Bilateral Hydronephrosis–Obstructive Uropathy

Congenital obstructive uropathy occurs in approximately 0.1% of pregnancies.¹² Congenital bilateral hydronephrosis results from fetal urethral obstruction at the bladder outlet, most often by posterior urethral valves in male fetuses or urethral obstruction in females. Other causes of fetal obstructive uropathy include obstruction at the ureteropelvic or vesicoureteric junction and a number of complex disorders in females (e.g., cloacal plate anomalies). These uropathies are easily detected by ultrasonography, which is often performed to investigate oligohydramnios from diminished fetal urine output. Severe obstructive lesions may lead to progressive renal dysplasia and dysfunction, bladder distention, and oligohydramnios and ultimately result in devastating developmental consequences, such as limb and facial deformities and pulmonary hypoplasia (Figure 7-1).¹³ Preterm delivery allows early urinary tract decompression ex utero, but fetal pulmonary immaturity limits the efficacy of this approach. Early intrauterine intervention with placement of a vesicoamniotic shunt allows drainage of urine from the fetal bladder into the amniotic cavity, thereby decompressing the urinary tract. In animal models, in utero relief of obstructive uropathy improves dysplastic renal histology, restores normal urine flow and amniotic fluid volume, and results in improved lung growth and development.¹⁴ The applicability of these findings to human fetal obstructive uropathy remains unclear and controversial.¹⁵

Vesicoamniotic catheter shunts have been used for intrauterine treatment of bilateral hydronephrosis since the early 1980s.¹⁶ These valveless, double-coiled catheters are placed percutaneously with ultrasonographic guidance, with one coil being left in the urinary bladder and the other in the amniotic space. Common problems associated with these catheters include (1) difficult placement, occlusion, and displacement; (2) fetal trauma, iatrogenic abdominal wall defects, and amnioperitoneal leaking; and (3) preterm PROM, preterm labor, and infection.¹⁷ Neonatal survival rates after fetal vesicoamniotic shunting (performed from the 1980s to 2001) varied from 50% to 90%, with approximately half of the survivors having normal renal function.^15,18,19 Currently a multicenter, randomized controlled trial is underway comparing the perinatal mortality and renal function of fetuses with lower urinary tract obstruction treated by either vesicoamniotic shunting or conservative noninterventional care.²⁰

Fetal cystoscopy is a more recent treatment option that allows direct visualization of the fetal urethra. Although not a viable treatment for urethral atresia, fetal cystoscopy facilitates diagnosis and treatment of lower urinary tract obstruction due to posterior urethral valves.²¹ Fetal cystoscopy with ablation of posterior urethral valves appears to provide a survival advantage over conservative therapy but has not been demonstrated to improve perinatal survival over vesicoamniotic shunting.²²

Current evidence supports fetal surgery for the correction of obstructive uropathy in selected fetuses in an effort to restore amniotic fluid volume, prevent pulmonary hypoplasia, and decrease perinatal mortality. However, the effects on long-term renal function and other morbidities remain unclear, and additional evidence is needed.

Congenital Diaphragmatic Hernia

Approximately 1 of 2500 live newborns has a congenital diaphragmatic hernia (CDH).²³ Without fetal intervention, this anomaly causes significant mortality from pulmonary hypoplasia and insufficiency. Survival rates have improved to between 60% and 85% over the past 20 years^24–27 and are closely associated with the degree of pulmonary hypertension and dysfunction.²⁴ Significant mortality occurs despite optimal postnatal surgical management at tertiary care medical centers (i.e., procedures involving removal of the herniated viscera from the chest, administration of surfactant, ventilation techniques that minimize lung trauma, use of ECMO, and closure of the diaphragm). Intrauterine correction of CDH has the potential to prevent the development of pulmonary hypoplasia and allow the fetal lung to develop before delivery.

The use of a fetal lamb model demonstrates that parenchymal hypoplasia and associated pulmonary vascular changes can be reversed by correction in utero.³ Primary repairs of human CDH in utero have been undertaken only for fetuses with severe disease, with limited success but many lessons learned, including the development of minimally invasive approaches.^28,29

Fetal lungs contribute to amniotic fluid volume by secreting more than 100 mL/kg/day of fluid that exits the trachea and mouth. Tracheal occlusion impedes the normal egress of fetal lung fluid and results in expansion of the hypoplastic lung, thereby inducing lung growth and cellular maturation in fetuses with CDH.^30,31 This occlusion technique, termed “plug the lung until it grows” (i.e., PLUG),^32,33 replaced primary repair in utero for the correction of the pulmonary hypoplasia associated with CDH. It is a less extensive, palliative fetal surgical procedure that enhances lung growth to improve postnatal survival, with postponement of the definitive repair until after birth.^30,34 Once the trachea is occluded, fetal pulmonary fluid slowly accumulates and expands the lung, pushing the viscera out of the thorax. A small detachable balloon for endoluminal tracheal occlusion is placed in the trachea via percutaneous endoscopic endotracheal intubation and is either left in place until delivery or deflated and removed earlier (Figure 7-2).^35,36

A prospective randomized trial (1999-2001) evaluated fetal tracheal occlusion for intrauterine treatment of severe CDH.³⁷ Inclusion criteria included (1) a gestational age of 22 to 28 weeks, (2) left-sided herniation of the liver into the hemithorax, and (3) a low lung-to-head ratio (LHR) (i.e., < 1.4). The LHR is a ratio of the contralateral lung size compared with head circumference and is correlated with the severity of pulmonary hypoplasia and survival for a given gestational age.³⁸ The trial was closed early (n = 11); fetal tracheal occlusion resulted in no improvement in survival compared with control (77% versus 73%) and no reduction in morbidity at 90 days. The rates of preterm PROM and preterm delivery were higher in the fetal intervention group.³⁷ However, the survival rate was unexpectedly high in the control group. It is possible that the LHR criterion of less than 1.4 was not sufficiently restrictive and allowed inclusion of fetuses in the study that were likely to survive with standard postnatal tertiary medical care. Table 7-1 notes outcomes of left-sided CDH fetuses treated in utero or with standard postnatal care.

More recently in Europe, the Fetal Endoscopic Tracheal Occlusion (FETO) Task Force began a collaboration among three medical centers for treatment of severe cases of CDH with a high risk for death.³⁹ FETO intervention criteria for fetuses at high risk included both LHR less than 1.0 and liver herniation into the hemithorax.⁴⁰ Use of smaller-gauge endoscopes and reversal of the tracheal occlusion before birth appear to show great promise for reduction in the risk for preterm delivery due to preterm PROM.^41,42 Owing to concern for tracheal damage by very early tracheal balloon placement,⁴³ the tracheal balloon is placed between 26 and 28 weeks’ gestation and removed before birth by a second fetoscopic procedure near 34 weeks (if the fetus is still in utero).⁴⁴ This second procedure is performed to minimize the risk of preterm labor, avoid the need for the EXIT procedure, and potentially improve lung growth and minimize the reduction of type II alveolar cells associated with prolonged tracheal occlusion. Outcomes for 210 cases (through 2008) of fetuses with a mean gestational age of 27 weeks, LHR less than 1.0, and primarily left-sided CDH (84%) were compared with those for historic postnatal treatment controls (1995-2004). Use of FETO significantly improved the survival rate (47% versus 20%), and delivery occurred at a median gestational age of 35.3 weeks.³⁹ However, the comparative results may represent selection bias or improvement in technique and clinical care over time. A more recent (2008-2010) randomized, controlled, single-institution trial compared cases of severe CDH (LHR < 1.0 with liver herniation) randomized to either FETO (n = 21) at 26 to 30 weeks’ gestation or standard postnatal care (n = 20).⁴⁵ The overall survival rate with severe CDH was significantly greater in the FETO intervention group than in the expectant management group (52.6% versus 5.3%). In 2009, a randomized Tracheal Occlusion To Accelerate Lung growth trial (TOTAL) was started.⁴⁶ It compared postnatal management to late (30 to 32 weeks’ gestation) FETO intervention for moderate lung hypoplasia and earlier FETO intervention (27 to 30 weeks’ gestation) for severe lung hypoplasia. In addition, it is now understood that LHR depends on gestational age⁴⁷ and that a ratio of observed to expected LHR is a better expression of CDH severity and likelihood of survival.^48,49 This ratio is used as part of the ongoing TOTAL trial. Results of this trial will help determine if and when FETO should be offered for cases of severe CDH.

Congenital Pulmonary Airway Malformation

Congenital pulmonary airway malformations (CPAM) are pulmonary tumors with cystic and solid components; these malformations were previously described as congenital cystic adenomatoid malformations (CCAM).⁵⁰ The incidence is approximately 1 in 25,000 pregnancies.⁵¹ The classification scheme for CPAM includes five subtypes, based on cyst size, characteristics of the epithelial lining, cyst wall thickness, and the presence of mucous cells, cartilage, and skeletal muscle.^50,52 Lesions are assessed by ultrasonography and described as containing cysts larger (macrocystic) and smaller (microcystic) than 5 mm in diameter. Lesions can either regress with minimal associated morbidity or progressively enlarge, often resulting in hydrops fetalis (fetal heart failure). Small lesions detected in utero or in the newborn are treated after birth by surgical excision of the affected pulmonary lobe. Large lesions can cause mediastinal shift, hydrops, and pulmonary hypoplasia and can interfere with fetal or neonatal survival; fetuses with untreated lesions associated with hydrops fetalis have a survival rate of less than 5%.⁵³ In a retrospective single-institution review of 71 cases, the initial antenatal ultrasonographic ratio of CPAM volume to fetal head circumference (CVR) was evaluated for the formation of hydrops fetalis and postnatal outcomes.⁵⁴ Fetuses with a CVR less than 0.56 were noted to have no adverse postnatal outcomes, whereas a CVR greater than 0.56 had a positive predictive value for adverse postnatal outcome of 33%. In addition, a CVR greater than 1.6 was associated with a greater risk for hydrops fetalis.⁵⁵

Depending on lesion size, location, and characteristics, CPAMs can be managed with either fetal intervention or postnatal resection. Macrocystic lesions have been decompressed in utero by placement of shunt catheters between the cysts and the amniotic cavity, resulting in sustained decompression and resolution of hydrops⁵⁶; these procedures are followed by postnatal surgery. However, not all lesions can be decompressed successfully because the cysts are not always contiguous (i.e., in communication with each other) and can refill rapidly. In addition, these thoracoamniotic shunts have associated risks, including malfunction, displacement, fetal hemorrhage, and chorioamnionitis.⁵⁷ CPAMs inappropriate for drainage can be resected with open fetal surgery. Fetal pulmonary lobectomy for lesions associated with hydrops fetalis has resulted in a 30-day postnatal survival rate of 50%, with tumor resection allowing for compensatory lung growth and resolution of hydrops fetalis.⁵⁸ Maternal administration of betamethasone also has been noted to improve hydrops fetalis and overall outcome in selected fetuses with CPAM.^59,60 A retrospective review of 24 fetuses with predominantly microcystic CPAM and the presence of hydrops fetalis found that corticosteroid treatment resulted in better survival than resection with open fetal surgery.⁶¹

Intralobar and extralobar pulmonary sequestrations (bronchopulmonary sequestrations) are rarer congenital lung anomalies than CPAM and involve nonfunctional lung tissue (disconnected from the bronchial tree). As with CPAM, therapeutic options depend on fetal morbidities including hydrops fetalis and pulmonary hypoplasia. Thoracoamniotic shunts have been successfully placed to decompress massive congenital pleural effusions caused by fetal chylothorax that otherwise result in hydrops fetalis, pulmonary compression, and fetal or neonatal death.⁶²

Sacrococcygeal Teratoma

The prevalence of sacrococcygeal teratoma (SCT) is approximately 1 in 20,000 to 40,000.⁶³ Some fetuses with SCT undergo massive tumor enlargement, experience hydrops fetalis and placentomegaly, and die in utero. These tumors function as large arteriovenous fistulas, and fetal demise results from high-output cardiac failure. Management of these tumors requires close surveillance because they can grow rapidly and reach a size as large as 1000 cubic centimeters.⁶³ Fetuses with large lesions are at risk for intrapartum dystocia or tumor rupture and hemorrhage; these fetuses may require cesarean delivery. Fetuses with lesions diagnosed before 30 weeks’ gestation have a poor prognosis but may benefit from surgical debulking in utero; surgical techniques have not reached the necessary level of sophistication to allow complete resection of lesions that deeply invade the pelvis. In utero radiofrequency ablation and thermocoagulation have been used to reduce the tumor blood supply, but the benefit remains unclear.⁶⁴ Catheterization of a fetal hand or umbilical cord vein for blood and crystalloid transfusion during tumor resection may be needed. To date, there has been no significant improvement in outcome with intervention in cases of SCT with hydropic fetalis.

Some SCT cases are accompanied by “maternal mirror syndrome” or Ballantyne syndrome, a hyperdynamic state (i.e., hypertension, peripheral and pulmonary edema) in which the maternal physiology mirrors the abnormal circulatory physiology of the hydropic fetus.⁶⁵ This syndrome is associated with a substantial increase in fetal mortality and maternal morbidity and requires aggressive management similar to that used for severe preeclampsia, a disease from which it must be distinguished. Platelet count, aspartate aminotransferase, alanine aminotransferase, and haptoglobin are typically unaffected in maternal mirror syndrome and may serve as diagnostic clues to rule out severe preeclampsia and HELLP (hemolysis, elevated liver enzymes, low platelets). Unfortunately, maternal mirror syndrome typically does not resolve quickly, even with rapid correction of the fetal pathophysiology, and severe maternal complications including pulmonary edema occur in about 20% of cases.⁶⁵

Myelomeningocele

Although not lethal, a myelomeningocele is a protrusion of the meninges and spinal cord through a congenital defect in the vertebrae and overlying muscles and skin. It can result in lifelong morbidity and disability, including paraplegia, bowel and bladder incontinence, hydrocephalus, Arnold-Chiari II malformation, and impaired cognition.⁶⁶ Myelomeningocele has an incidence of about 1 in 2000 live births but is becoming less common owing to folate supplementation in the maternal diet. In addition, detection by ultrasonography and alpha-fetoprotein screening of maternal blood has allowed for earlier diagnosis (i.e., second trimester) and consideration of pregnancy termination.

The specific cause of myelomeningocele remains unknown. Animal models have demonstrated improved neonatal neurologic function with fetal closure of the defect in utero.^67,68 The results associated with defect closure support a “two-hit” disease model in which the pathology is produced by failure of the fetal neural tube to form combined with prolonged exposure to the uterine amniotic fluid.⁶⁹ Mutations of the PAX3 gene and direct cord trauma may also play a role in the pathophysiology associated with a myelomeningocele.⁷⁰

The 5-year mortality of myelomeningocele is approximately 8% for live births; if it is not corrected in utero, surgical closure must be performed within a few days after birth.⁷¹ Ventriculoperitoneal shunting is required in 85% to 90% of uncorrected cases; however, despite successful shunting, permanent deficits such as central hypoventilation, vocal cord dysfunction, and oromotor and swallowing dysfunction can still occur from the associated Arnold-Chiari malformation.⁷² The average intelligence quotient in myelomeningocele patients who require ventriculoperitoneal shunting is 80 (low normal).⁷³

The purpose of fetal surgery for myelomeningocele is to improve function later in life.⁶⁹ Fetal surgery is primarily performed through an open fetal surgical technique. Preliminary results suggest that in utero repair successfully reverses the hindbrain herniation of the Arnold-Chiari II malformation, probably through normalization of cerebrospinal fluid flow, and decreases the need for ventriculoperitoneal shunt placement before 1 year of age.⁷⁴ More recently, a randomized, prospective, multicenter clinical trial completed between 2003 and 2010 examined the risks and benefits of open fetal surgery for myelomeningocele repair compared with standard postnatal repair in 183 patients.⁷⁵ Open fetal repair reduced the need for ventriculoperitoneal shunting and improved motor function at 30 months of age, but increased the risk for preterm birth and a partial or complete uterine dehiscence. Two perinatal deaths occurred in each group. Table 7-2 displays a subset of outcome measures that were significantly different between the prenatal and postnatal repair groups. Further data from the trial will include assessment of the long-term benefit from the prenatal intervention.

A recent study of endoscopic intrauterine myelomeningocele repair resulted in an extraordinarily high complication rate for both mothers and fetuses.⁷⁶ Of the 19 study patients, three fetuses died intraoperatively and another three procedures were stopped owing to severe hemorrhage from the procedure. Although this intervention was associated with spinal segmental neuroprotection, it resulted in significantly more complications; the authors concluded that, pending further advances, this technique is unsuitable as standard care.⁷⁶

Twin-to-Twin Transfusion Syndrome

An abnormal connection of chorionic blood vessels in the placenta between two monochorionic twins can result in twin-to-twin transfusion syndrome (TTTS). TTTS complicates 10% to 15% of monochorionic pregnancies, usually manifests at 15 to 26 weeks’ gestation, and is typically recognized at 20 to 21 weeks’ gestation.^77,78 Intertwin transfusion is common between monochorionic twins and is usually balanced by the presence of arterioarterial (AA) and venovenous (VV) connections; the presence of AA connections is associated with a ninefold reduction in TTTS.⁷⁹ By contrast, unidirectional and imbalanced blood flow through arteriovenous (AV) chorionic vessels results in TTTS. In normal fetoplacental vasculature, the umbilical artery branches at the placenta surface and traverses and then descends into the tissue, where it further branches into capillary divisions for gas and nutrient exchange. The arterial system is “paired” with venous vasculature, which returns blood to the umbilical cord. In TTTS, the umbilical artery similarly descends into the placenta and cotyledon, but rather than connecting with a paired vein it connects with a vein that transports blood to the other twin.⁸⁰

The twin serving as the recipient demonstrates polycythemia, polyuria, polyhydramnios, and hypertrophic cardiomyopathy; this twin is at risk for hydrops fetalis and fetal death. The twin serving as the donor is typically hypovolemic, growth restricted, and pressed against the endometrium in an oligohydramniotic sac (hence the designation “stuck” or “pump” twin) and often has a velamentous cord insertion; this twin is at risk for neonatal renal failure, tubular dysgenesis and dysfunction, and high cardiac output hydrops fetalis. Diagnosis is currently based on ultrasonographic findings that focus on differences in fetal size or amniotic sac fluid volume, presence of cardiac dysfunction in the recipient twin, umbilical cord size, and abnormal umbilical arterial flow velocity.^77,81 Twin size discordance is not always present and is no longer considered necessary to confirm the diagnosis. For unclear reasons, fetuses with TTTS are at risk for neurologic injury with white-matter lesions and long term disability; poor neurodevelopmental outcomes are associated with increased TTTS severity, delayed therapeutic interventions, and preterm delivery.⁸² If TTTS is untreated, it carries a greater than 80% mortality with 15% to 50% risk for significant morbidity in surviving neonates.⁸³

A variety of therapeutic management techniques have been developed, including (1) serial amnioreduction to control polyhydramnios and reduce the risk for preterm labor, (2) surgical septostomy to equalize amniotic pressures, (3) selective feticide to allow the other fetus to survive, and (4) selective fetoscopic laser photocoagulation (SFLP) of the vascular anastomoses between the two twins. Serial amnioreduction was demonstrated to improve placental perfusion and decrease the occurrence of preterm delivery.⁸⁴ In a retrospective review of 223 twin sets with TTTS, amnioreduction resulted in an overall birth survival rate of 78%, with 65% of recipient twins and 55% of donor twins alive at 1 month of age.⁸⁵ In a prospective randomized trial comparing serial amnioreduction to septostomy, there was no difference in the rate of survival between the two techniques.⁸⁶ Septostomy is rarely used for treatment because the creation of a single amniotic sac can increase the risk for umbilical cord entanglement.

The laser used for SFLP is typically inserted percutaneously through an endoscope (≤ 2.0 mm diameter) into the recipient twin’s amniotic sac. Maternal anesthesia is commonly managed with either neuraxial blockade or local anesthetic infiltration from skin to myometrium. Fetoscope placement is determined by placental location and is guided by ultrasonography. Vessels that cross the dividing membrane separating the amniotic sacs are visualized, and abnormal connecting vessels are selectively coagulated with the laser.^77,80 On occasion, based on anatomic constraints created by the location of the fetuses and placenta, nonselective laser ablation is performed; however, this type of ablation is associated with higher rates of intrauterine fetal demise.^77,87 Ablation of all abnormal connecting vasculature is not needed for success of the procedure.⁸⁸ After completion of the SFLP, amniotic fluid may be removed to reduce the degree of polyhydramnios and possibly decrease the risk for preterm labor.

A 2004 randomized multicenter trial compared laser therapy to amnioreduction for treatment of severe TTTS diagnosed between 15 and 26 weeks’ gestation.⁸⁹ Rates of at least one twin survival were significantly higher in the laser treatment group at both 28 days (76% versus 56%, P < .01) and 6 months of life (76% versus 51%, P < .01). In addition, neurologic outcomes were better in the laser treatment group. A subsequent prospective study of a large subgroup of survivors from this trial observed these children for 6 years and found no additional change in survival or long-term neurologic outcome from the original 6-month data.⁹⁰ A meta-analysis of studies published between 1997 and 2007 noted that treatment of TTTS with laser ablation resulted in a higher overall rate of fetal survival than amnioreduction⁹¹; similar findings were demonstrated in a Cochrane review of treatment for TTTS.⁹²

A more recent variation of the SFLP technique requires vascular laser ablation in a specific sequence. First, ablation occurs at the donor-to-recipient AV anastomoses, then at the recipient- to-donor AV anastomoses, then at the AA superficial anastomoses, and finally at the VV superficial anastomoses. The order of the procedure is designed to reduce the chance of hemodynamic compromise and hypotension during the procedure in the donor twin⁹³; however, it is associated with longer operative times and increased case difficulty, particularly with an anterior placenta. In a single-institution study of consecutive SFLP for treatment of TTTS, twins treated with this sequence had better survival rates than twins whose SFLP procedure was performed without a specific sequence.⁹³ A prospective multicenter trial found a significantly improved 30-day survival of both fetuses and improved donor survival with this sequential technique when compared with a cohort control group.⁹⁴

The most common complication of SFLP is preterm PROM with subsequent preterm labor and delivery. Other possible complications include placement of the trocar through the placenta, hemorrhage, and possible membrane perforation resulting in limb entrapment and ischemia.⁷⁷ In conclusion, trial results and meta-analyses provide evidence that SFLP results in superior outcomes than amnioreduction for the treatment of TTTS. Further research is needed to determine optimal techniques and timing of interventions for the treatment of TTTS.

Twin Reversed Arterial Perfusion Sequence

In monozygotic twins, one twin can also perfuse the other by retrograde blood flow though AA anastomoses. Twin reversed arterial perfusion (TRAP) sequence affects 1% of monozygotic twins and 1 in 30 triplets. Inadequate perfusion of the recipient twin via retrograde flow results in the development of a lethal set of anomalies that include acardia and acephalus. The normal (“pump”) twin perfuses both itself and the nonviable twin and is at risk for high-output congestive heart failure, polyhydramnios, and preterm birth. If untreated, TRAP sequence is associated with a risk for intrauterine death of the pump twin exceeding 50%.⁹⁵ Diagnosis is confirmed with ultrasonographic demonstration of reverse flow to the acardiac twin via the umbilical artery. Cardiovascular failure in the pump twin is the indication for intervention, and early diagnosis is beneficial for optimal treatment.

The goal of therapy is to interrupt the vascular communication between the two twins. In contrast to the treatment of TTTS, treatment of TRAP sequence results in the death of the anomalous nonviable fetus. Percutaneous endoscopic laser or radiofrequency coagulation of the umbilical cord and/or placental vascular anastomoses is the most viable therapeutic option.^96,97 Alternative therapies include sectio parva (selective cesarean delivery of one of multiple fetuses), percutaneous thrombosis of the acardiac twin’s umbilical cord with coils or other thrombogenic material, and alcohol-impregnated suture cord ligation. A retrospective review of 60 TRAP sequence cases from multiple European centers using endoscopic laser coagulation noted overall survival rates approaching 80% and a median gestational age of 37.4 weeks at delivery.⁹⁸ An additional study, using radiofrequency coagulation in 26 TRAP sequence cases at a single U.S. medical center, demonstrated a 92% survival rate of the viable twin with a mean gestational age of 35.6 weeks at delivery.⁹⁶ Both procedures are typically performed with infiltration of local anesthesia at the insertion site of the ablation device, although neuraxial anesthesia has also been used. The procedures are guided by ultrasonography, and absence of flow to the nonviable acardiac twin is confirmed with Doppler imaging at the end of the procedure and again 12 to 24 hours later.

Congenital Heart Defects

The most commonly performed closed fetal cardiac intervention for a congenital heart defect is an aortic valvuloplasty for mid-gestational aortic stenosis with evolving hypoplastic left heart syndrome. Technical success as high as 75% has been reported using an angioplasty balloon over a guidewire inserted percutaneously through an access cannula.⁹⁹ Approximately 40% of successful cases result in aortic regurgitation and minimal subsequent left ventricular growth; however, the physiology of the left ventricle improves and leads to improved aortic and mitral valvular growth. In about 30% of the successful cases, biventricular circulation is present at birth. Other congenital heart defects that may benefit from antepartum closed fetal cardiac intervention include (1) hypoplastic left heart syndrome with an intact or highly restrictive atrial septum and (2) evolving hypoplastic right heart syndrome with pulmonary atresia or stenosis without a ventricular septal defect.^100,101

More complex congenital cardiac defects that might benefit from open repair have only been repaired in animal models and require fetal extracorporeal circulatory bypass; the use of fetal cardiac bypass induces a catecholamine-mediated fetal stress response that increases vascular resistance and cardiac afterload and is poorly tolerated by the immature fetal myocardium. Fetal cardiac bypass can also result in severe placental dysfunction when the high-capacitance, low-resistance placenta is incorporated as the oxygenator, resulting in endothelial dysfunction and leukocyte and complement activation. After fetal cardiac bypass, increases in placental vascular resistance, reduced blood flow, impaired gas exchange, and fetal acidosis are frequently observed.^102,103 Correction of complex congenital cardiac defects, either open or closed, requires careful anesthetic and pharmacologic strategies for myocardial protection.

Anesthesia for Minimally Invasive and Percutaneous Procedures

Local anesthetic infiltration of the abdominal wall is sufficient to reduce maternal discomfort for many percutaneous procedures (e.g., amniocentesis, cordocentesis, intrauterine blood transfusion, needle aspiration of cysts, shunt placement into the fetal bladder or thorax, SFLP for TTTS). Supplemental maternal analgesia and anxiolysis can be achieved by maternal administration of an opioid, a benzodiazepine, or a low-dose propofol infusion, which may confer some fetal immobility and analgesia via placental transfer.¹¹¹ However, when larger needles or multiple attempts are necessary for percutaneous procedures or a minilaparotomy, maternal comfort can be difficult to achieve with local infiltration, even when supplemented with sedation. In these circumstances, neuraxial anesthesia is recommended.^77,91 General anesthesia is rarely necessary for percutaneous procedures unless placental location makes the procedure significantly challenging or uterine exteriorization is needed.

Fetal movement may be hazardous for the fetus in cases of intrauterine transfusion, cord blood sampling, or thoracic shunt placement, because displacement of the needle or catheter may lead to trauma, bleeding, or compromise of the umbilical circulation. Placental transfer of maternally administered opioids and benzodiazepines can reduce fetal movement⁸⁹ but cannot ensure fetal immobility for more stimulating procedures. A randomized controlled trial of the use of a maternal remifentanil infusion (0.1 µg/kg/min) demonstrated improved fetal immobility and operating conditions during fetoscopic surgery, when compared with maternal administration of diazepam.¹¹² Fetal immobility can be safely achieved with direct fetal intramuscular or umbilical venous administration of pancuronium or vecuronium (0.3 mg/kg intramuscularly or 0.1 to 0.25 mg/kg intravenously) using ultrasonographic guidance. The onset of fetal paralysis occurs in 2 to 5 minutes, with an approximate duration of 1 to 2 hours.¹¹³ For procedures that can cause noxious stimulation to the fetus, such as shunt catheter placement or cardiac septoplasty, an opioid (e.g., fentanyl 10 to 20 µg/kg) can be administered to the fetus intramuscularly or intravenously.^114,115 When general anesthesia is employed, placental transfer of a volatile halogenated agent is usually sufficient to immobilize and anesthetize the fetus.

The fetal surgical team should be prepared for treatment of fetal compromise with immediate availability of appropriate doses of atropine and epinephrine. The obstetric team should be prepared to perform an emergency cesarean delivery if the gestational age is compatible with extrauterine viability. The anesthesiologist should be prepared to emergently provide general anesthesia if required.

Tocolysis typically is unnecessary after cordocentesis or intrauterine transfusion. For more invasive percutaneous procedures (e.g., shunt catheter placement, endoscopic techniques), some fetal surgery groups administer prophylactic tocolytic agents.

Anesthesia for Open Fetal Surgery

When corrective fetal surgery or an intrauterine procedure requires surgical access through a hysterotomy, general anesthesia is typically administered. Unique considerations for open fetal procedures include the need for profound uterine relaxation, intraoperative fetal monitoring, fetal anesthesia or analgesia, and postoperative maternal analgesia and uterine tocolysis (Table 7-3). In addition, significant maternal and fetal blood loss may occur, and the anesthesiologist must be prepared to achieve maternal and fetal resuscitation. A high concentration of a volatile halogenated agent is typically administered to provide both maternal and fetal anesthesia as well as uterine relaxation; adequate uterine relaxation may require greater than twice the minimum alveolar concentration (MAC) of a volatile halogenated agent.¹¹⁶

Preoperatively, the mother receives agents for aspiration prophylaxis and uterine tocolysis (e.g., rectal indomethacin) and an epidural catheter is placed for postoperative analgesia. Minimal doses of preanesthetic and adjuvant anesthetic agents are given to allow significant doses of a volatile halogenated agent to be administered to achieve effective uterine relaxation; this approach may also reduce the occurrence of hypotension. The patient is placed in the supine position with left uterine displacement. After administration of 100% oxygen and denitrogenation of the lungs, a rapid-sequence induction of general anesthesia with cricoid pressure and endotracheal intubation is performed. Fetal heart rate (FHR) and umbilical cord blood flow are often monitored with ultrasonography during induction.

Initially, anesthesia is maintained with a low concentration of a volatile halogenated agent while further preparations for surgery are undertaken, including (1) obtaining additional maternal vascular access, (2) prophylactic antibiotic administration, (3) urinary bladder catheterization, and (4) ultrasonographic assessment of fetal presentation and placental location. Medications to provide fetal analgesia (e.g., fentanyl 10 to 20 µg/kg), immobility (e.g., vecuronium 0.3 mg/kg), and resuscitation (atropine 0.02 mg/kg, epinephrine 1 and 10 µg/kg, and crystalloid 10 mL/kg) are prepared in sterile labeled syringes; each syringe should contain a single weight-based unit dose. Crossmatched blood should be available for maternal transfusion. In addition, O-negative, cytomegalovirus (CMV)-negative, irradiated, leukocyte-depleted, maternally crossmatched blood should be available for fetal transfusion. For open procedures that have a high risk for significant fetal blood loss (e.g., mass resection), an intravenous catheter should be placed in the fetus to provide access for blood and fluid transfusions. An arterial catheter should be placed for maternal blood pressure monitoring if uterine tocolysis with a nitroglycerin infusion is planned. Intraoperative maternal intravenous fluids are restricted (< 2 L) to reduce the risk for postoperative pulmonary edema; some fetal surgery centers choose to limit fluids even further (< 500 mL), although no clinical trials have proven a benefit of fluid restriction in this setting.¹¹⁷ The use of tocolytic agents such as magnesium sulfate and nitroglycerin has been associated with maternal pulmonary edema in patients undergoing fetal surgery.

A final discussion (i.e., surgical time-out), prophylactic antibiotic administration, and an increase in the concentration of the volatile halogenated agent should occur before skin incision. Maternal blood pressure is maintained with a mean arterial pressure within 10% of baseline values and greater than 65 mm Hg; a phenylephrine infusion provides titratable blood pressure control with minimal changes in the fetal acid-base status.^118,119 Ephedrine or glycopyrrolate boluses also can be administered to maintain maternal heart rate and improve cardiac output.¹²⁰ Maternal administration of a nondepolarizing muscle relaxant is usually not needed owing to the profound depth of anesthesia but may be used to improve operative conditions. The intrauterine location and position of the fetus is confirmed with ultrasonography just before hysterotomy to optimize the incision location.

Before uterine incision, it is important to achieve an increased end-tidal concentration of the volatile halogenated agent to provide both fetal anesthesia and uterine relaxation (typically ≥ 2 MAC). Fetal well-being can be assessed with pulse oximetry, ultrasonography for FHR monitoring, echocardiography to assess fetal ventricular contractility, and electrocardiography (ECG).

The uterus is assessed both visually and by palpation for contractions or increased tone; further tocolysis can be achieved with the administration of additional halogenated agent (≤ 3 MAC) or intravenous nitroglycerin as an infusion or in small boluses (50 to 200 µg).¹²¹ Use of desflurane at 1.5 MAC with supplemental intravenous propofol and remifentanil has provided adequate uterine relaxation in one retrospective study.¹²² For circumstances in which volatile halogenated agents or general anesthesia must be avoided (e.g., family history of malignant hyperthermia), a spinal or epidural anesthetic can be used with an intravenous infusion of nitroglycerin in doses up to 20 µg/kg/min.¹²¹ This technique does not have a clear advantage for fetal outcome and may be associated with more morbidity. Nitroglycerin administration during open fetal surgery has been associated with maternal pulmonary edema.^117,123 Fetal intraventricular and periventricular hemorrhage and cerebral ischemia can result from changes in fetal cerebral blood flow, and concern has been raised regarding the use of tocolytic agents, which may affect vascular tone.¹²⁴

An opioid and a muscle relaxant are administered to the fetus intramuscularly, either before uterine incision with ultrasonographic guidance or after uterine incision with direct vision. Some anesthesiologists administer atropine at this time in an effort to prevent opioid-induced fetal bradycardia; a muscle relaxant with vagolytic effects also can be chosen to minimize this type of bradycardia. Further studies are needed to determine the optimal anesthetic technique for ensuring maternal and fetal cardiovascular stability, optimal uteroplacental perfusion, and adequate fetal anesthesia to cause immobility and blockade of the fetal stress response.

A small uterine incision is created remote from the location of the placenta. A stapling device with absorbable synthetic copolymer (Lactomer) staples is used to extend the incision to seal the membranes to the endometrium and prevent excessive bleeding.¹²⁵ During surgery, the exposed fetus and uterus are bathed with warmed fluids. The intrauterine temperature is closely monitored to prevent fetal circulatory compromise associated with hypothermia.¹²⁶

When uterine closure is initiated at the conclusion of the procedure, a loading dose of magnesium sulfate is administered (4 to 6 g intravenously over 20 minutes), followed by an intravenous infusion of 1 to 2 g per hour. As magnesium potentiates neuromuscular relaxation, close monitoring of twitch recovery is needed if a nondepolarizing muscle relaxant was administered. The volatile halogenated agent can be significantly decreased or discontinued after the magnesium sulfate bolus has been administered. The epidural analgesia can be initiated and maternal anesthesia is maintained with additional fentanyl, nitrous oxide, and/or propofol.

Postoperative concerns include maternal and fetal pain, preterm PROM, preterm labor, infection, and a variety of potential fetal complications, including heart failure, intracranial hemorrhage, constriction of the ductus arteriosus from indomethacin, and fetal demise. Postoperative maternal analgesia can be maintained with a continuous epidural infusion of a dilute solution of local anesthetic and opioid for several days. Effective analgesia may help prevent postoperative preterm labor.^127,128 Intravenous opioids can also be used for analgesia; however, decreased FHR variability may occur.¹²⁹

Management of postoperative preterm labor has been the “Achilles heel” of fetal surgery.⁷ Tocolysis is typically provided by an infusion of magnesium sulfate for at least 24 hours; however, supplemental agents may include indomethacin, terbutaline, or nifedipine. Magnesium most likely competes with calcium at voltage-operated calcium channels, indomethacin blocks the synthesis of prostaglandins, and beta-adrenergic agonists activate adenylate cyclase in the uterine muscle, thereby reducing intracellular calcium levels. Frequently two tocolytic agents are required to create uterine quiescence. Uterine activity and FHR are monitored closely during the first 2 to 3 postoperative days. The fetus is evaluated postoperatively by ultrasonography, and if indomethacin is used, periodic fetal echocardiography is conducted to determine whether premature closure of the ductus arteriosis has occurred.

Patients recovering from open fetal surgery should remain near the fetal treatment center after being discharged. These patients are at high risk for preterm PROM, preterm labor, infection, and uterine rupture. Unless preterm labor occurs, cesarean delivery is typically planned at 37 weeks’ gestation.

Anesthesia for the Ex Utero Intrapartum Treatment Procedure

Initially described as a method to remove the iatrogenic airway obstruction created for intrauterine treatment of CDH, the EXIT procedure has evolved into a technique useful for a number of fetal disorders that compress the airway and/or render neonatal tracheal intubation difficult or impossible. It is also useful when resuscitation and surgical intervention are required immediately before birth, while the fetus is still supported by the placentofetal circulation. Cases appropriate for an EXIT procedure include (1) thoracotomy for congenital pulmonary airway malformations and (2) tracheostomy and the removal of neck masses such as fetal teratoma. The use of an EXIT procedure may also assist the transition to ECMO for pulmonary insufficiency or the stabilization of conjoined twins prior to separation.^11,130,131 Similar to open fetal surgery procedures, sustained uterine relaxation and delay of placental separation are necessary for a successful EXIT procedure.

Anesthesia for EXIT procedures is most commonly performed with the use of general anesthesia, although neuraxial anesthesia, typically combined with intravenous nitroglycerin infusion to achieve uterine relaxation, may also be used. Multiple reviews of the surgical and obstetric considerations associated with the EXIT procedure have been published.^132–134

The conventional anesthetic approach is a modification of the general anesthetic technique used for cesarean delivery. Preparation for fetal monitoring, airway management, fetal/neonatal resuscitation, and postdelivery care should be completed before entering the operating room. A sterile pulse oximeter probe and an end-tidal carbon dioxide indicator or gas analyzer are used for fetal monitoring during the procedure; basic ultrasonography can be added to assess the fetal heart. Unit doses of atropine, epinephrine, and calcium for intramuscular fetal injection are transferred sterilely to the scrub nurse for possible emergency intraoperative resuscitation. Supplemental fetal anesthetic agents for intramuscular injection are also prepared and transferred to the scrub nurse (see later discussion). A sterile ventilation bag with an air/oxygen source and manometer is available for the fetus, along with multiple endotracheal tube sizes and devices for fetal tracheal intubation. Catheters for intravenous access as well as crystalloid and blood (O-negative, CMV-negative, leukocyte depleted, irradiated, maternally crossmatched) should be available for fetal volume resuscitation if needed.

A maternal epidural catheter may be placed preoperatively for postoperative analgesia. Anesthetic considerations for the mother are similar to those for cesarean delivery but should also include large-bore intravenous access, availability of uterotonic agents and crossmatched blood, and the ability to quickly obtain invasive maternal monitoring if needed.

Techniques of induction and tracheal intubation do not differ from those typically used for cesarean delivery. Although techniques for maintenance of anesthesia vary between medical centers, administration of 2 to 3 MAC of a volatile halogenated agent is often needed to achieve and maintain adequate uterine relaxation. Occasionally, nitroglycerin administered intravenously as a bolus dose (50 to 200 µg) or as an infusion is required to obtain appropriate uterine relaxation. Fetal anesthesia from the halogenated agent transferred across the placenta can be supplemented by direct fetal (intramuscular) administration of an opioid (fentanyl 5 to 15 µg/kg) and a paralytic agent (rocuronium 1 to 3 mg/kg or pancuronium 0.1 to 0.3 mg/kg); some practitioners also administer intramuscular atropine (20 µg/kg) to prevent fetal bradycardia.^132,134 Intramuscular agents can be administered to the fetus either before uterine incision with ultrasonographic guidance or after incision with direct visualization. Significant variability in serum fentanyl concentrations among fetuses has been documented in umbilical cord blood during EXIT procedures,¹³⁵ and similar variability may exist with muscle relaxants and other agents.

When adequate uterine relaxation has been achieved, the placental location and edges are confirmed by intraoperative ultrasonography. Similar to open fetal procedures, the uterine incision is extended with a stapling device to minimize blood loss. The fetal head and shoulders are delivered in preparation for tracheal intubation. For more extensive procedures, such as fetal thoracotomy, or when there is fetal bradycardia suggestive of umbilical cord compression, the fetus can be completely delivered and placed on the maternal chest and abdomen. In an effort to maintain fetoplacental circulation, warmed fluids are continuously irrigated into the uterine cavity and care is taken to avoid manipulation of the umbilical cord. The warmed irrigant maintains fetal euthermia and helps prevent decreased uterine volume, placental separation, and spasm of the cord vessels. The fetus is initially monitored with (1) a pulse oximeter probe placed on the fetal hand, (2) periodic ultrasonography, and (3) direct visualization.

The duration of the fetal procedure can range from a few minutes (e.g., bronchoscopy or intubation) to several hours (e.g., neck or thoracic mass resection, tracheostomy, or central intravascular access). Although the majority of procedures require less than an hour, the anesthetic technique should be capable of providing maternal, fetal, and uteroplacental stability over several hours.^136,137 Once surgery is completed and the trachea secured, surfactant is given if indicated. Fetal oxygen saturation is typically 40% to 60% at this time¹³⁸ but increases significantly to above 90% with ventilation of the fetal lungs. Failure to achieve a fetal saturation of at least 90% is an indication for the use of ECMO before clamping the umbilical cord and delivering the fetus.¹³⁹

After the umbilical cord is clamped, the maternal anesthetic technique is altered to help achieve uterine tone and diminish the risk for postpartum hemorrhage.^132,140 This is accomplished by (1) substantially reducing the inspired concentration of the volatile halogenated agent; (2) adding nitrous oxide, propofol, and/or an opioid to maintain anesthesia; (3) administering oxytocin and other uterotonic agents, if needed; and (4) initiating epidural analgesia via the epidural catheter that was placed before surgery.

An alternative technique for EXIT procedures involves the use of neuraxial anesthesia.^141–143 Although this approach avoids many of the risks associated with general anesthesia, large doses of intravenous nitroglycerin (1 to 10 µg/kg/min) are often required to achieve adequate uterine relaxation. Nitroglycerin can cross the placenta; however, a significant amount is metabolized at the placental interface, resulting in minimal fetal effects.^121,143 If nitroglycerin is to be used at significant doses for a prolonged period, an arterial catheter is recommended and the patient should be observed for evidence of pulmonary edema. Fetal analgesia and immobility can be achieved with fetal administration of intramuscular drugs (see earlier discussion); maternally administered intravenous remifentanil undergoes significant transfer across the placenta and may serve as an adjuvant for fetal analgesia and immobility.^144,145 Prospective trials are necessary to delineate the advantages and disadvantages of these various anesthetic techniques for the EXIT procedure.

Fetal Response to Surgical Stimulation

The subjective phenomenon of pain has not been, and perhaps cannot be, assessed adequately in human fetuses. Pain is a multidimensional, subjective, psychological construct that can exist in the absence of physical stimuli (e.g., phantom limb pain), and it includes emotional and affective components that require higher-level cortical processing. As such, although pain is commonly associated with physical noxious stimuli, it is more than nociception or a simple reflex activity associated with a withdrawal response.¹⁴⁶ Attempts have been made to correlate pain with surgical stress; however, the physiologic responses are not equivalent and reductions in stress hormones should not be interpreted as an indicator of adequate analgesia.¹⁴⁷ The stress response is mediated primarily in the spinal cord, brainstem, and/or basal ganglia, without involvement of the cortex. This distinction acknowledged, pioneering studies of preterm neonates undergoing surgery with minimal anesthesia have revealed circulatory, sympathoadrenal, and pituitary adrenal responses characteristic of stress (e.g., increased release of catecholamines, growth hormone, glucagon, cortisol, aldosterone, and other corticosteroids; decreased secretion of insulin).^148,149 Administration of adequate anesthesia in neonates blunts this stress response,¹⁵⁰ and attenuation of the stress response in preterm neonates has been associated with improved outcomes.¹⁵¹

In studies of intrauterine blood transfusion in the human fetus, surgical needling of the intrahepatic vein (in contrast to the insensate umbilical cord) is associated with evidence of a stress response, including increases in plasma beta-endorphin and cortisol concentrations and decreases in the middle cerebral artery pulsatility index (determined by Doppler imaging). These responses, which may redistribute blood flow to vital organs, including the brain,¹⁵² can be blunted by the administration of fentanyl 10 µg/kg.¹¹⁴ Human fetuses elaborate pituitary-adrenal, sympathoadrenal, and circulatory stress responses to noxious stimuli as early as 18 weeks’ gestation.^153–156 During late gestation, fetuses can respond to environmental stimuli such as noise, light, music, pressure, touch, and cold.¹⁵⁷

Nerve terminals for the detection of touch, temperature, and vibration (not pain) are present deep within the human skin at 6 weeks’ gestation and become more numerous and superficial (e.g., toward the skin surface) by 10 weeks’ gestation.¹⁵⁸ Immature skin nociceptors likely begin to emerge at 10 weeks’ gestation and are present by 17 weeks’ gestation¹⁵⁹; in internal organs, nociceptors develop slightly later.

Peripheral nerve fibers that control movement grow into the spinal cord at about 8 weeks’ gestation. When these fibers connect with nociceptors is unknown, but the timing of these connections is delayed when compared with other sensory inputs in nonhuman, mammalian models. One human study suggested that nociceptive nerve fibers do not enter the spinal cord before about 19 weeks’ gestation.¹⁶⁰ The cerebral cortex develops after the fetal spinal cord and brainstem. Thalamocortical axons reach the somatosensory cortex at 24 to 26 weeks’ gestation.¹⁴⁶

The developing cerebral cortex consists of transient fetal zones where neuronal proliferation, cell migration, apoptosis, axonal outgrowth, and synaptogenesis occur according to a highly specific timetable (see Chapter 10). Originating as a smooth layer without sulci and gyri, the cerebral cortex, like the thalamus, has no internal cellular organization.¹⁵⁸ The cortical subplate is a temporary structure that serves as a waiting and organizing zone for various afferents destined for the cortex; it develops about 13 weeks’ gestation and recedes after 32 to 34 weeks’ gestation. The insular cortex starts developing in humans at approximately 15 weeks’ gestation, with the cortical plate eventually developing into the six layers of the cerebral cortex.^161–163

The first fibers from the thalamus reach the subplate between 12 and 18 weeks’ gestation and remain until the maturation of the cortical plate. The gestational age at which thalamic pain fibers reach the human cortex only can be estimated from histologic studies of other thalamocortical circuits. Thalamic projections reach the visual subplate at 20 to 22 weeks’ gestation,^162,164 the visual cortex at 23 to 27 weeks’ gestation,¹⁶⁵ and the auditory cortical plate at 26 to 28 weeks’ gestation.¹⁶⁶ The subplate becomes thinner in the insula and in areas of the brain where early cortical folding occurs. Extensive brain growth and maturation occur after 34 weeks’ gestation, resulting in cortical sulci and gyri and an extensive network of pathways within the cortex and to the thalamus, midbrain, and spinal cord. Prior to the completion of the thalamocortical system, the midbrain reticular system is possibly responsible for pain awareness, as it is for consciousness.^162,167

Studies of fetal electroencephalograms (EEGs) at 24 weeks’ gestation demonstrate electrical activity only 2% of the time, predominantly in 20-second bursts with periods of inactivity lasting up to 8 minutes. At 30 weeks’ gestation, EEGs begin showing patterns of wakefulness and sleep, but these are not concordant with fetal behavior. By 34 weeks’ gestation, electrical activity is present 80% of the time, with more distinct wakefulness and sleep cycles similar to adult patterns.

Low levels of oxygenation in utero and endogenous neural inhibitors such as adenosine and pregnenolone may preclude optimal functioning of neural tissues and networks.¹⁶⁸ However, two studies using near-infrared spectroscopy in preterm infants demonstrated differences in cerebral oxygenation over the somatosensory cortex with noxious and non-noxious stimulation. This appears to indicate that noxious information is at least transmitted to the infant cortex by 25 weeks’ gestation^169,170; similarly, preterm neonates also have demonstrated cortical evoked potentials after a heel lance.¹⁷¹

Although initial fetal reactions to nociceptive stimulation are purely reflexive, and cortical processes can occur only after thalamocortical connections have been completed, nociceptive stimuli can activate a number of subcortical mechanisms; moreover, stress responses may influence the maturation of both the thalamocortical pathways and the cerebral cortex. Although the data on the consequences of fetal exposure to stressful stimuli are incomplete, recognition should be given to the possibility that noxious stimuli, which can be attenuated or blocked by anesthesia, may be associated with adverse long-term neurodevelopmental consequences.¹⁷² For example, circumcision in nonanesthetized neonates has been associated with increased pain responses to injections performed 6 months later¹⁷³; in addition, fetal stress has been associated with long-term adverse hormonal effects in young monkeys.¹⁷⁴

The exact onset of fetal sentience, the capacity to feel pain, is unknown (see Chapter 5). Because of this uncertainty, it seems best to err on the side of administering adequate fetal anesthesia.¹⁷⁵ Altogether, clinical observations of fetal and neonatal behavior, information about the development of mechanisms of pain perception, and studies of fetal and neonatal responses to noxious stimuli provide a compelling physiologic and philosophic rationale for the provision of adequate fetal anesthesia, especially after 24 to 26 weeks’ gestation. Noxious stimulation during fetal life causes a stress response, which could have both short- and long-term adverse effects on the developing central nervous system. Although the link between the stress response and pain is not always predictable, the threshold for pain relief is typically below that for stress response ablation, and the stress response to noxious stimulation is clear evidence that the fetal nervous system is reactive.¹⁷⁶ Administration of fetal anesthesia has been the standard practice worldwide since the inception of fetal surgery more than 30 years ago.^5,6,177,178 The importance of fetal immobility, cardiovascular homeostasis, analgesia, and perhaps amnesia have always been emphasized in fetal surgery practice.

Effects of Anesthesia on the Fetus

In fetal lambs, the concentration of halothane required to prevent movement in response to painful stimuli (1.0 MAC) is less than 40% of the concentration required in adult sheep.¹⁷⁹ Despite rapid placental transfer of volatile halogenated agents, experiments in sheep models have shown that fetal concentrations of these agents remain lower than maternal concentrations for significant periods after maternal administration (Figure 7-3).¹⁸⁰

Experimental studies of the fetal effects of maternal administration of a volatile halogenated agent in sheep have not produced uniform results.^180–183 Maternal administration of isoflurane and halothane resulted in variable changes in fetal blood pressure, heart rate, oxygen saturation, cardiac output, and acid-base status. A retrospective analysis of cardiac imaging from both open fetal cases and EXIT procedures noted severe left ventricular systolic dysfunction in the fetus with use of high concentrations of desflurane.¹²²

Maternal-fetal sheep studies, with no surgical stimulus to either the mother or fetus, show that deep maternal inhalation anesthesia (> 2.0 MAC) results in progressive fetal acidosis. By contrast, lower maternal inhalation anesthesia (1.0 MAC) also may be undesirable because it does not adequately block the fetal response to a painful stimulus, which includes increased fetal catecholamines, vasoconstriction, and redistribution of fetal blood flow.¹⁸⁴ However, the combined effect of adequate fetal anesthesia with a halogenated agent, intrauterine manipulation, and fetal stress on fetal cardiovascular stability and regional blood flow remains unknown. Brief fetal exposure to deep maternal inhalation anesthesia (i.e., 2.0 to 3.0 MAC) does not appear significantly detrimental to the fetus, with no fetal hypoxia, hypercarbia, or acidosis observed even after exposures of 2 hours if maternal arterial pressure is maintained.¹³⁶ However, others have seen acidosis after 45 minutes of fetal exposure to anesthesia.¹⁸⁵

Another concern is that anesthetic agents may result in neuronal apoptosis in the developing fetal brain (see Chapter 10). Initial evidence was found in 2003, when an anesthetic consisting of midazolam, nitrous oxide, and isoflurane was shown to alter neurons in the developing brain of 7-day-old rats and to cause long-term impairment of brain function.¹⁸⁶ Additional in utero rat studies noted that fetal exposure to isoflurane for 4 hours near mid gestation resulted in abnormal spatial memory acquisition.¹⁸⁷ Primate studies have demonstrated significant neurodegeneration in the neonatal period after exposure to isoflurane.¹⁸⁸ It is not currently known if anesthetic agents similarly affect human fetuses or newborns. Several retrospective human studies have produced inconclusive results. One retrospective cohort trial noted that a single short anesthetic exposure in children younger than age 2 years did not have long-term cognitive implications; however, exposure to multiple anesthetics was a significant risk factor for development of learning disabilities.¹⁸⁹ Another retrospective study examined anesthetic exposure among twins and found no causal relationship between early anesthetic exposure and learning disabilities.¹⁹⁰ Although there is concern for neurotoxicity in human fetuses or children exposed to anesthetic drugs, current findings remain inconclusive about the long-term effects of anesthetic agents on brain function in humans. A 2007 U.S. Food and Drug Administration advisory committee concluded that no change in clinical practice is justifiable based on current data.^191,192 The neurocognitive effect of anesthetic exposure on fetuses during fetal surgery remains unknown.

Fetal Monitoring

Maternal and fetal anesthesia, uterine incision, fetal manipulation, and surgical stress may adversely affect uteroplacental and fetoplacental circulation by several mechanisms. Maternal hypotension, increased uterine activity, and maternal hyperventilation and hypocarbia impair uteroplacental and/or umbilical blood flow. Fetal manipulation may affect fetal cardiac output, regional distribution of cardiac output, and umbilical blood flow. Direct compression of the umbilical cord, inferior vena cava, and mediastinum adversely affects fetal circulation. Current methods of intraoperative fetal monitoring include FHR monitoring, pulse oximetry, ultrasonography (including echocardiography and Doppler assessment of umbilical cord blood flow), and blood gas and acid-base analysis.

Intraoperative monitoring of the FHR during open fetal surgery was initially attempted with a standard fetal corkscrew electrode processed by a standard FHR cardiotachometer, but signal failure was frequent, secondary to low signal amplitude and movement artifact. With use of atrial pacing wires sutured to the fetus, proximal wire shielding, increased gain to allow signal amplification, and filter modification, a more reliable display of fetal ECG with visible P waves and QRS complexes was possible. Unfortunately, this technique did not eliminate motion artifact. Analysis of the fetal ECG using ST waveform analysis may prove beneficial for fetal surgery.¹⁹³ A fetal scalp electrode has been used successfully as part of an EXIT procedure to monitor FHR after the head was exposed.¹⁹⁴

Plethysmography combined with pulse oximetry has proved to be very useful, particularly for the EXIT procedure. The predictive value of pulse oximetry may be superior to FHR monitoring; bradycardia has been found to be a late sign of fetal compromise in fetal lambs subjected to umbilical cord compression (Figure 7-4).¹⁹⁵ However, bradycardia can also precede oxyhemoglobin desaturation during human fetal surgery (Figure 7-5).¹⁹⁶

Ultrasonography is a crucial intraoperative fetal monitoring device. The FHR can be determined with visualization of the heart or with Doppler assessment of umbilical cord blood flow. Fetal cardiac contractility and volume also can be assessed qualitatively by echocardiography. Unfortunately, the sterile transducer often cannot be positioned continuously because its location interferes with surgery.

New devices may allow monitoring of (1) fetal blood pressure and EEG; (2) fetal arterial blood oxygen saturation, PO₂, and PCO₂; and (3) fetal cerebral oxygenation, blood volume, and blood flow with near-infrared spectroscopy.