Anesthetic Techniques for Fetal Surgery: Effects of Maternal Anesthesia on Intraoperative Fetal Outcomes in a Sheep Model

WITH advances in prenatal diagnosis and improvements in fetal surgical techniques, fetal surgery has become an accepted alternative treatment option for certain fetal diseases. This was most recently exemplified by the Management of Myelomeningocele Study, a multicenter randomized trial that demonstrated a reduced need to shunt within the first year of life and improvement in short-term neurologic outcomes in patients undergoing prenatal repair of myelomeningocele.¹  During open fetal surgery, general inhalational anesthesia with a high minimum alveolar concentration (MAC) is typically used to promote adequate uterine relaxation for optimal surgical exposure of the fetus and to minimize the risk of placental separation.^2–4  The potent vasodilatory effect of high-concentration inhalational anesthetics (2–3 MAC) results in maternal and fetal hemodynamic instability and decreased uteroplacental blood flow.⁵  These deterioration effects may be very harmful to the fetus undergoing fetal surgery, who is often at higher risk of perioperative morbidity and mortality, because of coexisting disease, decreased cardiopulmonary reserves, and invasive surgical manipulation.

An alternative technique using propofol and remifentanil to supplement inhalational anesthesia has been used for open fetal surgery.⁶  This technique reduces the exposure to desflurane while providing adequate uterine relaxation. In our recent human fetal surgery retrospective outcomes study comparing two different anesthetic techniques, high-dose desflurane (HD-DES) versus supplemental IV anesthesia (SIVA), we observed a higher incidence of fetal cardiac depression (fetal bradycardia and left ventricular systolic dysfunction) and significant maternal hemodynamic instability during exposure to high concentrations of desflurane for long periods.⁶  However, our human fetal anesthesia study had the following limitations: (1) retrospective study, (2) nonrandom assignment of two anesthetic techniques, (3) varied skills and experiences of surgeons and anesthesiologists during the study period, and (4) inadequate time to perform detailed quantitative intraoperative fetal echocardiographic evaluations. To eliminate the above limitations and to evaluate the real effects of both anesthetic techniques on fetal outcomes after open fetal surgery, we designed this study with standardized and controlled experiments and evaluations using a fetal–maternal sheep model. The chronically instrumented fetal–maternal sheep model is a well-established model for fetal–maternal physiologic and hemodynamic studies.^7–9  The first aim of this prospective and controlled sheep study was to compare the maternal and fetal effects of two commonly used fetal anesthetic techniques, HD-DES and SIVA. We hypothesized that compared with HD-DES, the SIVA technique provides more stable maternal–fetal hemodynamic parameters, less fetal acidosis, and less fetal cardiac dysfunction. We specifically compared fetal cardiac function using fetal echocardiography and compared fetal acidosis using blood gas analysis.

Because exposure to high-concentration inhalational anesthetic technique is known to cause maternal hypotension and secondary fetal acidosis from maternal hypotension, the second aim of this study was to demonstrate whether fetal acidosis is a result of HD-DES exposure by itself or is a result of untreated maternal hypotension associated with HD-DES anesthesia. We hypothesized that fetal acidosis persists under HD-DES exposure even when maternal blood pressure is maintained in the normal range.

This maternal–fetal sheep study had the following two phases: (1) randomized, controlled, crossover study and (2) HD-DES anesthesia with aggressive management of maternal hypotension. The first phase of the study was a 2 × 2 or AB/BA randomized, controlled, crossover trial with A = HD-DES and B = SIVA. All animals were instrumented under general anesthesia, followed by a greater than 72-h recovery period, and then exposed to two different anesthetic techniques in random sequences (original set in fig. 1). There was a minimum 40-h washout period between experiments in which ewes were not exposed to any anesthesia. The method of permuted-block randomization was used to assign an equal number of sheep to either the AB or the BA experiment sequence group.

All procedures were carried out under humane care in compliance with the Guide for the Care and Use of Laboratory Animals by the National Academy of Sciences. After approval from the Committee for Animal Care at Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, 26 pregnant Dorset ewes at 105–110 days of gestation (term, 147–150 days) were instrumented. Among these animals, 19 ewes underwent (AB/BA) a randomized, controlled, crossover trial.

In the second phase, an additional seven animals were exposed to HD-DES as a single exposure after recovery from instrumentation (additional set in fig. 1). In this phase, maternal mean arterial pressure (MAP) was tightly maintained within 10% of baseline value throughout the experiment.

Pregnant ewes were fasted overnight and received 2 million units of penicillin G and 100 mg gentamicin 30 min before surgery. The ewes were premedicated with IV diazepam (0.25 mg/kg) and ketamine (10 mg/kg), then intubated and mechanically ventilated with 10.2% end-tidal desflurane and oxygen 1–3 l/min (Aestiva 5/7100; GE-Ohmeda, Helsinki, Finland). Ventilation parameters were adjusted to maintain end-tidal concentration of carbon dioxide (ETco₂) between 35 and 40 mmHg (Datex Capnometric; Datex-Ohmeda, Madison, WI). For analgesia, 0.3 mg of buprenorphine was given intramuscularly before the surgical procedure.

A venous line and an arterial line were placed into the maternal femoral vein and artery for fluid replacement and hemodynamic and blood gas monitoring, respectively. After midline laparotomy and minihysterotomy, the fetal hind limb was exposed. Catheters were placed in the fetal femoral artery and fetal femoral vein. Through the same hysterotomy, the fetus was repositioned and an umbilical flow probe (4–6 mm; Transonic Systems, Inc., Ithaca, NY) was placed around the common umbilical artery and secured in place for monitoring of umbilical blood flow. The fetal incision and the uterus were closed in layers, following replacement of amniotic fluid with warm saline containing 1 g of cefazolin antibiotic. Bilateral uterine arteries were identified, and then uterine flow probes (4–6 mm) were placed and secured on each vessel. All catheters and probes were tunneled to the flank of the maternal ewe and stored in a pouch sewn to the skin. The maternal abdominal wall was closed.

A Lap Disc Hand Access Device (Hakko Co., Tokyo, Japan) was implanted onto the contralateral side of the maternal abdominal wall to the instrument pouch through a paramedian incision to allow multiple intraabdominal exposures for transuterine fetal echocardiography without surgery. The maternal abdomen was automatically closed by twisting the outer part of this device, and locked. Additional sterile tape and abdominal support were applied over the device to ensure integrity.

Randomized, Controlled, Crossover Study Phase. Pregnant ewes were fasted overnight and received 2 million units of penicillin G, 100 mg of gentamicin, and 0.15 mg of buprenorphine intramuscularly. Baseline maternal–fetal hemodynamics and blood gases were measured before anesthesia to ensure that all animals were healthy for experiments. General anesthesia was then induced by 3 mg/kg propofol and 1.5 mg/kg succinylcholine via a maternal femoral venous line, followed by tracheal intubation and mechanical ventilation with 100% oxygen. Ventilation was adjusted to maintain ETco₂ between 30 and 40 mmHg. To minimize the confounder effect of inotrope and vasopressor on uterine blood flow and fetal hemodynamic status, maternal hypotension was treated by IV ephedrine only to maintain maternal MAP above 60 mmHg. Animals were then randomized to one of the two anesthetic protocols described below. After the first anesthetic experiment, the ewe recovered. There was a minimum 40-h washout interval between the two anesthetic experiments. Following the washout interval, for the second experiment, animals received the other/second anesthesia regimen and then they were euthanized. With the HD-DES anesthesia technique, the animal was exposed to 1.5 MAC of desflurane (10.2% end-tidal concentration) for the initial 60 min. Then, desflurane concentration was increased to 2.5 MAC (18% end-tidal concentration) for an additional 90 min to simulate conditions during midgestation or ex utero intrapartum human fetal surgery.⁶ 

With the lower dose desflurane anesthesia with SIVA technique, the animal was anesthetized with IV propofol (450 μg·kg⁻¹·min⁻¹) and remifentanil (0.5 μg·kg⁻¹·min⁻¹) for the first 60 min, followed by 1.5 MAC of desflurane (10.2% end-tidal concentration), propofol (75 μg·kg⁻¹·min⁻¹), and remifentanil (0.25 μg·kg⁻¹·min⁻¹).

Because anesthetic requirement is species specific, we used maternal electroencephalography to ensure comparable depth of anesthesia with both anesthetic techniques. Subcutaneous electrodes were placed over the frontal and parietal parts of the maternal head. The electrical signals were amplified, digitized, and recorded using a Nicolet One electroencephalography system (Viasys Healthcare, Madison, WI). Four animals were used to identify the above anesthetic regimens by continuously monitored maternal electroencephalography. Because MAC of desflurane in nonpregnant ewes is 9.5%,¹⁰  we estimated MAC of pregnant ewes at 70% that of nonpregnant ewes, or 6.8%. Burst suppression on maternal electroencephalography was used to guide comparable doses of desflurane, propofol, and remifentanil in the SIVA and desflurane in HD-DES anesthetic techniques by titrating the level of anesthesia until burst suppression developed on electroencephalography (fig. 2). Burst suppression has been used as an endpoint for measuring depth of anesthesia in sheep.¹¹  In humans, it is a stage of very deep anesthesia¹²  comparable to clinically validated bispectral index scores of approximately 30.¹³ 

HD-DES with Tightly Controlled Maternal Blood Pressure Phase. Pregnant ewes were fasted overnight and received 2 million units of penicillin G, 100 mg of gentamicin, and 0.15 mg of buprenorphine intramuscularly. Baseline maternal–fetal hemodynamics and blood gases were measured before anesthesia to ensure that all animals were healthy for the experiment. General anesthesia was then induced by 3 mg/kg propofol and 1.5 mg/kg succinylcholine via maternal femoral venous line, followed by tracheal intubation and mechanical ventilation with 100% oxygen. Ventilation was adjusted to maintain ETco2 between 30 and 40 mmHg. Each animal in this group was exposed to 1.5 MAC of desflurane (10.2% end-tidal concentration) for the initial 60 min. Then, desflurane concentration was increased to 2.5 MAC (18% end-tidal concentration) for an additional 90 min as in HD-DES technique. Maternal MAP was maintained within 10% of baseline. Ephedrine (5 mg, IV bolus every 2–3 min as needed) was the first-line treatment drug given to the ewe to maintain blood pressure. Phenylephrine (20 μg, IV bolus every 2–3 min as needed) was given if the maternal heart rate was higher than 180 beats/min or when there was inadequate response to ephedrine.

Hemodynamic Parameters. All hemodynamic parameters were recorded using a PowerLab data acquisition system (AD Instruments, Colorado Springs, CO). Maternal arterial blood pressure, maternal heart rate, fetal arterial blood pressure, fetal heart rate, umbilical blood flow, and bilateral uterine blood flows were recorded at baseline, and then continuously after induction of anesthesia until the end of the experiments.

Arterial Blood Gases and Lactate Analysis. Maternal and fetal arterial blood was drawn simultaneously at baseline, then every 30 min after induction of anesthesia until the end of the experiment (at 30, 60, 90, 120, and 150 min). The arterial blood gases (pH, arterial oxygen tension [Pao₂], arterial carbon dioxide tension [Paco₂], HCO₃⁻,and oxygen saturation) were analyzed by iSTAT (Abbott Labs, East Windsor, NJ). Lactate was measured by a YSI 2300-STAT analyzer (YSI Corp., Yellow Springs, OH).

Quantitative Fetal Echocardiography. Transabdominal fetal echocardiography was performed at baseline. Serial transuterine fetal echocardiography was also performed every 30 min after induction of anesthesia at 30, 60, 90, 120, and 150 min with the transducer placed through the Hand Access Device. By using two-dimensional and M-mode imaging, left and right ventricular transverse dimensions at end-diastole diameter and end-systole diameter were measured. The ventricular shortening fraction was calculated for left and right ventricular chamber. Using color flow Doppler examination, the degree of atrioventricular valve insufficiency was assessed qualitatively at the mitral and tricuspid valves. All video records of the fetal echocardiography were reviewed by a single experienced pediatric cardiologist (E.C.M.) who was blinded to the experimental procedure.

Randomized, Controlled, Crossover Phase. Statistical analysis was performed using SAS software (version 9.2; SAS Institute, Inc., Cary, NC). Linear mixed effect models with Kenward-Roger correction for small samples were used to analyze outcomes measured at different time points.¹⁴  For some small-sample repeated measures problems, estimates of precision and inference for fixed effects based on asymptotic distribution are known to be inadequate. We applied Kenward-Roger correction, which uses an adjusted estimator of the covariance matrix of the fixed and random effects and is shown to perform well in small sample settings.¹⁴  The models included treatment (anesthetic technique) and period as fixed effects; sheep as a random effect; and the corresponding baseline measurement, hematocrit, and ephedrine treatment for maternal hypotension as covariates. Treatment-specific least-squares means ± SE were reported for outcomes of interest. The differences of least-squares means between treatments were compared with two-tailed tests with a significance level of 0.05. We expected no carryover effect with a washout period of at least 40 h between two consecutive anesthetic experiments carried out on the same ewe. To validate this assumption, we compared baseline measures at the beginning of two periods using paired Student t tests and linear mixed effect models.

HD-DES with Tightly Controlled Maternal Blood Pressure Phase. Data from the animals under a tightly controlled maternal blood pressure protocol were compared to data from animal exposed to HD-DES in the first-phase experiments in the crossover trial while adjusting for corresponding baseline measurement. Group-specific (with and without tight control for maternal blood pressure) least-squares means ± SE were reported for outcomes of interest. The differences of least-squares means between the groups were compared with two-tailed tests with a significance level of 0.05.

The data from the randomized controlled crossover and the tightly controlled maternal blood pressure phases were combined and analyzed using mixed effect models to examine the overall treatment effect (HD-DES vs. SIVA) on maternal and fetal outcomes. A binary variable indicating the type of trial was included in the model in addition to the covariates considered in the analysis of the randomized, controlled, crossover trial.

We instrumented a total of 26 midgestational ewes. Four animals were used to identify comparable anesthetic regimens, and five animals did not complete the experimental protocol because of technical failure and/or fetal demise in utero. Data from 11 animals that underwent the randomized crossover phase and an additional six animals that underwent the tightly controlled maternal blood pressure phase with HD-DES exposure were analyzed.

All animals that underwent the crossover trial received both anesthetic regimens in random sequences. The SIVA technique was the first experiment in five animals and HD-DES was the first experiment in six animals. There were no statistically significant differences in baseline hemodynamics, arterial blood gases, or fetal echocardiography parameters between the first experiment and the second experiment.

Maternal MAP during HD-DES anesthesia was statistical significantly lower than SIVA anesthesia throughout the study period, with the mean difference being 19.53 mmHg (95% CI, 17.6–21.4; P < 0.0001) (fig. 3). Profound maternal hypotension occurred after exposure to 2.5 MAC of desflurane (after 60 min) and persisted throughout the study period with HD-DES. Maternal normotension was maintained in the first 60 min in the SIVA technique, before desflurane exposure. However, after adding 1.5 MAC of desflurane after the first 60 min in SIVA technique, maternal hypotension occurred. During the entire study period, the ewes undergoing the HD-DES technique required a higher dose of ephedrine for treatment of hypotension than those undergoing the SIVA technique (44 vs. 9 mg, P = 0.004). There were no statistically significant differences in maternal or fetal heart rate, or fetal MAP between anesthetic techniques.

Uterine blood flow was lower with the HD-DES technique compared with the SIVA technique (fig. 3). Uterine blood flow decreased in the same direction as maternal MAP, with the mean difference being 78.91 mmHg (95% CI, 67.43–97.39; P < 0.001). With the HD-DES technique, uterine blood flow decreased within the first 15 min after anesthesia, and decreased further after increasing desflurane to 2.5 MAC. In contrast, with the SIVA technique, uterine blood flow was stable until 1.5 MAC of desflurane began at 60 min of the experiment; then, uterine blood flow decreased with desflurane exposure. Umbilical blood flow declined from baseline in all animals. There was no statistically significant difference in umbilical blood flow between techniques (fig. 3).

Fetuses with HD-DES exposure (n = 2 [13.3%]) needed more resuscitation than those with SIVA exposure (n = 0 [0%]). Although this difference was not statistically significant because of our relatively small sample size (P = 0.24), this difference (13.3% vs. 0%) between both anesthetic technique is clinically important. Bradycardia and cardiac arrest in these two fetuses occurred after 2 h of exposure to HD-DES. The maternal–fetal hemodynamic measurement, blood gas measurement, and echocardiographic measurement after 120 min from these two animals were excluded from this study because of the resuscitative interventions.

With HD-DES exposure, maternal Pao₂ was statistically lower (Pao₂, 278 vs. 343 mmHg at 150 min; P = 0.025) and maternal Paco₂ was higher (Paco₂, 48 vs. 43 mmHg at 150 min; P = 0.046) than with SIVA exposure. Although these differences have attained marginal statistical significance, the differences of Pao₂ and Paco₂ values are clinically not significant because they do not convey any associated clinical concerns of hypoxia and/or hypercarbia requiring interventions. There were no statistically significant differences in maternal pH, base deficit, or lactate between the two anesthesia techniques.

Fetuses exposed to the HD-DES technique developed more acidosis than fetuses exposed to the SIVA technique. The fetal pH at 150 min after exposure to HD-DES was significantly lower than after SIVA exposure (pH 7.11 vs. 7.24 at 150 min; P < 0.001). The fetal Paco₂ and lactate were significantly higher with HD-DES than with SIVA exposure statistically, and these values increased over time with worsening acidosis. There was no significant difference in fetal Pao₂ between the two anesthetic techniques. Fetal Pao₂ increased only slightly despite enormous increases in maternal Pao₂ after anesthesia with 100% oxygen. The details of fetal–maternal blood gas and lactate under two different anesthetic exposures at different time points are demonstrated in tables 1 and 2.

Because of technical difficulties, quality transuterine fetal echocardiographic data acquisition was not feasible in all animals. Fetal echocardiographic data were obtained from nine animals undergoing the SIVA technique and eight animals undergoing the HD-DES technique. There were no statistically significant differences in fetal left ventricular function and right ventricular function as demonstrated by shortening fraction between the two anesthesia techniques (table 2). One fetus undergoing HD-DES and two fetuses undergoing SIVA exposure developed mitral regurgitation. Five fetuses with both HD-DES and SIVA exposure developed tricuspid regurgitation.

Six animals that underwent the tightly control maternal blood pressure trial with HD-DES exposure were compared to six animals exposed to HD-DES in the first crossover phase experiments without tightly maternal blood pressure control. An average dose of 120 mg of ephedrine and 317 μg of phenylephrine were given during a 150-min experiment to maintain maternal normotension under a tightly control maternal blood pressure trial.

Maternal MAP under tightly controlled blood pressure protocol during HD-DES anesthesia was higher than maternal mean arterial pressure without control for blood pressure as demonstrated in figure 3. Even with well-maintained maternal normotension, uterine blood flow decreased from baseline. There was no statistically significant difference in uterine blood flow under HD-DES exposure with tightly controlled maternal blood pressure compared with HD-DES exposure without aggressive control for blood pressure (fig. 3). There was no difference in umbilical blood flow between these two groups.

The animals under the tightly controlled maternal blood pressure protocol during HD-DES exposure had lower maternal pH (pH 7.34 vs. 7.42 at 150 min; P = 0.037), higher maternal Pao₂ (Pao₂, 400 vs. 248 mmHg at 150 min; P = 0.004), and higher base excess (base excess, −0.2 vs. 6.8 mm at 150 min; P = 0.035) than the animals without control for maternal blood pressure. Although these differences are statistically significant, they are not clinically concerning, as the values are within acceptable and physiologic ranges and did not require any interventions.

There were no statistically significant differences in fetal blood gas analysis between fetuses under tight control for maternal normotension in comparison with those without control of maternal blood pressure during HD-DES exposure. Fetuses in both groups developed acidosis over time. However, the extent of the increase in Paco₂ and the decrease in pH in fetuses without treatment for maternal hypotension was greater than that of the fetuses with tight maternal blood pressure control. The maternal and fetal blood gas analyses of animals under tight blood pressure control and without aggressive control for maternal blood pressure are demonstrated in tables 3 and 4.

There were no statistically significant differences in fetal left ventricular function and right ventricular function as demonstrated by shortening fraction between animals under the tightly controlled maternal blood pressure protocol during HD-DES exposure and animals without control for maternal blood pressure (table 4).

The fetal outcomes under the tightly controlled blood pressure protocol during HD-DES anesthesia were included in the outcomes under HD-DES exposure and then compared with SIVA. With HD-DES exposure (without and with tight maternal blood pressure control), the fetuses developed significantly more acidosis (pH 7.09 vs. 7.26; P = 0.002) and higher Paco₂ (83 vs. 60 mmHg, P = 0.015) than the fetuses exposed to SIVA technique at 150 min. There were no statistically significant differences in left and right ventricular shortening fraction between fetuses exposed to two anesthetic techniques (table 5).

Our study demonstrated that, in sheep, the SIVA anesthetic technique preserved stable maternal hemodynamics and fetal acid base status compared with the HD-DES anesthetic technique. By shortening the duration of desflurane exposure and minimizing the concentration of desflurane, the SIVA technique results in relatively stable maternal–fetal hemodynamics, uterine blood flow, and fetal acid–base physiologic status. Fetal acidosis persisted with HD-DES anesthetic technique even with tight control of maternal blood pressure compared with the SIVA technique.

Anesthesia for open fetal surgery is very different from anesthesia for cesarean delivery.²  Minimum exposure to 0.5–1 MAC inhalation anesthesia is recommended for cesarean delivery to maintain stable maternal–fetal hemodynamics and to ensure prompt hypertonic response of the uterus after delivery. For open fetal surgery, high-dose inhalational anesthesia (up to 2.5 MAC) is traditionally used to provide maximum uterine relaxation for better surgical exposure with access to the fetus and to decrease the risk of placental separation during fetal surgery. All potent inhalational anesthetic agents cross the placenta into the fetus. Among these agents, desflurane is preferable for open fetal surgery because of its low blood gas partition coefficient.¹⁵  This results in rapid placental transfer, faster change in tissue concentrations, easy titrability, and shorter duration of fetal exposure.

The high-dose inhalation anesthetic technique has some disadvantages. First, high-dose inhalational anesthesia is known to induce hemodynamic instability. Maternal hypotension results in decreased uterine blood flow, which impairs fetal perfusion and leads to fetal hypercarbia and acidosis.¹⁶  Second, high-dose inhalational anesthesia may result in fetal cardiac depression. The fetal anesthetic requirement is much lower than the maternal requirement. For instance, the MAC of halothane in the fetal lamb is approximately 50% that of maternal MAC (0.33% vs. 0.69%). However, maternal anesthesia with 1 MAC of halothane results in an end-tidal concentration of 0.45% in the fetus.¹⁷  Lower anesthetic requirement in the fetus in combination with high-dose exposure and rapid placental transfer of inhalational agents may contribute to direct fetal cardiac depressive effects. An alternative technique using supplemental IV anesthesia with propofol and remifentanil infusions and a lower dose (1–1.5 MAC) of desflurane (SIVA technique) is expected to minimize maternal hemodynamic instability and fetal cardiac depression, in comparison with traditional high-dose (2.5 MAC) desflurane anesthesia. Propofol crosses the placenta and induces vasodilation in the placenta without reducing placental blood flow.¹⁸  Remifentanil, an ultra–short-acting opioid, crosses the placenta and is rapidly metabolized without adverse neonatal or maternal effects.¹⁹ 

Previous studies in both midgestational and near-term pregnant ewes showed that high-dose (1.5–2 MAC) halothane, isoflurane, and sevoflurane decreased maternal blood pressure, cardiac output, and uterine blood flow.^5,20–22  This high concentration of inhalational agents also induced reversible fetal hypoxemia and acidosis because of poor uteroplacental perfusion. In our midgestational sheep model, desflurane decreased maternal blood pressure, decreased uterine blood flow, and induced fetal acidosis and fetal cardiac arrest in a dose-dependent fashion.

When we compared the effects of SIVA and HD-DES, we designed our protocol to maintain maternal MAP above 60 mmHg to minimize the use of vasoactive agents, the confounder on uteroplacental blood flow. In clinical practice, vasoactive drugs such as ephedrine and phenylephrine are promptly given to treat maternal hypotension during fetal surgery and cesarean delivery. We designed the additional set of experiment using the tightly controlled maternal blood pressure protocol to mimic fetal and obstetric anesthesia clinical practice. We chose ephedrine as a first-line drug to treat maternal hypotension because ephedrine has been shown to increase,²³  or not change,²⁴  uterine blood flow; in contrast, phenylephrine is known to reduce uterine blood flow.²⁵  Phenylephrine is often used when high doses of ephedrine become ineffective and cause significant tachycardia. We found that a small dose of ephedrine as given to HD-DES during the crossover design study had little effect on uterine blood flow. However, when large doses were used to maintain maternal blood pressure in the HD-DES with tightly controlled maternal blood pressure protocol, the increase in uterine resistance was more pronounced than the increase in maternal blood pressure, and uterine blood flow still declined.²⁶  As a result, fetal acidosis still occurred with better and nearly normal maternal blood pressures.

The decrease in fetal MAP and umbilical blood flow in both anesthetic regimens was minimal, even with substantially diminished maternal blood pressure and uterine blood flow, depicting the tremendous fetal reserve capacity for oxygen and nutrients. Our findings are consistent with the immaturity or the absence of α-receptors in the fetal and umbilical–placental vascular bed.²⁶ 

Besides uterine blood flow, maternal Paco₂ also affect respiratory gas exchange in the placenta.²⁷  In our sheep model, fetal hypoperfusion and accumulation of fetal Paco₂ were reflected by increased maternal Paco₂. Although ETco₂ was fixed in our protocol, we observed a progressive disparity between ETco₂ and maternal Paco₂. This finding was more profound in the HD-DES group than in the SIVA group. Uemura et al. found the same phenomenon in midgestational sheep anesthetized with 1.5 MAC of isoflurane.^16,22  Because ETco₂ inaccurately estimates maternal Paco₂, arterial blood gas monitoring is recommended in addition to standard monitoring during invasive surgical procedure in pregnant women.²⁸ 

Higher maternal Pao₂ did not increase fetal Pao₂ in our study, a phenomenon consistently observed in other fetal sheep studies.^7–9  We found no difference in fetal Pao₂ during exposure to two different anesthetic techniques, or in baseline Pao₂ values after approximately 40 h between HD-DES and SIVA techniques, suggesting no carryover effect. However, Jonker et al. noted a decrease in fetal Pao₂ and oxygen content at 5 days after exposure to higher dose halothane in comparison with supplemental IV diazepam and ketamine with lower dose halothane,²⁹  suggesting a detrimental long-term effect of high-dose inhalational anesthetic agents and the potential benefits of IV anesthetic drugs.

Fetal echocardiography is the standard for intermittent fetal cardiac monitoring during open fetal surgery. However, the image quality and the reliability of data depend on multiple factors, including maternal adiposity and fetal alignment.³⁰  The time for performing fetal echocardiography during human fetal surgery is also limited to minimize fetal surgical time. Even with more time for transuterine fetal echocardiography in our sheep model, a good acoustic view for data acquisition was not feasible in all animals. Fetuses with HD-DES exposure developed severe acidosis over time, and two fetuses developed significant bradycardia and required resuscitation. However, we found no statistically significant differences in left ventricular shortening fraction, right ventricular shortening fraction, and atrioventricular valve regurgitation between two anesthetic regimens. The inability to observe statistically significant echocardiographic differences might be attributable to (1) our selection of specific cardiac echocardiographic parameters or monitoring frequency, (2) relatively better tolerance of hypoxia and acute acidosis by fetal sheep myocardium,³¹  and (3) our small sample size and echocardiographic data insufficient to detect the difference. Subtle differences in fetal cardiac status might be detectable only with more sensitive technologies such as tissue Doppler imaging (which was not available to us during this study) and/or increased sample size. However, it is also possible that the echocardiographic techniques used in this study were not sensitive enough to identify early and minimal myocardial depression in the fetus. Additional studies are needed to evaluate more quantitative and sensitive assessment to measure and identify fetal cardiac deterioration during open fetal surgery.

The fetal–maternal sheep model is an established model for fetal surgery and maternal fetal hemodynamic study. It is not an ideal model with which to study uterine relaxation because the sheep uterus is much thinner than human uterus. In our retrospective human study, we previously demonstrated that the SIVA technique provides adequate uterine relaxation with a lower rate of fetal resuscitative interventions compare with the the HD-DES technique.⁶  In this study, we use a prospective, experimental animal model to eliminate the limitations associated with the retrospective design of our previous clinical study.

By using a sheep model, we were able to test the same animal with both anesthetics in a robust randomized crossover design; each fetal–maternal sheep was exposed to both HD-DES and SIVA regimens in random sequences. Each animal served as its own control, providing greater statistical power to analysis with a smaller study sample size. This type of study requires an adequate washout interval between the two anesthetic exposures to prevent a carryover effect. Previous studies have shown that fetal acidosis from inhalational anesthetic agents or impaired uterine blood flow was reversible within 2–24 h.^16,20  With more than a 40-h washout interval, we found no differences between the first and the second experiments in baseline hemodynamic and physiologic parameters, including fetal hematocrit. However, two fetal cardiac arrests with HD-DES exposure occurred during the second experiment. This could be attributable to hypovolemia or an undetected carryover effect of the first anesthetic exposure on fetal cardiac function.

Anesthetic requirement is species specific. Because the MAC of desflurane in nonpregnant ewes is 9.5%,¹⁰  we estimated the MAC of pregnant ewes at 70% that of nonpregnant ewes, or 6.8%. However, the ewes are insensitive to IV anesthetic agents and require a higher dose to provide equivalent general anesthesia with inhalation anesthetics. Andaluz et al. demonstrated general anesthesia in pregnant ewes by using a 400-μg·kg⁻¹·min⁻¹ propofol infusion as the sole anesthetic protocol.³²  The combination effects of propofol and remifentanil, or propofol, remifentanil, and desflurane, have never been studied in a maternal–fetal sheep model. To ensure comparable anesthetic regimens of SIVA and HD-DES techniques, we titrated the level of anesthesia to achieve burst suppression on the maternal electroencephalogram.

In conclusion, in sheep, the HD-DES anesthesia technique compared with the SIVA technique caused significant maternal hypotension and reduction in uterine blood flow, which resulted in significant fetal acidosis. When maternal blood pressure was maintained in the normal range under HD-DES anesthesia by using vasopressors, uterine blood flow still declined and fetal acidosis persisted. Our study in sheep suggests that prospective comparisons of these methods in humans are warranted. A more sensitive fetal echocardiographic techniques and/or a larger sample may be needed to assess fetal cardiac function and to promptly detect fetal deterioration during fetal surgery.

The authors thank John McCann, B.S. (Senior Research Assistant, Department of Anesthesiology, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio), and Scott Baker, B.S. (Research Associate, Division of Cardiothoracic Surgery, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio), for their technical assistance and instrumentation; Regina Keller, A.S., and Susan Gomien, B.S. (Fetal Sonographers, Division of Cardiology, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio), for performing fetal echocardiography; Ethicon Endo-Surgery, Inc. (Cincinnati, Ohio), for donation of Hand Access Devices; and the Veterinary Services at Cincinnati Children’s Hospital Medical Center (Cincinnati, Ohio) for taking great care of the study animals.