Anesthesia and surgery can cause derangements of maternal physiology that may result in hypoxia, hypercapnia, stress, and abnormalities of temperature and carbohydrate metabolism. These states can increase oxidative stress and be teratogenic or may enhance the teratogenicity of other agents.

Severe hypoglycemia and prolonged hypoxia and hypercarbia have caused congenital anomalies in laboratory animals, but there is no evidence to support teratogenicity after brief episodes in humans. Prolonged exposure to hyperglycemia is a known teratogen causing cardiac, CNS, and skeletal abnormalities, but any organ system can be affected. Anomalies are tightly linked with duration and degree of hyperglycemia, particularly in the first trimester, but it is unlikely that a brief period of intraoperative hyperglycemia has any teratogenic effects. Although severe hypoxemia causes structural teratogenesis in animals and humans, the chronic hypoxemia experienced by mothers at high altitudes results in small-for-gestational-age infants but without an increase in congenital defects. Maternal stress and anxiety have been associated with poor fetal neurocognitive development (behavioral teratogenicity). Studies following obstetric outcomes in women in war zones have demonstrated that maternal emotional trauma and posttraumatic stress disorder may affect fetal growth and increase the risk for miscarriage and hypertensive disorders of pregnancy, but these studies may be confounded by the lack of infrastructure to meet basic prenatal healthcare needs. Hypothermia is not teratogenic, whereas hyperthermia is teratogenic in both animals and humans. Congenital anomalies, especially involving the CNS, have repeatedly been associated with maternal fever during the first half of pregnancy. It must be remembered that fetal temperature is on average 0.5°C to 1°C higher than maternal temperature. Embryonic oxidative stress from reactive oxygen species has been implicated as one of the mechanisms involved in teratogenicity of many agents.

Ionizing radiation is a human teratogen that can result in an increased, dose-related risk for malignant disease, genetic disease, congenital anomalies, and/or fetal death. Radiation is expressed as grays (Gy) or milligrays (mGy) (1 Gy = 100 rad) and is evaluated as cumulative dose (i.e., background radiation and medical diagnostic radiation) throughout the entire pregnancy. No dose of radiation is considered safe, but the type and severity of effects vary with the radiation exposure to the uterus and fetus and with the gestational age of the fetus ( Table 18.2 ). Background radiation during gestation is 1.3 to 5.8 mGy. There is no evidence that radiation exposure below 50 mGy is associated with a teratogenic effect in either humans or animals. In contrast with the negligible risk for teratogenicity, observational studies suggest that there is a slightly higher risk for childhood cancer at radiation doses of 10 mGy or higher. The relative risk for childhood malignancy after maternal abdominal radiation exposure has been estimated to be 2.28 (95% confidence interval [CI], 1.31 to 3.97). The absorbed fetal dose for all conventional radiographic imaging inside or outside the abdomen and pelvis is negligible and is well below 50 mGy, with most falling under 5 mGy ( Table 18.3 ). However, direct radiographic examination using computed tomography or fluoroscopy of the abdomen and pelvis may result in significant fetal radiation exposure, with doses that may approach 100 mGy. Interventional radiologic procedures (e.g., cerebral angiography, cerebral embolization, endoscopic retrograde cholangiopancreatography) are frequently complex and prolonged, particularly when surgery is not an option. These procedures, especially the abdominal interventions, expose the fetus to a significant radiation dose, sometimes more than 50 mGy; thus, professional societies have developed guidance for the use of these technologies during pregnancy.

If the mother’s condition necessitates diagnostic testing and radiation exposure, and no other acceptable imaging modality is available, testing should not be withheld if the benefits to the mother (and, by extension, the fetus) are judged to outweigh the risks. Delayed diagnosis of a critical condition may result in significant harm to the maternal/fetal dyad. The consultation with the radiologist should include discussion of the optimal diagnostic modality that adheres to the ALARA principle (as low as reasonably achievable) for radiation and minimizes the absorbed dose, through abdominal shielding, scan length reduction, and milliampere modulation. The avoidance or minimization of IV radiopaque iodinated contrast media, which crosses the placenta and theoretically can cause neonatal hypothyroidism, should be considered as well.

Diagnostic ultrasonography during pregnancy has long been considered devoid of embryotoxic effects. In animals, ultrasound intensities up to 20 W/cm ² have been found to be safe. However, in rat models, when higher intensities (>30 W/cm ² ) have been used, or when repeated exposure has occurred during early pregnancy, postnatal neurobehavioral effects have been described. Ultrasound waves produced by modern diagnostic equipment can induce significant increases in fetal temperature, especially when imaging is prolonged. Although Miller et al. concluded that the range of temperatures produced could be teratogenic, no biologic effects have been documented from diagnostic ultrasonographic examinations in the pregnant patient, despite widespread use over several decades. Nonetheless, it must be stressed that these epidemiologic studies were conducted in an era when ultrasound equipment was less potent and associated with lower increases in temperature. Doppler interrogation emits significantly more acoustic intensity, and subsequently more heat, than pulse-echo imaging equipment. Therefore, ultrasonographic examinations should be used judiciously, keeping the exposure time and acoustic output to the lowest level possible.

Gadolinium-based MRI contrast agents readily cross the placenta, enter the fetal circulation, and are excreted by the fetal kidneys into the amniotic fluid, where they may be swallowed by the fetus or resorbed by the mother. Gadolinium can be detected in the brain, bone, and liver in mammalian models. A large retrospective study designed to evaluate the effects of MRI exposure interrogated a Canadian universal healthcare database of all births after 20 weeks’ gestation from 2003 to 2015. The relative risk for stillbirth was higher (1.68) among those pregnancies in which the fetus was exposed to gadolinium-MRI in the first trimester ( n = 1737) compared with no MRI ( n = 1,418,451); no conclusion could be made for MRI with and without gadolinium because of small sample sizes. Gadolinium can cause fetal bradycardia, late fetal heart rate (FHR) decelerations, and preterm labor. Adequately performed studies on the effects of gadolinium in pregnancy have not yet been performed. The American College of Radiology and the American College of Obstetricians and Gynecologists (ACOG) recommend that IV gadolinium be avoided during pregnancy and used only if absolutely essential.

Early rodent studies documented teratogenicity (CNS malformations, skeletal abnormalities, and growth restriction) after exposure to a variety of opioids; however, these conclusions have been challenged because opioids given in large bolus injections can cause respiratory depression and impaired feeding, which may be teratogenic. In a rat study designed to avoid such problems, Fujinaga and Mazze maintained clinically relevant concentrations of morphine and failed to observe structural anomalies, although fetal growth restriction was present, and offspring mortality was increased. Using the same methodology, Fujinaga et al. found fentanyl , sufentanil , and alfentanil completely devoid of teratogenic effects. Additional mouse studies have confirmed the absence of structural teratogenicity with opioids.

Use of tranquilizers and anxiolytics in pregnancy has been investigated less rigorously than opioids. Animal studies have demonstrated structural or behavioral teratogenesis after exposure to some of the barbiturate , phenothiazine , and tricyclic antidepressant agents. The reader is referred to standard teratology reference sources and package inserts for animal and human data related to specific drugs.

Studies in rodents and nonhuman primates indicate that exposure of the immature brain to anesthetic agents such as propofol , thiopental , and ketamine (i.e., agents classified as N -methyl-d-aspartate [NMDA] antagonists and gamma-aminobutyric acid [GABA] agonists) are associated with brain cell apoptosis and functional learning deficits (see Chapter 10 ). A metaanalysis by Bleeser et al. demonstrated that anesthesia-induced neurotoxicity during pregnancy is a consistent finding in animal studies, with neurodegeneration increasing with longer duration, greater dosage, and higher frequency of agent exposure.

However, recent animal and human evidence of exposure of the fetal brain to various anesthetic agents has been reassuring. Using a rabbit model, Van der Veeken et al. injected anesthetic agents and neuromuscular blocking agents into the fetus and observed no brain modifications; the addition of fetal surgery yielded no neurodevelopmental effects. In a similar rabbit model, 2 hours of maternal exposure to general anesthesia and laparotomy, with minimal organ and no fetal manipulation, resulted in slower offspring motor neurodevelopment, but the effect was almost undetectable by 7 weeks. By contrast, no fetal neuronal injury or neurobehavioral impairments were observed in an ovine model with maternal exposure to anesthesia for 5 hours. A rabbit study of maternal sevoflurane anesthesia for laparotomy produced neurodevelopmental effects, which were not prevented with the addition of xenon.

Teratogenesis has not been associated with any of the commonly used induction agents, including barbiturates , ketamine , and benzodiazepines , when administered in clinical doses during anesthesia. In a bidirectional cohort study, Bleeser et al. reported no neurodevelopmental changes in the fetus and neonate when exposed to anesthesia prenatally.

Studies of long-term sedative-hypnotic therapy have associated teratogenicity with some agents, but affirmative studies have been retrospective and have had methodologic flaws. In a prospective evaluation of children with birth defects, a higher risk for cleft lip was not found when mothers ingested diazepam during the first trimester. The present consensus among teratologists is that diazepam is not a proven human teratogen. Moreover, no evidence suggests that a single dose of a benzodiazepine (e.g., midazolam ) is harmful to the fetus.

Obstetric complications of prenatal opioid exposure are well documented and include preterm birth, small for gestational age, LBW, reduced head circumference, and sudden infant death (see Chapters 42 and 52 ). The evidence supporting structural teratogenicity of opioids in humans is of low quality, but in a systematic review of case-control and cohort studies, Lind et al. found a positive association between fetal opioid exposure and congenital anomalies (e.g., cleft lip, cleft palate, atrial septal defect, clubfoot) in 17 of 30 studies. Neonatal opioid withdrawal syndrome and impaired cognition, psychomotor skills, and behavioral development have been observed in infants and children exposed to opioids in utero . However, many studies did not correct for socioeconomic or environmental factors that might also influence neurobehavioral development.

Attempts should be made to reduce cumulative dose of opioid medications prescribed to pregnant women while ensuring adequate pain management. Analgesic and anesthetic techniques, such as regional, peripheral nerve, and field blocks, and surgical techniques, such as laparoscopy, that reduce the need for postoperative opioids should be used preferentially in pregnant women. Multimodal analgesia, including maximizing the use of nonopioid medications that are considered safe in pregnancy (e.g., acetaminophen ), should be considered.

Under certain conditions, the volatile halogenated anesthetic agents can produce teratogenic changes in chicks or small rodents. Long-term exposure to subanesthetic concentrations of halothane caused fetal growth restriction in rats but no increase in the incidence of congenital anomalies, whereas isoflurane had no adverse effects.

Significant reproductive effects have occurred with anesthetic concentrations of volatile anesthetic agents, including fetal skeletal abnormalities or death, following repeated or prolonged maternal exposure in rodents. Teratogenicity in these studies was most likely caused by the physiologic changes (e.g., profound hypothermia, hypoventilation) associated with anesthesia rather than by the anesthetic agents themselves. Moreover, some strains of mice are especially likely to demonstrate anomalies, such as a cleft palate. Mazze et al. exposed rats to 0.75 MAC of halothane , isoflurane , or enflurane , or 0.55 MAC of nitrous oxide , for 6 hours daily on 3 consecutive days at various stages of pregnancy. The animals remained conscious throughout the study, and normal feeding and sleep patterns were preserved. Under these conditions, no teratogenic effects were associated with any of the volatile agents. The only positive finding was a threefold increase in the rate of fetal resorption with nitrous oxide. No evidence has suggested reproductive toxicity with either sevoflurane or desflurane in clinical concentrations.

In contrast with the volatile halogenated agents, nitrous oxide is a weak teratogen in rodents under certain conditions, even when normal homeostasis is maintained. Rats continually exposed to 50% to 70% nitrous oxide for 2 to 6 days (starting on day 8 of gestation) had an increased incidence of congenital abnormalities. To exclude the possibility that adverse effects were a consequence of the anesthetic state, Lane et al. exposed rats to 70% nitrous oxide or to a similar concentration of xenon (a slightly more potent anesthetic devoid of biochemical effects) for 24 hours on day 9 of gestation; abnormalities occurred only in the nitrous oxide group. With the exception of one study in which extremely prolonged exposure to a low concentration of nitrous oxide had some minor effects, at least 50% nitrous oxide has been required to consistently produce anomalies. The threshold exposure time has not been rigorously determined, although exposure for at least 24 hours typically was necessary.

In vivo and embryo culture studies in rats have confirmed that nitrous oxide has several adverse reproductive effects, each of which results from exposure at a specific period of susceptibility. Fetal resorptions occurred after exposure on days 8 and 11 of gestation, skeletal anomalies after exposure on day 8 or 9, and visceral anomalies (including situs inversus ) only when exposure occurred on day 8.

Initially, teratogenicity associated with nitrous oxide was thought to result from its oxidation of vitamin B12, which interferes with its function as a coenzyme for methionine synthase. Transmethylation from methyl-tetrahydrofolate to homocysteine, with subsequent production of tetrahydrofolate (THF) and methionine, is catalyzed by methionine synthase ( Fig. 18.3 ). Thus, methionine synthase inhibition could cause a decrease in THF (with a resultant decrease in DNA synthesis) and lower methionine levels (with resultant impairment of methylation reactions). Nitrous oxide rapidly inactivates methionine synthase in both animals and humans. Prolonged human exposure to nitrous oxide leads to adverse neurologic and hematologic outcomes, the latter probably resulting from diminished DNA synthesis.

Pathway of Methionine Synthase Inhibition by Nitrous Oxide (N ₂ O) .

Metabolic consequences include impaired DNA synthesis and methylation reactions. THF , Tetrahydrofolate.

Illustration M. Fujinaga, Palo Alto, CA.

Considerable evidence indicates that methionine synthase inhibition and a consequent lack of THF are not solely responsible for the teratogenic effects of nitrous oxide. Additional studies have implicated alpha-1 adrenergic receptor stimulation in the production of situs inversus by nitrous oxide. Postulated mechanisms by which sympathetic stimulation might have adverse reproductive effects include a decrease in uterine blood flow and overstimulation of G protein–dependent membrane signal transduction pathways. There is also evidence that nitrous oxide may cause neuronal apoptosis in rats, in part, as a result of oxidative stress.

In summary, evidence suggests that the etiology of nitrous oxide teratogenicity in rats is complex and multifactorial. Determination of the relative roles of methionine deficiency and sympathetic stimulation or other mechanisms awaits further study. Although nitrous oxide is considered a weak teratogen in rats and mice, reproductive effects occur only after prolonged exposure to high concentrations that are unlikely to be encountered in humans during clinical anesthesia. Whether nitrous oxide administration is associated with neuronal apoptosis and learning impairments in humans remains to be determined.

Epidemiologic surveys dating to the early 1970s suggested that reproductive hazards (e.g., spontaneous abortion, congenital anomalies) were associated with occupational exposure to waste anesthetic agents , principally nitrous oxide , in the operating room and during dental surgery. Critical reviews of these studies questioned their conclusions, noting that response bias, inappropriate control groups, lack of verification of medical data, and exposure to multiple environmental factors made definitive conclusions impossible. By the late 1970s and 1980s, more rigorous studies had not confirmed an association between operating room work and exposure to anesthetic agents with higher reproductive risk.

Dimich-Ward et al. studied the risk for fetal anomalies in live and stillborn births in an offspring cohort study of registered nurses ( n = 22,611) in the British Columbia Vital Statistics Registry (1986 to 2000). In nurses exposed to the operating room environment, an increased risk for congenital anomalies was associated with exposure to halogenated anesthetic agents (odds ratio [OR], 1.49) or nitrous oxide (OR, 1.42). The Nurse’s Health Study-II ( n = 8461), published in 2012, compared the risk for spontaneous abortion in nurses exposed to anesthetic gases (e.g., nitrous oxide, halothane, enflurane, and isoflurane ) versus those who were not exposed, with correction for age, parity, shift work, and hours worked. The odds for abortion were not different between groups. A major limitation of these studies was the failure to correct for tobacco use or alcohol consumption; the relative risk for spontaneous abortion is increased with cigarette smoking (1.8), one or two alcoholic drinks per day (1.98), and more than three drinks per day (3.53).

Using a questionnaire-based survey of 744 pregnancies in 2028 female veterinarians, Shirangi et al. observed that the prevalence of preterm birth (before 37 weeks’ gestation) was higher in those exposed to unscavenged anesthetic gases (7.3%) compared with the general Australian population (5.7%) during the same period. The identity of the specific anesthetic gas used was not requested; however, halothane , nitrous oxide , enflurane , and methoxyflurane were commonly used during the study period. A hazards model predicted a 2.5-fold increase in preterm delivery risk in women exposed to unscavenged gas for one or more hours per week compared with an unexposed group.

It is possible that the higher waste levels of nitrous oxide encountered in dentists’ offices compared with operating rooms pose a reproductive risk, but overall the epidemiologic data do not support an increased risk for congenital anomalies with long-term occupational exposure to nitrous oxide. However, nitrous oxide is more potent than carbon dioxide in trapping heat in the ozone layer, with potential relevance to climate change.

Transient mild to moderate decreases in maternal PaO ₂ are well tolerated by the fetus, primarily because fetal hemoglobin has a high affinity for oxygen. Severe maternal hypoxemia results in fetal hypoxia and, if persistent, may cause fetal death. Any complication that causes profound maternal hypoxemia (e.g., difficult tracheal intubation, esophageal intubation, pulmonary aspiration, total spinal anesthesia, local anesthetic systemic toxicity [LAST]) is a potential threat to the fetus.

Studies of isolated human placental vessels have suggested that hyperoxia may cause uteroplacental vasoconstriction, with potential impairment of fetal oxygen delivery ; however, this concern has proved to be unfounded. Studies in pregnant women have demonstrated better fetal oxygenation with increasing maternal PaO ₂ ; however, fetal PaO ₂ never exceeds 60 mm Hg (8.0 kPa), even when maternal PaO ₂ increases to 600 mm Hg (80 kPa). This physiologic limitation exists, in part, because of the low diffusion capacity of oxygen in the placenta, as well as the admixture of blood returning to the fetus from the placenta with deoxygenated blood from the fetal inferior vena cava (see Chapter 4 ). Thus, intrauterine retrolental fibroplasia and premature closure of the ductus arteriosus cannot result from high maternal PaO ₂ . McClaine et al. observed that the maternal administration of general anesthesia for 4 hours in gravid ewes produced an initial—but not sustained—increase in fetal systemic oxygenation, but a sustained increase in cerebral oxygenation. The investigators hypothesized that the increase in fetal cerebral oxygenation resulted from greater cerebral perfusion, lower cerebral metabolic rate, or both. Histologic examination found no evidence of neurotoxicity.

Maternal hypercapnia can cause fetal acidosis because fetal PaCO ₂ correlates directly with maternal PaCO ₂ . Although mild fetal respiratory acidosis is of little consequence, severe acidosis can cause fetal myocardial depression and hypotension. Maternal hyperventilation with low maternal PaCO ₂ and high pH can adversely affect fetal oxygenation by several mechanisms. Respiratory or metabolic alkalosis can compromise maternal-fetal oxygen transfer by causing umbilical artery constriction and by shifting the maternal oxyhemoglobin dissociation curve to the left. In addition, hyperventilation, independent of changes in PaCO ₂ , may reduce uterine blood flow and cause fetal acidosis. This decrease most likely is a consequence of mechanical ventilation, whereby increased intrathoracic pressure reduces venous return and cardiac output, which in turn decreases uteroplacental perfusion. Thus, hyperventilation should be avoided in the pregnant surgical patient. Rather, the PaCO ₂ should be kept in the normal range for pregnancy.

Maternal hypotension from any cause can jeopardize uteroplacental perfusion and cause fetal asphyxia (see Chapter 3 ). The most common causes of hypotension in the pregnant patient undergoing surgery include (1) deep levels of general anesthesia, (2) sympathectomy with high sensory levels of spinal or epidural blockade, (3) aortocaval compression, (4) hemorrhage, and (5) hypovolemia. In monkeys, prolonged hypotension (i.e., systolic blood pressure < 75 mm Hg) caused by deep halothane anesthesia resulted in fetal hypoxia, acidosis, and hypotension. After as much as 5 hours of severe partial asphyxia in utero (pH < 7.00 for at least 1 hour), neonatal monkeys were depressed and experienced seizures. Postnatal survival was poor and pathologic brain changes included swelling, necrosis, and hemorrhage. The clinical course and neuropathologic findings in these animals resembled those in infants known to have suffered severe intrauterine asphyxia and who died within a few days of birth.

However, good fetal and neonatal outcomes have been reported following deliberate use of moderate degrees of hypotension, usually to facilitate performance of a neurosurgical procedure in otherwise healthy women ; maternal systolic blood pressures were maintained at 70 to 80 mm Hg, with transient periods as low as 50 mm Hg. These blood pressure parameters likely cannot be extrapolated to maternal conditions such as gestational hypertension, preeclampsia, or other conditions that predispose to uteroplacental insufficiency. In surgical circumstances necessitating maternal hypotension, the risk to the fetus must be balanced against the risk for uncontrolled maternal bleeding or stroke.

The multiple factors that influence uteroplacental blood flow are discussed in detail in Chapter 3 . Of relevance to the pregnant surgical patient are drugs that cause uterine artery vasoconstriction. Preoperative anxiety and light anesthesia increase circulating catecholamines, possibly impairing uterine blood flow. Drugs that cause uterine tachysystole (e.g., ketamine in early pregnancy in doses more than 2 mg/kg, toxic doses of local anesthetics ) may increase vascular resistance of the vessels perfusing the uterus and placenta, decreasing perfusion.

Although historically ephedrine was the preferred first-line agent for the treatment of hypotension during the administration of neuraxial anesthesia in obstetric patients, evidence now supports the use of an alpha-adrenergic receptor agonist such as phenylephrine as the drug of choice for this purpose (see Chapter 26 ). Human studies in women receiving neuraxial anesthesia, including those that directly evaluated uterine vascular resistance by Doppler ultrasonography, have demonstrated no differences in uterine blood flow parameters when ephedrine or phenylephrine infusions were used to maintain baseline mean arterial pressure. Moreover, a metaanalysis of randomized controlled trials comparing ephedrine with phenylephrine for the treatment of hypotension during spinal anesthesia for cesarean delivery concluded the following: (1) there was no difference between phenylephrine and ephedrine for the prevention and treatment of maternal hypotension, (2) maternal bradycardia was more likely to occur with phenylephrine than with ephedrine, (3) women treated with phenylephrine had neonates with higher umbilical arterial blood pH measurements than those given ephedrine, and (4) there was no difference between the two vasopressors in the incidence of true fetal acidosis (i.e., umbilical arterial blood pH < 7.20).

Ngan Kee et al. demonstrated that placental transfer was greater for ephedrine than phenylephrine, with less early metabolism and/or redistribution in the fetus. Fetal concentrations of lactate, glucose, and catecholamines were also higher with ephedrine than phenylephrine, thus supporting the hypothesis that the lower fetal pH observed with ephedrine is related to metabolic effects secondary to stimulation of fetal beta-adrenergic receptors. Recently, norepinephrine (noradrenaline) has been observed to successfully prevent and treat spinal anesthesia–induced maternal hypotension and appears as safe as other vasopressors.