Preservation of Fetal Viability During Noncardiac Surgery

Should the Pregnant Patient Undergoing Surgery During Pregnancy Be Tilted?

If a pregnant patient undergoes surgery in the supine position, there is the concern of the supine hypotensive syndrome. In this position, the gravid uterus compresses the aorta, decreasing blood flow to the uterus, and the vena cava, reducing venous return. The syndrome occurs when the gestation is greater than 20 weeks. Angiographic studies were performed in term pregnant women demonstrating the compression. Femoral arterial and venous pressures were measured, with an increase in femoral venous pressure and a decrease in femoral arterial pressure when the pregnant patient was placed in the supine position. In 100 term pregnant patients, 41% of the patients had a decrease in blood pressure of 10% or more when assuming the supine position. An increase in heart rate accompanied this decrease. Adjusting the position to left lateral position returned the hemodynamics to baseline. A full lateral position was not necessary; rather, hemodynamics returned to baseline with a tilt as little as 20–30 degrees. Many argue for the greatest tilt possible. The subsequent research in lateral tilt led to the practice of tilting the patient or using a wedge for left uterine displacement during surgery in pregnant patients. Several adverse effects from supine positioning have been described such as fetal acidosis and decreased delivery of oxygen to the fetus.

Twenty healthy women with a term gestation fetus had their stroke volume, arterial pressure, and cardiac output determined in various positions. In the standing position, there is no aortocaval compression and the cardiac output is the greatest. In the left lateral position of 45 degrees, cardiac output was greatest with the left position compared with the right lateral position. In 32 term pregnant women, cardiac output, stroke volume, and heart rate were measured using bioimpedance cardiography. Cardiac output and stroke volume were the greatest when women were in left lateral tilt of 15 degrees, with cardiac output being the lowest when women lay on their back (CO 7.0 L/min versus 6.5 L/min). Using suprasternal Doppler, cardiac output was measured in 157 nonlaboring term pregnant patients who were in 0 degree, 7.5 degrees, 15 degrees, and full lateral tilt. Cardiac output increased until 15 degrees left lateral with no statically significant increase when going from 15 degrees to full lateral tilt. The degree of aortocaval compression varies among pregnant patients, with not all patients being equally affected.

The concept of aortocaval compression has been challenged. Using magnetic resonance imaging, the aorta and vena cava in term pregnant women were examined. The positions of the women were supine, 15-degree, 30-degree, and 45-degree tilt. The table was not tilted, rather a wedge foam was used. The vena cava was compressed with the patient in the supine position. The compression did not improve with a 15-degree tilt; and it did not improve until a 30-degree tilt was achieved. The only means to prevent compression of the vena cava is to have the patient tilted to 30 degrees, a position in which it is not possible for surgeons to operate. Furthermore, in the supine position, the aorta was not compressed. The results of this study suggest aortocaval compression is only caval compression in the supine position. A subsequent study confirmed these results. In women undergoing elective cesarean delivery, there was no difference in umbilical cord pH as long as blood pressure was maintained at baseline regardless of whether the patient was placed supine or with a 15-degree tilt ( Table 30.1 ).

Body weight has a negative correlation with the umbilical cord pH of the neonate born to mothers during spinal anesthesia. In a retrospective study of 5742 women undergoing elective cesarean delivery during spinal anesthesia, body mass index (BMI) had a negative correlation with umbilical cord pH. For every 10-unit increase in BMI, the umbilical cord pH decreased by 0.01 and the base deficit increased by 0.26 mmol/L. Those pregnant patients with a BMI of 40 kg/m ² or greater had a mean umbilical cord pH of 7.22. In this group, 7.7% of the morbidly obese patients had an umbilical cord pH < 7.10. Aortocaval compression may be exaggerated in obese pregnant patients. Examining anesthetic outcomes in 142 morbidly obese pregnant patients receiving spinal anesthesia revealed a greater incidence of profound hypotension in the morbidly obese pregnant patients. Also, guidelines for cardiopulmonary resuscitation for the pregnant patient recommend chest compressions be performed at a rate of 100/min at a depth of 2 inches and a compression:ventilation ratio of 30:2 with left uterine displacement being maintained manually.

Although left uterine displacement is recommended for surgery during pregnancy, it is not mandatory. If the surgeon is unable to perform the procedure with uterine displacement, it is acceptable to accomplish the surgery with the patient in the supine position as long as pressure is maintained.

Does Anesthesia During Pregnancy Have an Effect on the Fetus?

It is estimated that 1 in 500 pregnant patients will require surgery not related to the pregnancy. The greatest concern for the pregnant patient is the effect of the anesthetic agents on the fetus, both in regard to teratogenicity and to learning. All anesthetic agents have been implicated as teratogens in animal studies. However, the animal model does not replicate the clinical situation because exposure is greater than would be used clinically. For a teratogenic effect to occur, the mother must be exposed to a given level of drug for a specific period at a specific point in the gestation. To assess the teratogenicity of general anesthesia, population studies must be used.

When considering teratogenesis, surgery during the first trimester is the major focus because this time period is when the organs of the fetus are being formed. Snider and Webster were the first to study the effects of anesthesia and surgery during pregnancy. These authors evaluated the medical records of 9073 women who delivered infants between July 1959 and August 1964. Of these women, 147 women (1.6%) had surgery during pregnancy. There was no increase in congenital anomalies in the group who received surgery; note that none of the drugs used at that time are currently used. Brodsky and colleagues had a different approach, mailing a questionnaire to dentists and dental assistants to identify pregnant women who underwent surgery during pregnancy and to identify pregnant patients who had occupational exposure to nitrous oxide or volatile agents. They identified 287 women who had surgery during pregnancy. Of these women, 187 had surgery during the first trimester. A large number, 3624 women, had occupational exposure only. There was no major difference in the incidence of congenital anomalies in infants born to women who had surgery during pregnancy compared with a control group who did not have surgery. Furthermore, there was no increase in congenital anomalies in infants born to women with occupational exposure. Using health insurance data from the province of Manitoba, Duncan and colleagues confirmed the previous results. Mazze and Kallen performed the largest study by examining cases from three Swedish health-care registries for the years 1973 to 1981. These authors identified 5405 women who underwent surgery during pregnancy. Of these women, 65% received general anesthesia and 2248 had surgery during the first trimester. The authors found no increased incidence of congenital anomies. Another approach is to perform a case control study. Of the 20,830 pregnant women who had offspring with a congenital anomaly, 31 patients had surgery and anesthesia. This percentage did not differ from the 35,727 women who had babies without defects, 73 had surgery during pregnancy. There was no higher incidence of surgery and anesthesia for any congenital anomaly.

Although none of the previous studies was able to link anesthesia and surgery with a congenital anomaly when performed during pregnancy, these studies must be repeated because the choice of intravenous agents, inhalation agents, and neuromuscular blockers change. The anesthetic agents used for anesthesia currently differ from those included in the study by Mazze and Kallen. In a study of 6,486,280 pregnant women admitted between 2002 and 2012, 47,628 of these women had surgery during pregnancy. There was no increase in congenital anomalies. However, this study confirmed previously held convictions. Surgery and anesthesia during pregnancy, particularly abdominal surgeries, increased the risk of stillbirth, preterm delivery, and prolonged hospital stay. Although the risk of surgery during pregnancy is low, there is a higher risk of infection, primarily urinary tract infection.

Despite these data, the status of nitrous oxide as a reproductive toxin continues to be debated. Nitrous oxide is a teratogen in animals. It inhibits methionine synthetase, an enzyme necessary for folate metabolism. Despite the previous studies, other authors have postulated a link. Kallen and Mazze reexamined their database and noted six infants who had neural tube defects. This number is much higher than the expected number, 2.2 (an incidence of 1 per 1000 births). The authors postulated nitrous oxide as a possible cause, although the numbers and exposure do not support this proposition. Another study examined infants born with central nervous system defects in Atlanta between 1968 and 1980 who were matched to controls by race, birth hospital, and period of birth. Of the 694 mothers of infants with central nervous system defects, 12 reported first-trimester anesthetic exposure (34 of 2984 control mothers reported such exposure), yielding an odds ratio of 1.7 (confidence interval [CI], 0.8 to 33; not statistically significant). However, when examining infants with hydrocephalus and eye defects, the odds ratio increased to 39.6 (CI, 7.5 to 208.2). Although the odds ratio suggests statistical significance, it is important to examine the actual incidence. There were eight infants with this defect, three of whom had a first-trimester exposure to anesthetic agents. This result is more due to a small number and inappropriate application of statistics.

In 2016, the US Food and Drug Administration (FDA) released a statement regarding anesthetic use in children and pregnant women: “repeated or lengthy use of general anesthetic and sedation drugs during surgeries or procedures in children younger than 3 years or in pregnant women during their third trimester may affect the development of children’s brains.” The FDA then said that it is probably safe after research review. In 2017, the statement was updated to include information about neuronal loss in animal models and lengthy or multiple exposures “may negatively affect brain development.” This statement brought the potential of anesthetic-related neurotoxicity to the forefront and led many to wonder how it would affect clinical practice ( Table 30.2 ).

In 1999, Ikonomidou and colleagues reported that N-methyl-d-aspartate (NMDA) receptor blockade in developing rats induced increased neuronal apoptosis in an age-dependent fashion, with the most vulnerable time corresponding to a human 20–22-week gestational age fetus. Coupled with findings that in the fetal rat brain, alcohol, a gamma-aminobutyric acid (GABA _A ) receptor agonist, induced neuronal degeneration, the question of anesthetic-induced neurotoxicity was raised. NMDA receptor antagonism and GABA-potentiation implicated drugs such as ketamine and inhaled anesthetics as potential neurotoxins because the observed neuronal degeneration was in excess of the physiologic apoptosis that occurs in brain development. Jevtovic-Todorovic et al. anesthetized 7-day-old rats for 6 hours with isoflurane, nitrous oxide, and midazolam. Rats exposed to the three anesthetic drugs displayed deficits in learning and spatial memory. On histopathology examination of brain sections, large sections of neuronal apoptosis were identified in the anesthetic exposed group. Proposed mechanisms of this degeneration include anesthetic induced calcium overload, mitochondrial damage, and impaired myelogenesis leading to impaired synaptic transmission and potentiation.

Dexmedetomidine was investigated as a potential modulator of general anesthetic induced neuronal apoptosis. In several studies, it appeared to decrease neuronal apoptosis following isoflurane and ketamine exposure in fetal sheep and monkeys. The T-REX clinical trial is currently enrolling children under 2 years of age undergoing a general anesthetic for longer than 2.5 hours. The patients are randomized to sevoflurane or sevoflurane-remifentanil-dexmedetomidine and then undergo IQ testing at the age of 3 years to assess whether the combination anesthetic technique is associated with better neurodevelopmental outcome.

With this new evidence, researchers investigated whether the rodent model of anesthetic induced neuronal apoptosis could be similar to nonhuman primates. Newborn rhesus monkeys exposed to continuous ketamine infusions displayed increased neuronal apoptosis and persistent deficits in learning, motivation, and cognitive development. Further research in rhesus monkeys showed that multiple, but not single, exposures to isoflurane and sevoflurane were associated with increased anxiety behaviors to stressors several months later.

Applying animal data to clinical obstetric and pediatric anesthesia is challenging. In many animal models, healthy animals were anesthetized and no surgery was performed, which is quite unlike clinical practice. There is also variability between the ages of animals, anesthetic technique, and the level of monitoring used to detect hypoxia, hypercarbia, and hypotension. Clinical application of the findings is also difficult because no single structural change after anesthesia was tied to specific behavioral findings; studies identified affected regions in the cortex, hippocampus, and thalamus.

Further work on the role of general anesthesia exposure in pregnant women and children has been carried out by retrospective, ambidirectional, and randomized control trials. Early research mostly consisted of retrospective population-based studies to investigate a link between general anesthesia and diagnosis of learning disability, academic performance, or results of neuropsychological testing. Sprung and colleagues examined records of children in Olmsted County, Minnesota, to compare exposure to either general or regional anesthesia for cesarean delivery or vaginal delivery and found no difference in rates of later diagnosis of learning disability. Several large population-based cohort studies in Australia, Denmark, and the Netherlands found no differences in average academic scores; however, children exposed to anesthesia were over-represented in the lowest scoring percentiles. Among studies using clinical diagnosis as the outcome measure, results are conflicting. DiMaggio and colleagues examined a birth cohort in New York and concluded that anesthetic exposure compared with no anesthetic exposure was associated with increased risk of diagnosis for behavioral or developmental disorder. However, others found no difference in ADHD diagnosis with a single exposure but did conclude there was increased risk with multiple exposures.

Limitations of research based on diagnosis of learning disability or academic records include confounding variables such as maternal education, socioeconomic status, absenteeism, and conditions that may predispose the child to anesthetic exposure and a diagnosis of learning disability such as premature birth. Using neuropsychologic testing as a primary outcome measure allows researchers to decrease such confounding variables. Further analysis of an Australian cohort found no difference in behavior or cognitive ability among children exposed to general anesthesia aged 3–10 years. However, a smaller study found children exposed to general anesthesia under the age of 4 years scored lower in language comprehension and IQ testing and on MRI scan had lower gray matter density in the cortex than those with no exposure.

To address limitations of retrospective analysis and to study children undergoing anesthesia with modern techniques, the multisite Pediatric Ambidirectional Neurodevelopment Assessment (PANDA) study used a sibling-matched cohort of healthy children undergoing inguinal hernia repair before the age of 36 months. The PANDA study found no significant difference in verbal, spatial, executive function, and memory. When adjusted for sex, there were no differences in language scores. Final results of another ambidirectional long-term cohort study, the Mayo Anesthesia Safety in Kids, are pending because children who received anesthesia before the age of 3 years are being followed and tested until the age of 19 years. The General Anesthesia compared with Spinal anesthesia (GAS) trial is a randomized control trial that examined 722 infants under 60 weeks’ postconceptual age undergoing inguinal hernia repair. The infants received either spinal anesthesia or sevoflurane general anesthesia with the primary outcome being IQ testing, with full results being published after the study subjects undergo further testing at the age of 5 years. At 2 years old, there was no difference in five domains of testing including language, motor, social, cognitive, and adaptive.

As a result of confounding factors of variability in anesthetic technique, monitoring, socioeconomic status, sex, health status, and surgical effects, it is difficult to say definitively whether general anesthetics are neurotoxic. In addition to pediatric anesthesia, anesthesia for fetal surgery must be considered because often high doses of inhaled anesthetics are used to maintain uterine relaxation. For this reason, the International Fetal Anesthesia Database, a registry of fetal anesthetic outcomes, has been established. Randomized control trials and sibling-matched cohort studies decrease some of these confounding variables, but difficulties remain in the design of a study of healthy infants undergoing long anesthetics or to control for postprocedural sedation in an intensive care unit.

If there is an effect on neurodevelopment caused by general anesthetics, the risks of anesthesia must be weighed against the real risks of delaying a necessary procedure such as repair or palliation of congenital heart disease. Furthermore, many infants are not undergoing operations for trivial reasons and it may be riskier, for example, to delay myringotomy because chronic otitis may lead to hearing loss and a resulting potential language delay. Results from the randomized controlled trials and further research are awaited and they are expected to add more valuable information for clinicians and all those caring for pregnant and pediatric patients.

Management of Hypotension During Surgery on Pregnant Patients

The purpose of the placenta is to deliver oxygen and nutrients to the developing fetus. Uteroplacental perfusion is provided by the uterine and ovarian arteries as they join and penetrate the myometrium to form the arcuate arteries. In the nonpregnant state, uterine blood flow accounts for less than 5% of the cardiac output. During pregnancy, uterine blood flow increases progressively, reaching approximately 500 to 800 mL/min (10% of the cardiac output) at term. Uterine blood flow is directly proportional to the mean maternal arterial pressure and inversely proportional to uterine vascular resistance. Thus, it is important to maintain arterial blood pressure and to prevent increases in uterine vascular resistance. By maintaining arterial pressure, oxygen delivery to the fetus is ensured.

Wollman and Marx were the first to demonstrate the importance of avoiding hypotension when caring for pregnant patients. Their original intent was to examine the effect of crystalloid prehydration on the incidence of hypotension after spinal anesthesia. In this study, no vasopressors were administered. In patients who developed hypotension, the Apgar scores of the infants were lower, infants look longer to initiate respiration, and the umbilical cord pH was lower (a reflection of decreased uterine blood flow). These authors confirmed that hypotension is not well tolerated by the fetus. Given that maternal hypotension is bad for the fetus, the rapid return of maternal blood pressure to baseline is desirable. When treating hypotension, the anesthesiologist may use a direct agonist (such as phenylephrine, which binds directly to the alpha receptor) or an indirect acting sympathomimetic (such as ephedrine, which acts by causing the release of norepinephrine). The indirect acting drugs have both alpha and beta effects. When treating maternal blood pressure with either of these drugs, their effect on both maternal blood pressure and uterine blood flow must be considered. If the drugs result in vasoconstriction, the increase in blood pressure may not be sufficient to negate the effect on uterine artery vascular resistance.

The first study comparing different drugs for the treatment of blood pressure examined 14 pregnant ewes undergoing 80 treatments. All the sheep received both general and spinal anesthesia and uterine blood flow was measured in all animals. The authors studied metaraminol (an alpha agonist), ephedrine (an indirect acting drug), and mephentermine (an indirect acting drug). In this model, a decrease in blood pressure resulted in a direct, proportional decrease in uterine blood flow. All three drugs increased blood pressure and uterine blood flow. However, mephentermine and ephedrine increased it much more than metaraminol. The authors concluded that spinal anesthesia decreases maternal blood pressure and uterine blood flow. Indirect acting agents restore uterine blood flow to a greater extent. The clinical conclusion from this study was that if vasopressors are required, drugs whose mode of actin lies in cardiac stimulation rather than peripheral vasoconstriction should be used. This study was the one that initiated the recommendation that ephedrine should be used for the treatment of hypotension during pregnancy. In another sheep model study, uterine blood flow was unaffected with ephedrine, reduced 20% with mephentermine, and reduced 62% with methoxamine. This study confirmed the recommendation for ephedrine.

Ephedrine remained the standard of treatment for hypotension until 1988, when intravenous ephedrine was compared with intravenous phenylephrine to treat hypotension resulting from epidural anesthesia. In this study, parturients undergoing cesarean section were randomized to receive either ephedrine 5 mg or phenylephrine 100 μg to treat hypotension. An impedance cardiograph was used to measure stroke volume, ejection faction, and end-diastolic volume. Both vasopressors restored maternal blood pressure. Also, there was no difference in Apgar score or umbilical cord pH. More importantly, there was no difference in stroke volume and ejection fraction between the two medications. The major points from this study were that the cardiac effects of ephedrine were not as important and that the effects of phenylephrine were not as detrimental as previously thought.

The previous study examined cesarean section during epidural anesthesia. A similar study was conducted using spinal anesthesia. In this study, there was no difference in maternal blood pressure or neonatal Apgar score. However, the umbilical cord pH was higher in the phenylephrine group (7.33 vs 7.28). Subsequent studies comparing the two medications confirmed this finding. There was no benefit to phenylephrine but also no detriment; either vasopressor resulted in similar maternal effects on blood pressure and neonatal effects on umbilical cord pH and Apgar score in women with preeclampsia. In another study, Doppler echocardiography showed no difference in cardiac output between phenylephrine and ephedrine when used to treat hypotension occurring during spinal anesthesia.

Subsequent studies have confirmed this statistically significant, but probably clinically insignificant, difference in umbilical cord arterial pH. The hypotension alone cannot be responsible for the additional acidosis. It has been postulated that the difference is a result of ephedrine gaining access to the fetus, increasing catecholamine levels. This increase in catecholamine levels leads to an increase in oxygen consumption and an increase in lactate concentration.

A quantitative, systematic review of studies comparing phenylephrine and ephedrine was conducted. Seven randomized controlled trials were identified with a total of 292 patients. There was no difference between phenylephrine and ephedrine in the ability to correct maternal hypotension, but a higher incidence of maternal bradycardia occurred if phenylephrine was used. In regard to the neonate, there was no difference in the incidence of true fetal acidosis, but neonates whose mothers received phenylephrine had higher umbilical arterial pH values if the mother did not have preeclampsia.

Phenylephrine may be administered as intermittent bolus or as a continuous infusion. There is no difference between the two methods with regard to the effect on the neonate. A continuous infusion results in less variability in maternal blood pressure and less nausea and vomiting. The use of norepinephrine for the treatment of hypotension has been investigated. The benefit of norepinephrine is less depression on maternal heart rate and cardiac output. It seems to have no effect on the fetus and is a viable alternative. An international consensus statement was prepared on the management of hypotension during cesarean section. This statement advocates the use of α-agonists for the prevention and the treatment of hypotension. Phenylephrine is currently recommended because of the number of studies supporting its use. Ephedrine is also appropriate combined with phenylephrine if the heart rate is low.

Clearly, the choice of vasopressor for treating hypotension in pregnant patients has undergone much study and change. All of the studies have been conducted in women undergoing cesarean delivery. When discussing surgery during pregnancy, it must be assumed that the same principles apply. The initial thought that ephedrine was the better agent was based on an animal model. Subsequent study on parturients did not support this finding. In fact, a statistically significant, but clinically insignificant, difference in umbilical arterial pH was found for phenylephrine. Either drug is acceptable for the treatment of hypotension, although a continuous infusion of phenylephrine results in less maternal nausea and vomiting.