Pneumoperitoneum

When faced with providing anesthesia for the pregnant patient undergoing laparoscopic surgery, the anesthesiologist must focus not only on maternal/fetal issues and prevention of preterm labor, but also on patient positioning during surgery and the physiologic and mechanical effects of the carbon dioxide (CO ₂ ) pneumoperitoneum. During laparoscopy, pneumoperitoneum can cause cardiovascular and respiratory alterations in nonpregnant patients, which become accentuated in the parturient. Adding pneumoperitoneum to an enlarged uterus further limits diaphragm expansion and is associated with an increase in peak airway pressure, decrease in FRC, increased ventilation/perfusion (V/Q) mismatching, increased alveolar-arterial O ₂ gradient, decreased thoracic cavity compliance, and increased pleural pressure. Pneumoperitoneum and Trendelenburg positioning moves the carina cephalad, which can convert a low-lying tracheal tube to an endobronchial position. The Trendelenburg position increases intrathoracic pressure and accentuates all the respiratory-related physiologic changes.

The combination of pregnancy and CO ₂ pneumoperitoneum predisposes the parturient to hypercapnia and hypoxemia. Insufflation of CO ₂ results in CO ₂ absorption across the peritoneum and into the maternal bloodstream. Elimination depends on an increase in minute ventilation; however, mechanical hyperventilation can reduce uteroplacental perfusion, probably because of decreased venous return. Although end-tidal CO ₂ concentration (eT co ₂ , P etco ₂ ) correlates well with arterial CO ₂ tension (Pa co ₂ ) in healthy patients, these are poor guides to Pa co ₂ in sicker patients. Any increase in maternal Pa co ₂ or decrease in Pa o ₂ can affect fetal well-being. The cardiovascular changes associated with CO ₂ insufflation include reduction in cardiac index and venous return, which can be exacerbated by reverse Trendelenburg positioning. The observed increase in intracardiac filling pressures is probably secondary to an increase in intrathoracic pressure. The combination of reverse Trendelenburg position, general anesthesia, and peritoneal insufflation can decrease the cardiac index (CI) by as much as 50%.

The hemodynamic effects of aortocaval compression by the gravid uterus could further accentuate the hemodynamic effects of pneumoperitoneum and reverse Trendelenburg positioning, resulting in significant hypotension. Steinbrook and Bhavani-Shankar studied the cardiac output changes in four pregnant patients (17-24 weeks’ gestation) undergoing laparoscopic surgery using thoracic bioimpedance cardiography. IV ephedrine (10 mg) was given if SBP decreased by more than 20% with respect to baseline. The authors noted a 27% decrease in CI after 5 minutes of CO ₂ insufflation. CI remained 21% below baseline after 15 minutes of insufflation. The authors’ aggressive management of blood pressure during anesthesia (treating any decrease in BP approaching 20% of baseline measurements with IV ephedrine to minimize decreases in uterine blood flow) may have resulted in the somewhat smaller CI reduction during CO ₂ insufflation in their pregnant patients (27%) compared with 30% to 50% in nonpregnant subjects in most studies. Mean arterial pressure (MAP) and systemic vascular resistance (SVR) increased in these study subjects during CO ₂ insufflation, which is similar to that generally observed in nonpregnant subjects during laparoscopic surgery.

Monitoring

With the large number of physiologic changes associated with pregnancy, as well as the cardiovascular and pulmonary changes induced by laparoscopic surgery, optimal perioperative monitoring is unclear. The main debate is whether perioperative monitoring of arterial blood gases (ABGs) and fetal and uterine activity is necessary in parturients undergoing laparoscopic surgery. The Society of American Gastrointestinal Endoscopic Surgeons (SAGES) published guidelines for laparoscopic surgery during pregnancy that include perioperative monitoring of ABGs, as well as perioperative fetal and uterine monitoring, as echoed by other authorities. Amos et al. reported four fetal deaths in seven pregnant women who underwent laparoscopic cholecystectomy or appendectomy. During the same period, no fetal deaths occurred in patients who underwent pelvic surgeries by laparotomy. Even though no ABG data were collected, fetal demise could have resulted from prolonged respiratory acidosis, despite maintaining ET co ₂ in the physiologic range (low to mid-30s mm Hg). These concerns stem from previous studies indicating that elevation in maternal Pa co ₂ could impair fetal CO ₂ excretion across the placenta and could exacerbate fetal acidosis. Other risk factors were present for fetal loss in this series, however, including perforated appendix and pancreatitis.

Steinbrook et al. reported a case series of 10 pregnant women, gestational age 9 to 30 weeks, undergoing laparoscopic cholecystectomy; ABGs or perioperative fetal and uterine activity were not monitored. The patients underwent general anesthesia with controlled ventilation, with ET co ₂ maintained at 32 to 36 mm Hg. Fetal heart rate (FHR) and uterine activity were assessed preoperatively and immediately postoperatively. All patients had an uneventful recovery and did not need postoperative tocolysis, and no adverse maternal or fetal outcomes were noted. Seven patients were followed to delivery and had normal infants. The authors concluded that standard monitors recommended by the American Society of Anesthesiologists (ASA) are sufficient for the safety and well-being of the parturient and the fetus.

Based on a series of 45 laparoscopic cholecystectomies and 22 laparoscopic appendectomies performed during all three trimesters, Affleck et al. supported the use of noninvasive monitors and maintenance of the ET co ₂ within the physiologic range. They also recommended preoperative and postoperative FHR and uterine activity monitoring and no prophylactic tocolysis. No fetal loss or uterine injuries or spontaneous abortions occurred. There was no significant difference in preterm delivery rate, Apgar scores, or birth weights between the open and laparoscopic surgery groups. As in previous reports, the operative groups (both open and laparoscopic appendectomies and cholecystectomies) had a slightly higher rate of preterm labor compared with the general population. Furthermore, multiple case reports have reported successful outcomes with noninvasive monitoring.

Bhavani-Shankar et al. prospectively evaluated the Pa co ₂ -ET co ₂ difference in eight parturients undergoing laparoscopic cholecystectomy with CO ₂ pneumoperitoneum. The intra-abdominal pressures were maintained at about 15 mm Hg. These women underwent surgery with general anesthesia during the second and third trimester. After adjusting minute ventilation to maintain the ET co ₂ at 32 mm Hg, ABGs (alpha-stat method) were measured at fixed surgical phases: before insufflation, during insufflation, after insufflation, and after completion of surgery. The authors found no significant differences in either mean Pa co ₂ -ET co ₂ gradient or Pa co ₂ and pH during the various phases of laparoscopy. During the surgical phase the maximal Pa co ₂ -ET co ₂ difference detected was 3.1 mm Hg (range, 1.1-3.1). It appears that ET co ₂ correlates well with arterial CO ₂ , and adjusting ventilation to maintain ET co ₂ also maintains optimal maternal Pa co ₂ . These results do not support the need for ABG monitoring during laparoscopy in pregnant patients.

Laparoscopic procedures have been performed safely during all trimesters of pregnancy. However, some authors advocate reserving semielective, nonobstetric surgery during pregnancy only during the second trimester. During this period, organogenesis is complete, and spontaneous abortions are less common than in the first trimester. Furthermore, procedures during the third trimester have been associated with more preterm labor and potential difficulty in visualization with an enlarged uterus.

Anesthetic technique

Table 19-2 summarizes recommended anesthetic and surgical interventions for laparoscopy during pregnancy.

Anesthetic issues

Transvaginal ultrasound–guided oocyte retrieval (TUGOR) averages 10 to 20 minutes and can be performed with the patient under conscious sedation, paracervical block, neuraxial blockade, or general anesthesia. Therefore, short-acting agents are desired to minimize recovery time. Monitored anesthesia care and conscious sedation rely on adequate local anesthesia. However, they are inadequate to anesthetize the ovary. Patient discomfort, motion caused by pain, and a deep level of sedation leading to airway obstruction are serious risks. In addition, significant discomfort may leave patients with bad memories and may discourage future attempts at IVF. Therefore, we prefer to use neuraxial techniques or intravenous general anesthesia (IVGA).

A target-controlled propofol infusion delivered by nonanesthesiologists in the United Kingdom, initiated and supervised by a consultant anesthesiologist, required considerable medical input, especially in the early stages, to ensure the safe provision of care. Incremental alfentanil was used for analgesia. The successful use of propofol and alfentanil by patient-controlled pump was previously reported in IV sedation for egg retrieval. The main concern with propofol is distinguishing among conscious sedation, monitored anesthesia care (MAC), and general anesthesia. An anesthesiologist or nurse anesthetist under the supervision of the reproductive endocrinologist or anesthesiologist should be present at all times with the use of MAC. Therefore, it is essential to maintain verbal contact with patients during the use of conscious sedation. Even though propofol has been used by emergency medicine physicians and gastroenterologists in the United States for conscious sedation and by nonphysician providers in the United Kingdom, no data describe propofol use by nurses under reproductive endocrinologist supervision in the United States. Therefore, we discourage the practice of propofol use by nonanesthesia providers for egg retrieval.

Embryo transfer (ET) is a simple procedure that occurs on day 3 or 5 after TUGOR, relies on a fertilized oocyte, and rarely requires any anesthetic involvement. After speculum insertion into the vagina and examination of the cervix, a flexible catheter loaded with embryos and culture medium is advanced past the cervical os and injected into the uterus. Conscious sedation or MAC may be necessary in cases of significant discomfort with speculum insertion, or when there is difficulty advancing the flexible catheter past the cervical opening.

Gamete intrafallopian transfer (GIFT) is an alternative to IVF-ET that was more common before improved embryo culture techniques and successful pregnancies with IVF-ET. After hormone stimulation and TUGOR, unfertilized oocytes are mixed with sperm and transferred shortly after retrieval into the fallopian tube. Laparoscopy performed under general anesthesia is preferred so as to have direct visualization of the flexible catheter and fallopian tubes in GIFT. Although spinal anesthesia is rarely used for laparoscopic procedures because of concerns of shoulder discomfort and difficulty breathing with CO ₂ , one report highlights the safety of spinal anesthesia for laparoscopic oocyte retrieval. Another technique is performed with a minilaparoscopic approach, reducing intraperitoneal pressure and CO ₂ and obviating the need for general anesthesia. Pregnancy rates are similar between IVF-ET and GIFT; therefore the less invasive IVF-ET is more often performed. GIFT allows for the oocyte fertilization in vivo and may be acceptable for couples with religious beliefs that preclude IVF. Other transfer options include zygote intrafallopian transfer (ZIFT), pronuclear stage tubal transfer (PROST), and tubal embryo transfer (TET). Although fertilization is confirmed before embryo transfer, all these techniques require TUGOR to aspirate the follicular fluid and laparoscopically guided transfer into the fallopian tube 24 to 48 hours after fertilization. Similar pregnancy rates and the need for two procedures and anesthetics have led to a marked decline in the performance of these techniques.

In earlier reports of IVF when it was significantly longer, general endotracheal anesthesia was used with a combination of inhalation agents, with or without N ₂ O. General endotracheal anesthesia is now rarely used, except in cases of laparoscopic oocyte retrieval or when dictated by the patient’s condition. Concern about the use of N ₂ O originated from earlier reports suggesting that it had a teratogenic effect and caused fetal death in rats when used during organogenesis. In addition, lower DNA and RNA content and morphologic abnormalities were demonstrated in the embryos of pregnant rats exposed to N ₂ O during organogenesis. This potential teratogenicity has been attributed in part to the inactivation of methionine synthase. Short exposures to clinical concentrations of N ₂ O, isoflurane, and halothane had no deleterious effect on IVF and early embryonic growth up to the morula stage in the mouse. Despite the deleterious effect of N ₂ O in some rat studies, no significant differences between rates of fertilization or pregnancy were demonstrated in humans undergoing laparoscopic oocyte retrieval and isoflurane/N ₂ O or isoflurane/air general anesthesia. Inhaled agents have not been shown to possess a teratogenic or embryo effect. Furthermore, halothane can protect against N ₂ O-induced teratogenicity and spontaneous abortions in rats. In addition, higher pregnancy rates have been demonstrated in women undergoing laparoscopic PROST under isoflurane/N ₂ O compared with propofol/N ₂ O anesthesia.

Propofol is an ideal induction and maintenance agent because of its short-acting half-life and antiemetic properties. However, early reports demonstrated that propofol diffuses into follicular fluid, with greater levels observed with higher doses. Even though follicular fluid concentrations are higher in the last follicle than the first follicle, no differences were found in the ratio of mature to immature follicles or in fertilization, cleavage, or embryo cell number. In addition, a report on the use of propofol for IVGA for TUGOR of donor oocytes demonstrated a lack of negative effect on the oocyte, as evaluated by cumulative embryo scores and rates of implantation and pregnancy. Use of propofol (with N ₂ O) for transfer of fertilized embryos resulted in fewer pregnancies compared with an isoflurane, N ₂ O-based anesthesia. However, higher maternal serum concentrations were needed in this study to provide anesthesia for laparoscopic PROST compared with the use of propofol for IVF-ET procedures. Another study on mouse oocytes found that high levels of propofol in the follicular fluid may affect pregnancy rates. The use of thiopental and thiamylal for laparoscopic egg retrieval has also been associated with accumulation in follicular fluid ; a comparison of thiopental and propofol used for laparoscopic GIFT demonstrated similar pregnancy rates. A case-control study comparing propofol IVGA to paracervical block showed no difference in fertilization rates, embryo cleavage characteristics, or pregnancy rates between the two groups. Neither group received premedication; both groups received 0.5 mg of alfentanil at anesthesia induction, and the propofol group received a full induction dose (2 mg/kg), followed by a continuous infusion without additional anesthetic. The results of this study are compelling because an IVGA group was compared to a local anesthetic group without premedication. In addition, no studies demonstrate a teratogenic effect of propofol. Overall, although it may appear in follicular fluid when used for IVGA for brief IVF procedures, data support that propofol has no adverse effect on pregnancy rates.

Fentanyl, alfentanil, and midazolam, when used as premedications before TUGOR, reach low intrafollicular levels and have no effect on rates of implantation or pregnancy. The absolute concentration of intrafollicular levels is extremely low compared with plasma levels. Alfentanil had the lowest follicular fluid/plasma ratio (1:40) compared with midazolam (1:20) and fentanyl (1:10). Remifentanil, a relatively new analgesic agent, has a fast onset and a very short recovery, suitable for IVF procedures. A comparison of propofol/fentanyl anesthesia to a midazolam/remifentanil technique found a decreased need for manual ventilation and a faster recovery of function in the latter group. More patients in the propofol/fentanyl group experienced intraoperative awareness and did not enjoy the anesthetic, but time to discharge did not vary. Other studies have compared a propofol-based anesthetic with a sedative combination of ketamine and midazolam without demonstrating a difference in the recovery profile, embryo transfers, or pregnancy rates. Of note, there are sparse data on the safety of ketamine or remifentanil on ART.

Nonsteroidal anti-inflammatory drugs (NSAIDs), such as IV ketorolac, would be ideal for the acute visceral pain during and after TUGOR. However, there is reluctance to use them because prostaglandins (PGE ₂ , PGF _2α , PGI ₂ ) in the embryo and endometrium are important for implantation. Prostaglandin H synthase, also known as cyclo-oxygenase (COX), is an essential enzyme in prostaglandin synthesis and primarily localized in the endometrial epithelium. Despite these concerns, no animal or human data demonstrate any changes produced by COX inhibitors on the embryo or on implantation rates. Furthermore, implantation does not occur until 3 to 5 days after egg retrieval. Some UK centers routinely use NSAIDs without any known effects on endometrial lining or implantation rates. We prefer to use NSAIDs for egg donors or for patients with pain refractory to significant doses of opioids until further data are available. Future studies should help to clarify some of these concerns.

Nausea and vomiting are the most common complications of general anesthesia but is reduced with the use of propofol, low doses of opioids, and the avoidance of inhaled anesthetic agents. We prefer to avoid metoclopramide in patients undergoing IVF; the risk of affecting embryo implantation and a successful pregnancy is greater than its benefit in patients not at significant risk for acid aspiration syndrome. Metoclopramide, a dopamine receptor antagonist, causes elevated prolactin levels that may be associated with inhibition of pulsatile gonadotropin-releasing hormone (GnRH) secretion, a hypoestrogenic state, and ovulatory dysfunction. Although not helpful for gastric motility, ondansetron use for the treatment or prevention of nausea and vomiting is not contraindicated during IVF. Serotonergic agents, unlike serotonin (5-HT ₃ ) receptor antagonists such as ondansetron, may also increase prolactin levels. We prefer to use a neuraxial technique for patients at increased risk for postoperative nausea and vomiting or acid aspiration syndrome.

During TUGOR, a transvaginal approach is used to puncture the ovary and aspirate the follicular fluid. Both sympathetic and parasympathetic nerves supply the ovaries. Although most of the sympathetic nerves are derived from the ovarian plexus that accompanies the ovarian vessels, a minority are derived from the plexus that surrounds the ovarian branch of the uterine artery. Acute visceral pain is often diffuse in distribution, vague in location of origin, and referred to remote areas of the body. Paracervical block (PCB) has been utilized with and without conscious sedation for TUGOR to improve pain relief. PCB anesthetizes the vaginal mucosa, uterosacral ligaments, and peritoneal membrane over the pouch of Douglas. Although the ovaries are not anesthetized, their pain sensitivity is the lowest compared with the rest of the internal female genital organs. PCB with 150 mg of lidocaine reduced abdominal pain by one half compared with placebo. The linear visual analog pain score (VAPS; 0-100 mm) decreased from 43.7 to 21.2 mm when evaluated 4 hours after TUGOR. Another study demonstrated no difference in VAPS when 50 mg of lidocaine was compared with 100 and 150 mg for PCB. Assessing VAPS immediately after the procedure found median abdominal pain levels of 30 to 32 mm. Although small concentrations of lidocaine in the follicular fluid can have adverse effects in mouse oocyte fertilization and embryo development, these levels do not affect embryo implantation or pregnancy rates. PCB alone is not sufficient to provide complete analgesia because of its 10% to 15% failure rate and lack of interference with afferent sensory fibers originating from the ovarian plexus. This finding is reflected in the 2.5 times higher vaginal and abdominal pain levels with PCB alone versus PCB with the addition of conscious sedation.

Neuraxial techniques have also been used for TUGOR and are more likely to anesthetize the ovary, vaginal mucosa, and peritoneal membrane. A thoracic dermatomal level of the tenth thoracic vertebra (T10) or higher is needed to anesthetize the ovaries. Spinal anesthesia is more likely to be beneficial because of its increased reliability and fast onset. It requires minimal to no conscious sedation and can be tailored to minimize high sensory levels and motor blockade. The optimal spinal anesthetic should allow adequate surgical anesthesia with minimal side effects, a fast onset, a short recovery time, and a similar rate of successful pregnancies as with other anesthetic techniques. Earlier reports described the use of 60 mg of 5% lidocaine for spinal anesthesia, but long recovery times and the finding of transient neurologic symptoms caused some concerns. In an effort to decrease recovery times and keep patients comfortable, Martin et al. decreased the dose to 45 mg of lidocaine and evaluated the benefit of adding 10 μg of fentanyl to the spinal anesthetic. A comparison of these studies demonstrated decreased time to ambulate, void, and discharge in the lower-dose lidocaine group. The addition of fentanyl to the lidocaine resulted in improved analgesia during the procedure and, postoperatively, a decreased opioid consumption and no change in side effects or ability to ambulate, void, or be discharged. The addition of increased amounts of fentanyl to the spinal technique and surgical improvements leading to shorter egg retrieval led to a further decrease in the dose of lidocaine to 30 mg.

Although we have had a good success with the subarachnoid use of 30 mg of lidocaine combined with 25 μg of fentanyl, controversy about lidocaine and transient neurologic symptoms led to the evaluation of bupivacaine as an alternative. However, a comparison of 30 mg of lidocaine with equipotent doses of bupivacaine (3.75 mg) demonstrated a longer time to micturition and recovery with bupivacaine. Of note, patients undergoing IVF procedures demonstrate decreased serum albumin and α ₁ -acid glycoprotein levels during supraphysiologic estrogen states at oocyte retrieval. This may lead to an increased free fraction of highly protein-bound drugs such as bupivacaine. However, this may only be significant when using larger doses of bupivacaine during epidural anesthesia, which is rarely used during IVF. At our institution, spinal anesthesia even with low doses of local anesthetic and opioid is associated with longer times to voiding and discharge compared with IVGA. This finding and short surgical time led us to use IVGA as our standard anesthetic for TUGOR. Spinal anesthesia with 30 mg of lidocaine and 25 μg of fentanyl is used at the patient’s request, in those with significant gastroesophageal reflux disease (GERD) or morbid obesity, when the patient has eaten and the oocytes must be retrieved before spontaneous ovulation occurs, and when indicated because of severe side effects to IVGA, such as postoperative nausea and vomiting.

Male factor infertility is the most common form of infertility. New variations of IVF include direct sperm harvesting and a single sperm injection into the cytoplasm of the oocyte, called intracytoplasmic sperm injection (ICSI). ICSI has greatly increased pregnancy rates in patients with male factor infertility caused by low sperm counts and is often combined with direct sperm aspiration from the epididymis or testicular biopsy. Earlier reports described a more invasive, microepididymal sperm aspiration (MESA) with open surgical aspiration of the scrotum. Recent work has pioneered less invasive techniques, such as percutaneous epididymal sperm aspiration (PESA) and testicular epididymal sperm aspiration (TESA). These two techniques have been done under local anesthesia of the superior and inferior spermatic nerves and the genital branch of the genitofemoral nerve, without premedication. We prefer to use spinal anesthesia or IVGA for these procedures to minimize patient discomfort and movement.

Anesthetic issues

In vitro fertilization consists of different stages, including suppression therapy, stimulation therapy, trigger or ovulation therapy, egg retrieval, fertilization, postovulation therapy with progesterone, and embryo transfer. Therapy with leuprolide acetate (Lupron), a GnRH agonist, causes suppression of gonadotropins (follicle-stimulating hormone and luteinizing hormone) and results in a lack of production of estrogen and progesterone. Stimulation therapy is conducted with FSH- and LH-containing human menopausal gonadotropin (hMG) and causes ovarian follicle growth. Human chorionic gonadotropin (hCG) causes ovulation to occur within 36 hours, and TUGOR is performed at this time. Supplemental progesterone is given after embryo transfer.

Stimulation therapy with gonadotropins (e.g., hMG or FSH preparations) may lead to very high estrogen levels and ovarian hyperstimulation. High estrogen levels place patients at risk for thromboembolic phenomena. In its more severe form, ovarian hyperstimulation syndrome (OHSS) may lead to increased vascular permeability with leaky capillaries and findings such as weight gain, intravascular volume depletion, ascites, pleural effusions, electrolyte changes, and renal dysfunction. These patients usually experience a state of fibrinolysis, with higher fibrinogen, plasmin/α ₂ -antiplasmin, thrombin/antithrombin complexes, and d -dimer levels than women with lower estrogen levels. In addition, tissue factor increases with high estrogen levels and is a powerful trigger of the extrinsic pathway of the coagulation cascade. Treatment is usually supportive, with intravascular volume expansion, analgesics, bed rest, and thrombosis prophylaxis. More invasive methods, such as paracentesis and thoracentesis, are more helpful for relief of symptoms, such as abdominal pain and shortness of breath. Conscious sedation is often required for these procedures and should take into account the increased sensitivity to medications caused by intravascular volume contraction. Caution should be used if regional anesthetic techniques are a consideration for these patients, because of the potential for anticoagulation.

The retrieval of oocytes from the follicles is not considered an elective procedure; failure to retrieve them may lead to spontaneous ovulation and a wasted cycle. In addition, failure to empty the follicles may lead to OHSS with all its known complications. Our preference is to perform TUGOR under spinal anesthesia with minimal sedation in patients who did not follow preoperative fasting guidelines. Aspiration prophylaxis is recommended with sodium citrate and metoclopramide, despite the concerns for increased prolactin levels with metoclopramide.

Although the rate of success continues to increase with ARTs, a significant number of cycles still do not result in a live birth. Therefore, close scrutiny is expected of variables that may affect oocyte retrieval or embryo transfer. It is essential to understand the impact of different anesthetic techniques and medications on IVF and carefully study any data that links these factors with implantation or pregnancy rates. The anesthetic may have an impact on ART, and vice versa.

Anesthetic Management

An anesthesiologist should assess the patient at about 28 weeks’ gestation to determine the effect of pregnancy on various systems ( Box 19-2 ). A multidisciplinary approach should be instituted, depending on the systemic findings. Careful evaluation of airway should be performed, and the anesthetic plan should be formulated well in advance and communicated to the patient as well as the obstetrician. Regional anesthesia is most appropriate for labor and delivery. Early initiation of epidural anesthesia is recommended, to provide ample time to negate difficulties encountered during epidural placements. Continuous spinal anesthesia is a reasonable alternative. These two techniques provide satisfactory analgesia and anesthesia as needed for cesarean delivery, urgent or otherwise. More recently, ultrasound is being used to increase the success of regional anesthesia in morbidly obese pregnant women.

If general anesthesia is contemplated, an assistant is advantageous, and airway backup equipment should be available. Difficult intubation should be anticipated, and a contingency backup plan should be initiated as needed. Great care must be exercised in appropriately positioning morbidly obese pregnant women for cesarean delivery. Hypoxia may occur in the supine position, which may require elevation of the back rest of the operating table. Retraction of the pannus to facilitate surgery can result in exaggerated supine hypotensive syndrome of pregnancy, which can result in cardiac arrest. A multidisciplinary approach is the key to a successful outcome of pregnancy in morbidly obese women.

Clinical presentation

Cardiorespiratory collapse was present in most patients with amniotic fluid embolism, as seen in the Morgan series. However, the presenting symptom in 51% of patients was respiratory distress. In the remainder, the first indication of a problem was hypotension in 27%, a coagulopathy in 12%, and seizures in 10%. On the other hand, Clark et al. found that, of those women presenting before delivery, 30% had seizures or seizurelike activity, whereas 27% complained of dyspnea. Fetal bradycardia (17%) and hypotension (13%) were the next most common presenting features. Of the 13 patients who developed symptoms after delivery, seven (54%) presented with an isolated coagulopathy manifested by postpartum hemorrhage. Several additional case reports have suggested that the presentation of amniotic fluid embolus can vary with regard to timing, presenting symptoms, and subsequent course. Therefore the differential diagnosis must be considered carefully while maintaining a high index of suspicion for amniotic fluid embolism ( Box 19-3 ).

Etiology

Intact fetal membranes isolate amniotic fluid from the maternal circulation. After delivery, uterine vessels on the raw surface of the endometrium become exposed to amniotic fluid. Normally, uterine contractions are effective in collapsing these veins. Therefore, in addition to ruptured membranes, for amniotic fluid embolism to occur there must be a pressure gradient favoring the entry of amniotic fluid from the uterus into the maternal circulation. Although the placental implantation site is one potential portal of entry, particularly with partial separation of the placenta, this is otherwise unlikely if the uterus remains well contracted. On the other hand, small tears in the lower uterine segment and endocervix are common during labor and delivery and are now thought to be the most likely entry points. In support of this concept, Bastien et al. reported a case of amniotic fluid embolus where postmortem examination revealed marked plugging of both the cervical vasculature and the lungs by various amniotic fluid elements.

The misconception that amniotic fluid routinely enters the maternal circulation at delivery arose from the belief that the presence of squamous cells in the pulmonary vasculature signaled amniotic fluid entry into the maternal circulation. Studies now show that squamous cells can appear in the pulmonary blood of heterogeneous populations of both pregnant and nonpregnant patients who have undergone pulmonary artery (PA) catheterization. The presence of squamous cells is thought to have resulted from contamination by either exogenous sources during specimen preparation or epithelial cells derived from the entry site of the PA catheter. Because it is difficult to differentiate adult from fetal epithelial cells, the isolated finding of squamous cells in the pulmonary circulation of pregnant patients without amniotic fluid embolus is most likely a contaminant and not indicative of maternal exposure to amniotic fluid. Furthermore, although squamous cells may be present in both groups (clinical evidence with and without amniotic fluid embolus), only the patients with amniotic fluid embolism had evidence of other fetal debris, such as mucin, vernix, and lanugo. In these patients, squamous cells and other granular debris were frequently coated with leukocytes, suggesting a maternal reaction to foreign material. Where other occasional unidentifiable debris was detected, the material present in the patients who did not have an amniotic fluid embolism was “clearly different” from that seen in the sample.

Additional confusion centers on the importance of trophoblastic embolization to the maternal lung. Trophoblastic cells are normally free floating in the intervillous space and therefore have direct access to the maternal circulation. Thus their presence in the maternal peripheral or central vascular circulation is neither surprising nor indicative of an amniotic fluid embolism. Further evidence that amniotic fluid does not normally enter the maternal circulation can be found from autopsies of parturients who died of various complications of pregnancy. A comparison of lung specimens from 20 toxemic patients with an equal number who had clinical evidence of amniotic fluid embolus, using a specific stain for acid mucopolysaccharide, confirmed the presence of mucin in the lung secretions from all the amniotic fluid embolism patients; no section from the toxemic patients stained positive.

In summary, the presence of squamous or trophoblastic cells in the maternal pulmonary vasculature must not be equated with the entry of amniotic fluid into the maternal circulation. No evidence suggests or supports that amniotic fluid embolism is a common physiologic event.

Pathophysiology

Once the amniotic fluid enters the maternal circulation, a number of physiologic changes occur that contribute to the syndrome observed. The pathophysiology is multifactorial, and the clinical presentation will depend on the predominant physiologic aberration.

Hemodynamic changes

Animal models have suggested that severe pulmonary hypertension was the major pathophysiologic change with amniotic fluid embolism. This was believed to result from either critical obstruction of the pulmonary vessels by embolic material or pulmonary vasospasm secondary to the response of the pulmonary vasculature to fetal debris, resulting in acute asphyxiation, cor pulmonale, and in turn sudden death or severe neurologic injury. However, human hemodynamic data do not support sustained periods of pulmonary hypertension. In fact, left ventricular (LV) failure seems to be the pathognomonic feature in humans.

Clark reviewed the available hemodynamic data from the published cases of amniotic fluid embolus in humans and found only mild to moderate increases in PA pressure, whereas all patients had evidence of severe LV dysfunction. Calculation of pulmonary vascular resistance further revealed that, with one exception, all were either normal or in a range that was reflective of isolated LV failure. In an attempt to reconcile clinical and animal experimental findings, Clark proposed a biphasic model to explain the hemodynamic abnormalities that occur with amniotic fluid embolus, suggesting that acute pulmonary hypertension and vasospasm might be the initial hemodynamic response. The resulting right-sided heart failure and accompanying hypoxia could account for the cases of sudden death or severe neurologic impairment. Those patients who survive the initial phase of pulmonary hypertension, which is transient, proceed to the next stage of LV failure. Mechanisms that contribute to the later phase of LV failure include hypoxia, leftward shift of interventricular septum secondary to right-sided heart failure (resulting in a decrease in cardiac output, leading to impaired coronary artery perfusion), and the direct myocardial depressant effect of amniotic fluid itself. Endothelin, which is in amniotic fluid in abundance, has been cited as the cause of LV failure.

Other humoral factors, including proteolytic enzymes, histamine, serotonin, prostaglandins, and leukotrienes, may contribute to the hemodynamic changes and consumptive coagulopathy associated with amniotic fluid embolism. Because of the clinical resemblance to sepsis and anaphylaxis, Clark et al. suggested that the syndrome of amniotic fluid embolism is caused by anaphylactoid reaction to amniotic fluid and named the syndrome “anaphylactoid syndrome of pregnancy.” Antigenic potential can vary in individuals and therefore can lead to different grades of the syndrome. For example, women carrying a male fetus are more likely to be affected. Similarly, fluid containing thick meconium may be more toxic than clear amniotic fluid. Human data have shown that, although most patients dying of amniotic fluid emboli have had clear amniotic fluid, there is a shorter time from the initial presentation to cardiac arrest and an increased risk of neurologic damage or death in the presence of meconium or a dead fetus. Further indirect evidence for an immunologic basis is the occurrence of fatal amniotic fluid emboli during first-trimester abortions. This suggests that under the right circumstances, maternal exposure to even small amounts of amniotic fluid can initiate the syndrome. Confirming the theory of “anaphylactoid reaction” to amniotic fluid requires further research using tryptase markers.

Coagulopathy

Consumptive coagulopathy is often associated with amniotic fluid embolism. In Morgan’s review, 12% of patients presented with a bleeding diathesis, with subsequent development of a bleeding diathesis in an additional 37%. More recent reviews, however, found an even higher incidence. Clark et al. reported that 83% of the cases in the national registry had either clinical or laboratory evidence of consumptive coagulopathy. The remaining 17% died before the clotting status could be assessed by either clinical or laboratory technique. Similarly, in 15 cases of fatal amniotic fluid embolism associated with induced abortion, two patients presented with coagulopathy, and an additional 75% of initial survivors went on to develop disseminated intravascular coagulation (DIC). It now appears that amniotic fluid embolism is almost always associated with some form of DIC, with or without clinically significant bleeding. Isolated DIC causing maternal hemorrhage may be the first indication of the problem in a small number of patients. The current laboratory evidence also supports that amniotic fluid embolus is associated with coagulation changes. Harnett et al. studied the effect of varying concentrations of amniotic fluid (10-60 μL added to 330 μL of whole blood) on thromboelastographic variables and found that amniotic fluid is procoagulant even with a 10-μL study sample.

The etiology of the coagulopathy remains somewhat obscure; investigations have yielded inconclusive and contradictory results. Although amniotic fluid contains activated coagulation factors II, VII, and X, their concentrations are well below those found in maternal serum at term. On the other hand, amniotic fluid has been shown to have a direct, factor X–activating property and thromboplastin-like effect; both increase with gestational age. The thromboplastin-like effect is likely caused by substantial quantities of tissue factor in amniotic fluid. Potential sources include sloughed fetal skin and epithelial cells derived from the fetal respiratory, GI, and genitourinary tract mucosa. Tissue factor activates the extrinsic pathway by binding with factor VII. This complex in turn triggers clotting by activating factor X. Lockwood et al. speculated that once clotting was triggered in the pulmonary vasculature, local thrombin generation could then cause vasoconstriction and microvascular thrombosis, as well as secretion of vascular endothelin. This vasoactive peptide can depress both myometrial and myocardial contractility and may primarily or secondarily contribute to the hemodynamic changes and uterine atony generally associated with this syndrome.

Diagnosis

There are no diagnostic criteria to confirm the presence of amniotic fluid embolism. The differential diagnosis includes air or thrombotic pulmonary emboli, septic shock, acute myocardial infarction, cardiomyopathy, anaphylaxis, aspiration, placental abruption, eclampsia, uterine rupture, transfusion reaction, and local anesthetic toxicity. In the presence of central venous access, blood from the pulmonary vasculature should be collected using the method described by Masson. He suggested that to minimize the possibility of maternal or exogenous contamination, a more representative sample of the pulmonary microvasculature can be obtained if blood is drawn from the distal lumen of a wedged pulmonary artery catheter. After discarding the first 10 mL of blood, an additional 10 mL is drawn, heparinized, and analyzed with Papanicolaou’s method. The presence of components of amniotic fluid, including squamous cells and mucous strands, reinforces the diagnosis. Although pulmonary vasculature preparations may occasionally be contaminated by maternal squames, when squamous cells are found in large numbers in such a sample, it is clinically significant and strongly supportive of the diagnosis of amniotic fluid embolus. This is particularly true if the squamous cells are coated with neutrophils or accompanied by other fetal debris, such as mucin or hair. A more reliable method of confirming the diagnosis might center on the identification of other amniotic fluid elements in the maternal pulmonary vasculature, as opposed to squamous cells.

Recent progress in the diagnosis of amniotic fluid embolus has centered on the attempt to develop simple, noninvasive, sensitive tests using peripheral maternal blood. Kobayashi et al. studied maternal serum sialyl Tn antigen levels in four women with clinical amniotic fluid embolism and compared them to both pregnant and nonpregnant controls. Sialyl, a mucin-type glycoprotein that originates in fetal and adult intestinal and respiratory tracts, is present in both meconium and clear amniotic fluid. Using a sensitive antimucin antibody, TKH-2, the authors found no difference in the serum levels of pregnant patients throughout gestation or in the early postpartum period compared with healthy nonpregnant controls. However, the antigen levels were elevated in the amniotic fluid embolism group. Nonetheless, this test appears promising, although it needs further evaluation. Study of a second marker of diagnosis that involves the measurement of zinc coproporphyrin, a characteristic meconium component, found plasma concentrations were higher in patients with amniotic fluid embolus.

Management

The management of the pregnant patient with amniotic fluid embolism is basically symptomatic and directed toward the maintenance of oxygenation, circulatory support, and correction of coagulopathy. Depending on the circumstances, a full cardiopulmonary resuscitation (CPR) protocol may be required. If the fetus is sufficiently mature and is undelivered at the time of cardiac arrest, cesarean delivery should be instituted as soon as possible.

Treatment of hemodynamic instability includes optimization of preload with rapid volume infusion. Direct-acting vasopressors may be required in restoring aortic perfusion pressure in the initial stages. Once this is attained, other inotropes such as dopamine and dobutamine can be added to improve myocardial function. When clinically feasible, PA catheterization can be instituted to help guide therapy. Diuretics may be required to mobilize pulmonary edema fluid. Treatment of the coagulopathy associated with amniotic fluid embolus involves blood component therapy. Amniotic fluid embolism is associated frequently with massive hemorrhage, requiring replacement with packed red blood cells (PRBCs). O-negative or group-specific blood can be used if crossmatched blood is unavailable. Plasma and platelets are given to replace the clotting factors. Ongoing therapy is generally guided by the clinical condition of the patient and laboratory evidence of coagulopathy.

Although cryoprecipitate is not first-line therapy for treating coagulopathy, it may be useful in circumstances in which fibrinogen is low and volume overload is a concern. Cryoprecipitate was also found useful in a patient with severe acute respiratory distress syndrome (ARDS) secondary to amniotic fluid embolus. After administration of cryoprecipitate, the patient’s cardiopulmonary and hematologic status improved dramatically, suggesting that cryoprecipitate may be useful when conventional medical therapy appears unsuccessful in maintaining blood pressure, oxygenation, and hemostasis. The recommendation was based on similar treatment protocols for severely ill patients with multiple trauma, burns, and postoperative sepsis. These patients are believed to have impaired clearance of circulating microaggregates and immune complexes by the reticuloendothelial system, leading to cardiopulmonary insufficiency and DIC. Cryoprecipitate is rich in opsonic α ₂ surface-binding glycoprotein, also known as fibronectin, which facilitates the reticuloendothelial system in the filtration of antigenic and toxic particulate matter. Depleted levels of this glycoprotein have been reported in severely ill patients, with marked improvement in the clinical status after repletion of fibronectin levels.

Isolated reports exist for other modalities of treatment for amniotic fluid embolism. In one patient, a serine proteinase inhibitor, FOY, was used to treat associated DIC. Nitric oxide and aerosolized prostacyclin have been used to treat refractory hypoxemia. Clark et al. suggested the use of high-dose corticosteroids and epinephrine as useful therapeutic adjuvants, in light of the similarities of amniotic fluid embolism to anaphylaxis. In management of coagulopathy of amniotic fluid embolism, Leighton et al. recently determined that patients receiving recombinant factor VIIa had significantly worse outcomes than cohorts who did not receive rVIIa, thus recommending that rVIIa be used in patients only when the hemorrhage cannot be stopped by massive blood component replacement. Recent evidence from clinical case reports also suggests that institution of cardiopulmonary bypass (CPB) may improve the outcome of patients with severe amniotic fluid embolism. Use of transesophageal echocardiography (TEE) during circulatory collapse confirmed the presence of intracardiac embolus, leading to CPB and emergent thrombectomy.

Management of eclampsia

Supplemental oxygen should be delivered immediately during seizure. A soft nasopharyngeal airway may facilitate oxygenation during seizure. Ventilation may be assisted once seizures end. Simultaneously, precautions should be observed to minimize chances of gastric aspiration. Midazolam in incremental doses up to 20 mg may be necessary, either to suppress seizures or to facilitate further treatment in a combative patient. Occasionally, propofol and succinylcholine may be required to facilitate oxygenation and ventilation. Immediate monitoring should include pulse oximetry, electrocardiogram (ECG), and blood pressure (BP) recordings. Left uterine displacement should be maintained throughout the resuscitative effort and until delivery of the infant.

Magnesium sulfate is the preferred drug for the definitive treatment of seizures. After an immediate loading dose of 4 to 6 g infused intravenously (IV) over 20 to 30 minutes, a maintenance dose of 1 to 2 g/hr is initiated, assuming that the patient has adequate urine output. Hourly monitoring of urine output, regular evaluation of deep tendon reflexes, and observation of respiratory rate should be implemented to guard against magnesium toxicity.

Unless otherwise contraindicated, eclamptic patients should undergo expeditious delivery. The common indications for cesarean delivery include fetal distress, placental abruption, prematurity with an unfavorable cervix, persistent seizures, and persistent postictal agitation.

Regional anesthesia to facilitate labor and delivery can be considered in patients who are seizure free, conscious, and rational in behavior, with no evidence of increased intracranial pressure (ICP) and absence of coagulopathy. Moodley et al. found no difference in maternal and neonatal outcomes when comparing epidural anesthesia with general anesthesia for cesarean delivery in conscious women with eclampsia. Unconscious or obtunded patients, or those with evidence of increased ICP, should have general anesthesia in line with neurosurgical anesthesia recommendations. Hyperventilation can be initiated soon after the delivery of the infant to minimize the effect of low Pa co ₂ on the uterine arteries. The patient can be extubated at the conclusion of surgery if awake and conscious. On the other hand, if general anesthesia was undertaken in a woman who was not conscious initially, consider leaving the patient intubated and transferring her to intensive care for BP control and controlled weaning from assisted ventilation while assessing neurologic recovery. Prolonged unconsciousness should prompt further evaluation with CT. Magnesium should be continued until BP normalizes and central nervous system (CNS) hyperexcitability disappears.