71 and 5.55 pg ml⁻¹ AVP) were 4.1 and 9.8 %, respectively. The anti-AVP antiserum was highly specific for [Arg⁸]vasopressin with cross-reactivities of < 0.05 % against oxytocin, [Lys⁸]vasopressin and [Arg⁸]vasotocin.

Data collection and analyses

Fetal arterial blood pressure was corrected for amniotic pressure. Fetal femoral and umbilical blood flows were measured with a T201 or T206 flow meter (Transonics Inc.). Fetal heart rate was triggered from the flow pulse. Changes in femoral and umbilical vascular resistances were calculated according to Ohm's principle by dividing arterial blood pressure (corrected for amniotic pressure) by either femoral or umbilical blood flow (Giussani et al. 1993). All analogue signals for calibrated fetal arterial blood pressure, amniotic pressure, heart rate, and femoral and umbilical blood flows were recorded continuously throughout the study using a data acquisition system. All signals were digitised, displayed and subsequently stored at 8 s intervals on disk by custom software (NI-DAQ, National Instruments, Austin, TX) running on a PC. Files were subsequently analysed using Microsoft Excel spreadsheets.

The automated compression regulator

The arrangement of the automated compression regulator is shown in Fig. 1. The output (1) from the Transonics flow box measuring UBF was connected in parallel to a data acquisition system (2a), which registered and logged UBF at 8 s intervals on an Microsoft Excel spreadsheet file, and the automated compression regulator unit (2b). The compression regulator unit (The Physiological Laboratory, University of Cambridge) was linked to an infusion/withdrawal pump (Model 2000, Harvard Instruments, USA) carrying a 60 ml syringe filled with sterile saline (3). Luer-lock connectors attached the syringe to the tubing of the saline-filled occluder cuff (4a) and to a pressure transducer (4b; Cobe, Argon Scientific, TX, USA). From the calculated 24 h baseline value for UBF any desired degree of compression could be entered on a precalibrated dial on the automated compression regulator unit. The unit would then drive the pump to either infuse saline, resulting in an inflation of the cuff and a concomitant reduction in UBF, or withdraw saline, reducing the constriction and raising UBF. Information from the pressure transducer was also registered by the automated compression regulator unit (5). If pressure inside the occluder cuff exceeded the pre-set value of 1100 mmHg indicating severe compression (based upon inflation of the cuff in air), the unit would automatically trigger a depressurisation solenoid (6). While this mechanism served as a security measure it was never used during the present series of experiments.

Figure 1. The automated servo-controlled compression regulator unit.

UBF, umbilical blood flow.

Statistical analyses

Values for all variables are expressed as means ± s.e.m. All measured variables were tested for normality of distribution and compared between a baseline value and values at +1, +2, and +3 days of umbilical cord compression/sham compression and at +1 day of recovery using two-way ANOVA with repeated measures (SigmaStat; SPSS Inc., Chicago, IL, USA) comparing the effect of time (days of compression/recovery), group (control vs. UCC) and interactions between group and time. For cardiovascular data, summary measures for statistical analyses were derived from the mean of each 24 h period throughout the experimental protocol (Matthews et al. 1990). For endocrine and metabolic data, measured values at each time point were included in the statistical analyses. Where a significant effect of time was indicated, within-group comparisons were made using one-way ANOVA with repeated measures. Where a significant effect of group was indicated multiple comparisons between groups were made with Student's t test for unpaired data. In addition, cosinor analysis was performed on hourly average values for cardiovascular variables to determine the presence of 24 h rhythms (Nelson et al. 1979). In brief, analysis of rhythms was performed only on those data that demonstrated a sinusoidal-type rhythmic pattern according to a posteriori examination of individual data for arterial blood pressure and heart rate. Among animals that revealed a significant 24 h rhythm, values for mesor (value about which oscillation occurs), peak times (timing of high point) and amplitudes (half the difference between highest and lowest values) were compared with Student's t test for unpaired data.

For all comparisons statistical significance was accepted when P < 0.05.

RESULTS

Compression of the umbilical cord by automated inflation of the occluder cuff

Baseline unilateral UBF, calculated over the 24 h preceding umbilical cord compression, was similar in control (191 ± 25 ml min⁻¹) and UCC fetuses (204 ± 22 ml min⁻¹; Fig. 2). Automated servo-controlled inflation of the occluder cuff, programmed to reduce UBF by 30 % from baseline, reduced UBF by 30.2 ± 1.7 % over the 3 days of umbilical cord compression (Fig. 2). UBF was maintained throughout the 3 days of umbilical cord compression to within 13 ± 4 ml min⁻¹ representing a coefficient of variation of 6.5 ± 1.1 % for the duration of the challenge. After deflation of the cord occluder, UBF returned toward baseline values in four fetuses and to a level significantly greater than baseline in the remaining four fetuses. However, when data were combined, the effect fell outside significance (Fig. 2).

Figure 2. Umbilical blood flow and umbilical vascular resistance.

Values are means minus s.e.m. of umbilical blood flow and means plus s.e.m. of umbilical vascular resistance (30 min average) for a baseline period (1 day prior to umbilical cord compression), for 3 days of umbilical cord compression/sham compression and for 1 day of recovery after umbilical cord compression/sham compression in sham control (n = 8) and UCC (n = 8) fetuses. *P < 0.05, baseline vs. compression.

Arterial blood gas and metabolic status

Maternal

Arterial blood gas status was similar between control and UCC ewes during the baseline period (control, n = 8: pH_a, 7.50 ± 0.01; P_a,CO2, 34.4 ± 0.8 mmHg; P_a,O2, 99.7 ± 2.2 mmHg, O₂SatHb, 93.5 ± 1.4 %; UCC, n = 8: pH_a, 7.49 ± 0.01; P_a,CO2, 35.3 ± 1.7 mmHg; P_a,O2, 98.1 ± 2.2 mmHg; O₂SatHb, 93.3 ± 1.3 %) and remained unaltered from baseline throughout the duration of the experimental protocol. Similarly, baseline values for maternal arterial blood glucose and lactate concentration between control and UCC ewes were not different (control, n = 8: glucose, 2.39 ± 0.18 mmol l⁻¹; lactate, 0.352 ± 0.068 mmol l⁻¹; UCC, n = 8: glucose, 2.71 ± 0.18 mmol l⁻¹; lactate, 0.290 ± 0.026 mmol l⁻¹) and remained unaltered from baseline throughout the duration of the experimental protocol.

Fetal

Baseline values for fetal arterial blood gases, pH_a, blood glucose and lactate were similar in control and UCC groups (Fig. 3 and Fig. 4). In addition, baseline C_a,O2 (3.77 ± 0.19 vs. 3.48 ± 0.19 mmol l⁻¹), O_2cap (55.6 ± 1.5 vs. 52.4 ± 2.9 mmol l⁻¹) and hindlimb O_2del (96.5 ± 10.0 vs. 112 ± 15 μmol min⁻¹) were not significantly different between control and UCC fetuses. Arterial blood gas and metabolic status remained unchanged from baseline throughout the experimental protocol in control fetuses (Figs 3–5). In contrast, inflation of the occluder cuff in UCC fetuses produced mild fetal asphyxia for the duration of the challenge with significant falls in pH_a, P_a,O2 and O₂SatHb together with a significant increase in arterial P_a,CO2 (Fig. 3).

Figure 3. Fetal arterial blood gas status.

Values are means ± s.e.m. for control (○, n = 8) and UCC (•, n = 8) fetuses. Fetal arterial blood samples were collected for measurement of blood gases during the baseline period at -1 day and -1 h, at 1 h, 8 h, 1 day, 2 days, 3 days of umbilical cord compression/sham compression and after 1 day of recovery from umbilical cord compression/sham compression. * P < 0.05, baseline vs. umbilical cord compression or recovery. †P < 0.05, control vs. UCC fetuses. P_a,CO2, arterial CO₂ partial pressure; P_a,O2, arterial O₂ partial pressure, O₂SatHb, percentage saturation of O₂ in haemoglobin.

Figure 4. Fetal blood glucose and lactate concentration.

Values are means ± s.e.m. for control (○, n = 8) and UCC (•, n = 8) fetuses. Fetal arterial blood samples were collected for measurement of blood glucose and lactate concentration during the baseline period at -1 day and -1 h, at 1 h, 8 h, 1 day, 2 days, 3 days of umbilical cord compression/sham compression and after 1 day of recovery from umbilical cord compression/sham compression. * P < 0.05, baseline vs. umbilical cord compression or recovery. †P < 0.05, control vs. UCC fetuses.

Figure 5. Fetal blood haemoglobin concentration, blood oxygen content and hindlimb oxygen delivery.

Values are means ± s.e.m. for control (○, n = 8) and UCC (•, n = 8) fetuses. Fetal arterial blood samples were collected for measurement of blood gases during the baseline period at -1 day and -1 h, at 1 h, 8 h, 1 day, 2 days, 3 days of umbilical cord compression/sham compression and after 1 day of recovery from umbilical cord compression/sham compression. * P < 0.05, baseline vs. umbilical cord compression or recovery. †P < 0.05, control vs. UCC fetuses. C_a,O2, arterial blood O₂ content; O_2del, arterial O₂ delivery to the fetal hindlimb.

Cord compression elicited an immediate and significant (P < 0.05) increase in [Hb]_a (from 8.4 ± 0.4 to 9.6 ± 0.4 g dl⁻¹; Fig. 5) which offset the reductions in P_a,O2 and O₂SatHb such that arterial blood oxygen content (3.48 ± 0.19 mmol l⁻¹) remained similar to baseline throughout the 3 days of the challenge. However, a significant reduction in hindlimb O_2del occurred by 24 h of cord compression (from 113 ± 16 to 85 ± 11 μmol min⁻¹; Fig. 5). In addition, a transient increase in blood glucose concentration occurred by 8 h of compression in UCC fetuses, which returned to basal values by 2 days of compression (Fig. 4). After 3 days of compression blood glucose concentrations were again significantly greater than baseline (Fig. 4). Hindlimb glucose delivery was significantly greater in UCC fetuses than in controls throughout the study period (26.2 ± 5.3 vs. 20.6 ± 2.1 μmol min⁻¹ for UCC and control fetuses, respectively, P = 0.004). Compression of the umbilical cord produced an increase in fetal blood lactate concentration at 8 h compression which was sustained relative to baseline, albeit at a reduced level, for the duration of the challenge (Fig. 4). After 1 day of recovery from cord compression, values for all variables, with the exception of blood lactate concentration, had returned to baseline levels. Blood lactate concentration returned to basal values by 2 days after the compression period in UCC fetuses (data not shown).

Fetal cardiovascular measurements

During the 24 h baseline period all cardiovascular variables were similar in control and UCC fetuses and appropriate for fetuses at this gestational age (Fig. 6). In control fetuses, cardiovascular variables remained unchanged from baseline with the exception of fetal heart rate, which declined progressively over the 3 days of the experiment (Fig. 6). Cosinor analysis on 24 h rhythms of cardiovascular data in control fetuses revealed significant rhythms in arterial blood pressure in 5 of the 8, and in heart rate in 4 of the 8, fetuses. In UCC fetuses, compression of the umbilical cord produced a significant increase in mean arterial blood pressure and significant decreases in femoral and umbilical blood flow for the duration of the challenge (Fig. 6). Consequently, calculated mean femoral vascular and umbilical vascular resistances increased by 24 ± 13 % and 49 ± 9 %, respectively, during cord compression and in UCC fetuses (Fig. 2 and Fig. 6). The maximum increase in femoral vascular resistance (46 ± 16 %) in cord-compressed fetuses occurred during the last 12 h of compression. In contrast to control fetuses, values for heart rate remained unaltered from baseline in UCC fetuses (Fig. 6). Cosinor analysis on 24 h rhythms of cardiovascular data in UCC fetuses revealed significant 24 h rhythms in heart rate in 6 of the 8 fetuses. The amplitude of the 24 h rhythm in heart rate was significantly greater in UCC relative to control fetuses (2.0 ± 0.6 vs. 6.9 ± 0.6 beats min⁻¹ for control (n = 4) and UCC (n = 6) fetuses, respectively). Following deflation of the occluder cuff, all cardiovascular variables returned to baseline levels within 1 day of recovery.

Figure 6. Fetal cardiovascular responses to 3 days umbilical cord compression.

Values are means ± s.e.m. of cardiovascular data (30 min average) for the baseline period (1 day prior to umbilical cord compression), for 3 days of umbilical cord compression/sham compression and for 1 day after umbilical cord compression/sham compression for control (n = 8) and UCC (n = 8) fetuses. * P < 0.05, baseline vs. umbilical cord compression/sham compression or recovery.

Fetal endocrine measurements

Baseline concentrations of noradrenaline (431 ± 89 vs. 505 ± 100 pg ml⁻¹), adrenaline (152 ± 41 vs. 88 ± 25 pg ml⁻¹) and AVP (6.55 ± 1.45 vs. 8.05 ± 2.02 pg ml⁻¹) were similar in control and UCC fetuses, respectively. Plasma concentration of all three hormones did not change significantly from baseline by the end of the sham compression period in control fetuses. In contrast, while plasma adrenaline and AVP concentration remained unchanged from baseline, a significant increase in plasma noradrenaline occurred in UCC fetuses during the period of cord compression (Fig. 7).

Figure 7. Fetal plasma noradrenaline concentration during cord compression.

Values are means ± s.e.m. for control (○, n = 8) and UCC (•, n = 8) fetuses. Fetal arterial blood samples were collected for measurement of plasma noradrenaline and adrenaline by HPLC during the baseline period at -1 day and -1 h, at 1 h, 8 h, 1 day, 2 days, 3 days of umbilical cord compression/sham compression and after 1 day of recovery from umbilical cord compression/sham compression. * P < 0.05, baseline vs. umbilical cord compression or recovery.

DISCUSSION

The aims of the present study were to develop a technique for inducing prolonged but reversible periods of adverse intrauterine conditions produced by compression of the umbilical cord, and to characterise the fetal responses to the challenge. In the present study the combination of a servo-controlled compression regulator with the implantation of an inflatable occluder around the umbilical cord and a Transonics flow probe around an umbilical artery inside the fetus permitted the extent of any reduction in umbilical flow, and thus of fetal compromise, to be adjusted continuously and maintained precisely for the duration of the study. The technique developed is reversible, with umbilical flow returning to baseline values within an hour after deflation of the occluder cuff. The data show that compression of the umbilical cord, to reduce umbilical blood flow by 30 % of basal values for 3 days, produces significant alterations in fetal cardiovascular, metabolic and endocrine functions.

Previous studies investigating the fetal responses to umbilical cord compression have mostly considered the fetal responses to compression of the umbilical cord in the short term (3-15 min) and to a greater degree of compression (50-100 % cord compression; Itskovitz et al. 1983a,b; 1987; Rudolph et al. 1989; Iwamoto et al. 1991). The fetal cardiovascular responses to an acute episode of umbilical cord compression include bradycardia, a rise in arterial blood pressure and peripheral vasoconstriction (Giussani et al. 1997) and the magnitude of these responses is related to the severity of the compression (Itskovitz et al. 1983b; Iwamoto et al. 1991). For example, while acute (up to 15 min), mild (25 % reduction from baseline) umbilical cord compression had no effect on fetal arterial blood pressure, heart rate or blood gas status (Itskovitz et al. 1983a), 50-75 % compression of the umbilical cord produced fetal asphyxia, an increase in mean arterial blood pressure, pronounced bradycardia and a redistribution of combined ventricular output towards essential circulations such as the fetal heart and brain (Itskovitz et al. 1983a, 1987; Iwamoto et al. 1991).

In the present study the magnitude of the umbilical cord compression was comparable to the mild-moderate (25-50 %) degree of acute cord compression previously reported by Itskovitz et al. (1983a,b) and Iwamoto et al. (1991). Indeed many of the initial (up to 1 h) fetal cardiovascular and metabolic responses to this degree of cord compression were similar between these studies: fetal asphyxia as indicated by a reduction in fetal P_a,O2 and percentage saturation of haemoglobin with an increase in P_a,CO2, no change in arterial blood pressure or oxygen content or in blood glucose and lactate concentrations. However, it is clear that the duration of compression has an impact on the magnitude of the fetal haemodynamic, endocrine and metabolic responses and that study of only the first 1-5 min of cord compression, as in the studies of Itskovitz et al. (1983a,b, 1987), does not reveal a number of developing fetal responses. For example, while fetal exposure to 5 min of cord compression had no effect on fetal haemoglobin, blood glucose and lactate concentration, P_a,CO2, or pH_a (Itskovitz et al. 1987), when the compression is sustained in a controlled manner, as in the present study, then increases in fetal haemoglobin, blood lactate concentration and P_a,CO2 do occur by 8 h of compression. Furthermore, sustained, partial compression of the umbilical cord led to significant increases in mean arterial blood pressure and femoral vascular resistance and a decrease in femoral blood flow, suggesting that blood flow was being redistributed away from the periphery to the more vital cerebral, myocardial and adrenal circulations (Rudolph, 1984; Giussani et al. 1994b).

The increase in fetal haemoglobin concentration and thus in fetal blood oxygen carrying capacity in the present study, to a large extent, counterbalanced the reduction in the percentage oxygen saturation of haemoglobin such that fetal blood oxygen content did not change significantly from baseline. The rapid increase in fetal haemoglobin concentration after cord compression is an interesting finding, although the mechanism(s) mediating the increase remain uncertain. It is unlikely that the increase reflects increased fetal erythropoiesis and red cell production within the first hour of cord compression and available data suggest that the late-gestation fetus has no significant splenic reservoir of erythrocytes (Brace et al. 1983). One possibility may be haemoconcentration through altered maternal-fetal fluid exchange, due to increased hydrostatic pressure within the placenta, but this cannot be addressed in the present study. Although fetal blood oxygen content was not significantly different from baseline values, the combination of a trend towards a reduction in oxygen content and a significant reduction in femoral blood flow produced a significant reduction in oxygen delivery to the fetal hindlimb in the order of 20-30 %. The reduction in oxygen delivery to the fetal hindlimb may explain the fetal lactacidaemia induced by cord compression, since the fetal hindlimbs have been suggested to represent a major source of lactate production during an imposed stress (Boyle et al. 1992). An increase in blood/tissue glucose availability may further provide substrate to fuel elevated fetal lactate production (Lawrence et al. 1982). However, alternative factors may contribute towards elevated fetal plasma lactate concentration such as reduced hepatic lactate uptake, which has been reported to occur during severe umbilical cord compression (Rudolph et al. 1989), or increased placental production of lactate (Sparks et al. 1982). Current data, however, suggest that the contribution of lactate from the placenta to the circulating pool of fetal lactate is of negligible importance, at least during the type of adverse intrauterine conditions produced by reduced uterine blood flow, since the flux of lactate across the feto-placental unit favours uptake by the placenta rather than by the fetus (Hooper et al. 1995).

The gradual fall in fetal heart rate measured by the end of the present study in control fetuses has been noted previously in many experimental animals including sheep (Boddy et al. 1974; Kitanaka et al. 1989; Unno et al. 2000), horses (Colles et al. 1978; Forhead et al. 2000) and the chick embryo (Pappano, 1977). Although the precise mechanisms mediating the fall in fetal heart rate with advancing gestation are unknown, it is thought to be due, in part, to maturing parasympathetic influences on the fetal heart (Walker et al. 1978; Dawes, 1985). In cord-compressed fetuses there was no fall in heart rate by the end of the experimental period. However, the amplitude of the 24 h rhythm in fetal heart rate was enhanced in cord-compressed fetuses relative to sham controls. The absence of any decline in heart rate combined with greater amplitude of the 24 h chronotrophic rhythm in cord-compressed fetuses may be due to a net increase in sympathetic tone to the fetal heart. Janssen and colleagues (1991) reported that the 24 h rhythm of heart rate in spontaneously hypertensive rats was under sympathetic control.

To address possible mechanisms mediating the cardiovascular and metabolic changes in cord-compressed fetuses, plasma samples were taken during the cord compression period for measurement of plasma noradrenaline and adrenaline concentrations, as changes in plasma catecholamines are known to contribute to positive chronotrophic effects on the fetal heart (Jones et al. 1988), an increase in femoral vascular resistance (Giussani et al. 1993) and glycaemic and lactacidaemic responses to hypoxaemia (Jones, 1980) in the sheep fetus during late gestation. In addition, plasma samples taken during the cord compression period were also analysed for changes in plasma vasopressin concentration, another vasoconstrictor hormone which has previously been reported to aid the redistribution of blood flow away from the periphery during adverse intrauterine conditions (Iwamoto et al. 1979, Perez et al. 1983).

The present study shows that sustained, partial compression of the umbilical cord led to a significant increase in plasma noradrenaline concentration but no change in either plasma adrenaline or vasopressin concentration. This supports an increase in sympathetic tone mediating, at least in part, the cardiovascular and metabolic responses in cord-compressed fetuses. The source of the increased plasma noradrenaline concentrations in cord-compressed fetuses may be increased sympathetic neuronal spill-over and/or increased synthesis and release of noradrenaline from the adrenal gland. The former is the more likely explanation as we have previously reported that cord-compressed fetuses also show a moderate increase in fetal plasma cortisol concentrations (Gardner et al. 2001), and cortisol is known to stimulate phenylethanolamine N-methyltransferase (PNMT) activity in late gestation and thus favour adrenaline, rather than noradrenaline, secretion from the adrenal medulla (Kvetnansky et al. 1995).

The effect on the fetus of the present method for producing prolonged fetal hypoxaemia by umbilical cord compression has interesting parallels with those of a method to induce reduced uterine blood flow (RUBF) developed by Clark et al. (1982). Indeed many of the endocrine responses are similar despite a greater degree of hypoxaemia and acidaemia being achieved during RUBF (Challis et al. 1989; Hooper et al. 1990). Both methods induced sustained increases in fetal plasma noradrenaline and cortisol concentrations and a transient increase in fetal plasma concentrations of ACTH (seeChallis et al. 1989; Hooper et al. 1990; Gardner et al. 2001). However, in contrast to the present study, RUBF induced significant increases in fetal plasma adrenaline and vasopressin concentrations (Hooper et al. 1990) but had no effect on fetal basal arterial blood pressure or heart rate (Bocking et al. 1988).

While this is the first study to produce controlled, reversible compression of the umbilical cord for a prolonged period, there have been a series of studies by Anderson and colleagues which have investigated the fetal responses to compression of the distal aorta in fetal sheep (Anderson & Faber, 1984; Anderson et al. 1986, 1987). Indeed some, but not all, of the cardiovascular and metabolic responses are similar. For example distal aortic compression for a variable period (range 2-14 days) resulted in a 2-3 mmHg reduction in fetal P_a,O2, a reduction in pH_a from 7.37 ± 0.02 to 7.32 ± 0.04 and an increase in [Hb]_a from 8.8 ± 0.2 to 10.8 ± 0.3 g dl⁻¹ (Anderson et al. 1987, 1993). However, arterial blood pressure was reported to either remain unchanged or increase in the upper body (carotid artery) and either transiently or persistently decrease in the lower body (femoral artery; Anderson et al. 1987, 1993). Clearly, compression of the whole umbilical cord versus compression of the fetal distal aorta and specifically supra- versus subrenal distal aortic compression will have differential effects on fetal cardiac preload and afterload, oxygen delivery from the placenta to the fetus, and in the clearance of waste products from the fetal to the placental circulations. Furthermore, compression of the whole umbilical cord in the present study represents a physiological challenge known to occur commonly during gestation (Rayburn et al. 1981; Clapp et al. 1999, 2000).

In conclusion, a method for producing prolonged, measured, reversible reductions in umbilical blood flow in late gestation fetal sheep has been developed. Partial compression of the umbilical cord, leading to a controlled reduction in UBF by 30 % of baseline, produced mild and sustained fetal hypoxaemia, a decrease in oxygen delivery to the fetal hindlimb and an increase in femoral vascular resistance which was associated with a consistent increase in blood lactate concentration. The duration of cord compression was sufficient to produce a significant increase in mean arterial blood pressure and a significant decrease in femoral blood flow, and to prevent the normal decline in heart rate over the period of the study. In addition, umbilical cord compression produced significant elevation in fetal plasma noradrenaline concentration but had no effect on fetal plasma adrenaline or vasopressin levels. The increase in fetal plasma noradrenaline concentration in cord-compressed fetuses may reflect an increase in sympathetic drive which mediates, in part, the fetal cardiovascular and metabolic responses to sustained, reversible adverse intrauterine conditions.