Epidemiology

The incidence of amniotic fluid embolism is difficult to establish because (1) AFE is a diagnosis of exclusion, (2) there is no universally accepted definition for identifying cases of AFE, and (3) differing ascertainment methods yield divergent rates of AFE. Registry data from the United Kingdom suggest an event rate between 0.8 and 2 per 100,000 deliveries.^4,5 Cross-sectional analyses of administrative data in Australia and the United States suggest higher rates—3.3 and 7.7 per 100,000 deliveries, respectively.^6,7 A systematic review published in 2009 estimated that the pooled incidence of AFE in North America was approximately 1 : 15,200 (95% confidence interval [CI], 1 : 13,900 to 1 : 16,700), whereas the incidence in Europe was three times lower at 1 : 53,800 (95% CI, 1 : 48,800 to 1 : 59,900).⁸ This heterogeneity likely reflects variations in ascertainment procedures, rather than true differences by continent of origin. One limitation of analyses of secondary databases is that the diagnosis of AFE may not have been validated. For example, one regional surveillance system in Australia has developed the capacity to systematically review records for all cases identified from administrative data. By only counting those women who experienced one of the cardinal symptoms of AFE, with no other potential explanation, the reported incidence decreased from 6.3 to 3.3 cases per 100,000 pregnancies.⁹

Risk Factors

Maternal demographic factors such as older age and race or ethnicity have been associated with AFE in population-based studies.^5,7 Other obstetric factors such as abnormal placentation, placental abruption, eclampsia, multiple gestation, induction of labor, artificial or spontaneous rupture of membranes, and operative delivery have also been associated with AFE.^5,7,10–12 Because nonreassuring fetal heart rate (FHR) tracings can complicate AFE in labor, cesarean delivery may be a consequence, rather than cause, of intrapartum AFE. Nonetheless, a strong association between cesarean birth and postpartum AFE persists among women in the United Kingdom Obstetric Surveillance System (UKOSS) dataset (adjusted odds ratio, 8.8).⁵ The proportion of excess postpartum AFE events associated with cesarean delivery (the population-proportional attributable risk for cesarean delivery) from the UKOSS dataset was estimated to be 62%.⁵

Pathophysiology

In 1941, Drs. Steiner and Lushbaugh, two pathologists from the University of Chicago, described a case series of eight autopsies after fatal intrapartum shock.³ Examination of lung tissue from these cases revealed embolic material of squamous cells, mucin, meconium, and amorphous eosinophilic material.³ Because all of these patients were described as having tumultuous labors with stronger than usual uterine contractions, it was presumed that the forceful contractions loosened or tore the placenta and forced the emboli into the maternal circulation.³ Yet, periods of uterine tachysystole are the least likely times for maternoplacental exchange of embolic material to occur as uterine blood flow ceases.¹³ The tachysystole is probably a result of endogenous norepinephrine release and, therefore, is likely temporally related to, but not causative for, embolic material transfer.¹³

Large intravenous boluses of human meconium suspended in human amniotic fluid can precipitate cardiovascular collapse in rabbits and dogs,³ but injection of autologous amniotic fluid fails to reproduce the AFE syndrome in many animal models.¹⁴ The passage of fetal squames, lanugo hair, and mucin into the maternal pelvic vasculature appears to be a common event at term,¹⁵ and fetal material has been identified in pulmonary arterial samples aspirated from critically ill women who did not have AFE.^16–18

The exact trigger for the reaction in women with AFE is not known but may be a rare pathologic fetal antigen or a common antigen presented in an unusual way—in amount, timing, or frequency of entry into the maternal circulation.¹⁹ AFE appears to be a systemic inflammatory response associated with the inappropriate release of endogenous inflammatory mediators.²⁰ Whatever the trigger, several maternal endogenous mediators appear to play an important role in the initial reaction, including arachidonic acid metabolites (i.e., thromboxane, prostaglandins, leukotrienes, endothelins).²⁰

Approximately 40% of women in a United States national AFE registry had a history of allergy or atopy, leading some authors to suggest an anaphylactoid mechanism.¹² In support of this theory, the symptoms of AFE could be blocked by the administration of a leukotriene inhibitor in a rabbit model.²¹ However, a series of case reports now suggests that levels of tryptase and histamine are not dramatically or universally elevated among women experiencing the AFE syndrome.^22,23 Although mast cell degranulation may contribute to the pathophysiology of AFE, this effect may be a secondary product of the inflammatory cascade, rather than causal, and thus the term anaphylactoid syndrome of pregnancy may be a misnomer.²³

Other immune-mediated mechanisms have also been implicated in AFE. A case-control study demonstrated low levels of the complement components C3 and C4 among women diagnosed with AFE compared with controls, suggesting that complement activation may also contribute to the inflammatory cascade.²² Whether complement plays a primary¹⁹ or secondary role²⁴ in the AFE syndrome is unknown. A heat-stable pressor agent in meconium has been suggested as another possible mediator of AFE; in a goat model this agent caused a circulatory response similar to that seen in human AFE.²⁵

Alternatively, the hemodynamic consequences of AFE could derive from activation of the coagulation system, mediated through platelet activation. Fetal squamous cells and syncytiotrophoblasts display high concentrations of tissue factor and phosphatidylserine.^26–28 Tissue factor irreversibly aggregates platelets,²⁹ leading to platelet degranulation, which releases thromboxane, serotonin, and additional mediators that amplify the immunologic systems. Serotonin is a potent pulmonary vasoconstrictor that produces vasodilation in the systemic vasculature and may lead to early right-sided heart failure in the AFE syndrome.^27,30,31

Coagulopathy develops in the majority of women who survive the initial cardiovascular collapse.^12,32 One proposed mechanism involves tissue factor. As pregnancy progresses, increasing amounts of tissue factor, a potent procoagulant, accumulates in the amniotic fluid.^27,28 Tissue factor binds factor VII, thus activating the extrinsic pathway and triggering clotting by activating factor X, with the subsequent development of a consumptive coagulopathy.²⁷ A second possible mechanism is that amniotic fluid has a thromboplastin-like effect, which induces platelet aggregation, releases platelet factor III, and activates the clotting cascade.³³ Other components of amniotic fluid, the amniochorion, and the placenta have also been implicated in contributing to the coagulopathy. Uterine atony, whether due to a specific myometrial depressant or from uterine hypoperfusion, may exacerbate hemorrhage and consumptive coagulopathy in AFE.³⁴

There is conflicting evidence regarding whether the bleeding seen in AFE is due primarily to a consumptive coagulopathy versus massive fibrinolysis. In one in vitro study, investigators added 10 to 60 µL of clear autologous amniotic fluid to 330 µL of blood from volunteers undergoing uneventful cesarean delivery.³⁵ Thromboelastographic analysis of the specimens revealed accelerated clot formation but no evidence of fibrinolysis. However, because of the rapid and severe hypofibrinogenemia observed during AFE, some investigators have suggested that severe hyperfibrinolysis is also present.³⁶ Future work is necessary to further refine our understanding of the coagulopathy that accompanies AFE.

Clinical Presentation

Whereas the classic presentation of AFE includes acute respiratory distress, cardiovascular collapse, and coagulopathy near the time of delivery, a broad range of AFE syndromes have been described in situations in which other diagnoses were excluded (Table 39-1).^12,20,37

Currently, the UKOSS provides the most comprehensive prospective surveillance for amniotic fluid embolism in the world. All hospitals in the United Kingdom with a consultant-led maternity unit report all suspected cases of AFE, along with monthly delivery volume data, for central review and reporting.^5,38 Cases of AFE are defined using either clinical criteria that include acute hypotension, cardiac arrest, acute hypoxemia, and/or coagulopathy in the absence of any other potential explanation for the symptoms and signs observed, or pathologic evidence indicating the presence of fetal squames or hair in the maternal lungs. All women in the UKOSS registry demonstrated at least one cardinal feature of AFE, including shortness of breath, hypotension, coagulopathy, maternal hemorrhage, or premonitory symptoms (see Table 39-1).⁵

A national registry in the United States accepted voluntary submissions of medical records for patients with suspected AFE beginning in 1988. The diagnosis of AFE was confirmed based on meeting all entry criteria outlined in Box 39-1. Data for 46 confirmed cases of AFE were last published in 1995.¹²

AFE most often occurs during labor; intrapartum events comprise 56% of cases in the UKOSS registry⁵ and 70% in the U.S. registry.¹² In the U.S. registry, seizure and dyspnea were the two most common presenting symptoms in women who collapsed before delivery.¹² Maternal symptoms may precede FHR changes, as shown in Figure 39-1. Fetal bradycardia, or the abrupt onset of variable decelerations that progress to fetal bradycardia, may also herald AFE in labor.¹²

AFE can present after abdominal trauma,³⁹ after first-trimester abortion,⁴⁰ in the second trimester,⁴¹ at the time of delivery,¹² and in the postpartum period.⁴² Although acknowledging that AFE has been reported up to 48 hours postpartum,³⁴ both the U.S. and U.K. registries require the diagnosis of AFE to be made within 30 minutes of delivery.^4,12

Close inspection of hemodynamic data reveals a biphasic cardiovascular response during AFE. During the initial phase, acute pulmonary hypertension results in right ventricular dilation, a decrease in cardiac output, and ventilation-perfusion () mismatch resulting in oxygen desaturation. Early arterial blood gas analysis may demonstrate evidence of profound shunt.¹² In the U.S. registry, 11 of 17 patients with a blood gas sample drawn within 30 minutes of the acute event demonstrated an initial PaO₂ less than 30 mm Hg while breathing an FIO₂ of 100%.¹² Release of endogenous catecholamines may produce a brief period of systemic hypertension and uterine tachysystole that precedes hypotension or cardiac arrest.¹² Electrocardiographic findings are nonspecific and vary from ST-wave and T-wave abnormalities to arrhythmias or asystole. A chest radiograph may show diffuse bilateral heterogeneous or homogenous areas of opacity.

Echocardiography typically demonstrates a dilated, akinetic right ventricle, pulmonary hypertension, and a normally-contracting left ventricle with a nearly obliterated cavity.^43,44 Initially, right ventricular failure leads to right ventricular dilation, which compresses the left ventricle and impedes left ventricular filling and cardiac output.^43,44 A second phase commences when right ventricular function improves,⁴⁴ typically 15 to 30 minutes after the initial event. At this point, left ventricular failure may persist due to ischemic injury to the left ventricle,⁴⁵ or direct myocardial depression,³⁴ and is accompanied by decreased systemic ventricular resistance, decreased left ventricular stroke index, and pulmonary edema.^46,47 Women who survive to the second phase may also experience hemorrhage and disseminated intravascular coagulopathy. Laboratory analysis may reveal anemia, thrombocytopenia, prolonged prothrombin time or activated partial thromboplastin time (aPTT) or both, and decreased fibrinogen levels.^12,34 In a minority of cases there is massive hemorrhage and disseminated intravascular coagulopathy without preceding cardiopulmonary collapse.³⁷

Because many of the signs and symptoms are nonspecific, the differential diagnosis for AFE is extensive and should include nonobstetric, obstetric, and anesthetic causes (Box 39-2). Even though the time course and clinical presentation of many of these competing diagnoses are similar, only amniotic fluid embolism and placental abruption result in a relatively sudden onset of consumptive coagulopathy after maternal collapse. However, partial AFE syndromes have been described in the literature.²⁰ Therefore, the absence of coagulopathy should not exclude the possibility of AFE.

Confirmatory Tests

To date, there is no definitive test to confirm the diagnosis of AFE, although the UKOSS considers the finding of fetal material in the maternal pulmonary vasculature at autopsy to be pathognomonic for AFE.⁵ However, fetal squamous cells and trophoblasts are commonly found in the maternal circulation of healthy parturients. Furthermore, differentiating between maternal and fetal cells histologically is challenging.¹⁸ Clinicians should therefore not place invasive monitors solely for the purpose of aspirating cells of fetal origin.

Several biochemical markers have been suggested. Although some of these may be promising based on preliminary data and theoretical understanding of AFE, studies of test performance are limited by delayed sample acquisition and small sample size owing to the rarity and unpredictability of the AFE syndrome.

Several markers suggest an anaphylactic/anaphylactoid mechanism. Mast cells release tryptase and histamine during degranulation; tryptase has been used as a marker for anaphylaxis because its half-life is longer than that of histamine. Elevations in serum tryptase have been reported in some parturients with AFE.²³ However, a case series found normal tryptase and urinary histamine levels in nine women with presumed AFE.²² Pulmonary mast cell counts have also been suggested; however, this measurement can only be obtained using immunohistochemistry at autopsy and therefore is of limited applicability in the clinical setting. In one observational study, the mean pulmonary mast cell count per fixed area for parturients who died of AFE was similar to that of parturients who died of anaphylactic shock but was higher than the mean counts in both pregnant and nonpregnant control patients.⁴⁸ These data, together with the finding of minimal to no elevation in serum tryptase, suggest that pulmonary mast cell degranulation may be a secondary process in AFE.²³

Complement activation may cause mast cell degranulation, and there is some evidence supporting widespread complement activation, and depressed C3 and C4 levels, in AFE.²² However, because complement is activated during acute respiratory distress syndrome and other inflammatory states, results are not specific for AFE. Zinc coproporphyrin⁴⁹ and sialyl Tn antigen²² are two biomarkers that are components of meconium and have been associated with the AFE syndrome. Sialyl Tn is a mucinous glycoprotein that originates in the fetal gastrointestinal tract and is also associated with mucinous gastrointestinal tumors. Most recently, insulin-like growth factor–binding protein-1 (IGFBP-1) was identified as a sensitive and specific biomarker for AFE in a case-control study conducted in 13 delivery centers in France.⁵⁰ IGFBP-1 levels exceeded 104.5 µg/L in 23 of 25 women with AFE and remained below 95 µg/L in all patients with postpartum hemorrhage due to atony, thrombotic pulmonary embolism, or uncomplicated labor. A single false-positive result was found in a woman with acute fatty liver of pregnancy. Based on the median concentration of IGFBP-1 in the amniotic fluid and blood samples, the investigators estimated that 6 to 92 mL of amniotic fluid passed into the maternal circulation in women who experienced the AFE syndrome.⁵⁰

Management

Although no single intervention has been shown to reliably reverse the AFE syndrome, prompt recognition and aggressive resuscitation may improve maternal and fetal outcomes. Maternal resuscitation should focus on three priorities: (1) maintenance of oxygenation; (2) hemodynamic support; and (3) correction of coagulopathy (Box 39-3).

On diagnosis of AFE, 100% oxygen should be delivered to the patient. Given the risk for coagulopathy and hemorrhage, large-bore intravenous access is warranted. An arterial line and central venous pressure catheter may facilitate hemodynamic monitoring, blood sampling, and vasopressor administration. Transesophageal echocardiography may be useful to guide volume resuscitation and selection of appropriate vasopressor therapy.

In addition to standard resuscitative measures, other management strategies have been reported for AFE. The use of cardiopulmonary bypass, extracorporeal membrane oxygenation, continuous hemofiltration, and exchange transfusions have all been described in the literature.^5,43,51 It is speculated that these technologies may filter amniotic fluid or vasoactive mediators from the systemic circulation. Strategies for management of the early right-sided heart failure seen in AFE include inhaled nitric oxide, prostacyclin, right ventricular assist devices, and vasopressors such as vasopressin, dobutamine, and milrinone.^52,53 The use of extracorporal membrane oxygenation and intra-aortic balloon counterpulsation has also been reported for management of left-sided heart failure.⁵⁴

In AFE, intact neonatal survival is related to the time interval from the onset of maternal compromise to delivery.¹² In the event of maternal cardiopulmonary arrest, the American Heart Association recommends that delivery of the fetus should occur within 5 minutes to increase the probability of good outcomes for both the mother and her neonate.⁵⁵ Although the operating room may provide a more favorable environment for surgery and resuscitation, simulation studies have suggested that resuscitation quality and the arrest-to-delivery interval both suffer with maternal transport^56,57; therefore, strong consideration should be given to performing a bedside perimortem cesarean delivery.

Because coagulopathy will likely ensue for the majority of AFE survivors, the obstetric and anesthesia providers should activate a massive transfusion protocol as soon as AFE is suspected. Blood and component therapy should be guided by the clinical presentation (see Chapter 38). Close communication with the blood bank is paramount because large quantities of blood products may be needed. Analysis of the U.K. registry demonstrated that patients who developed a coagulopathy received between 12 and 106 units of blood products.⁴ Potential pharmacologic therapies for the coagulopathy associated with AFE include antifibrinolytic agents (e.g., tranexamic acid, aprotinin), recombinant factor VIIa (rVIIa), prothrombin complex concentrate, and fibrinogen concentrate.^36,58,59

The use of rVIIa to treat intractable hemorrhage in the AFE syndrome is controversial. The use is off-label in obstetric hemorrhage (see Chapter 38). Recombinant factor VIIa binds to tissue factor and initiates clotting via the extrinsic pathway. Although individual case reports have described improvement in hemostasis, rVIIa has been associated with thrombotic complications. A meta-analysis of 25 trials involving 3849 nonobstetric bleeding patients without hemophilia found a significant increase in arterial thromboembolic events among patients who received rVIIa (relative risk [RR], 1.45; 95% CI, 1.02 to 2.05).⁶⁰

A case-controlled study based on a systematic review of case reports of women with the AFE syndrome identified a twofold increase in risk for death or permanent disability among 16 women treated with rVIIa compared with 28 control patients who underwent surgery to control bleeding but did not receive rVIIa.⁵⁸ Among survivors, treatment with rVIIa was associated with more permanent disability (risk ratio, 4.0; 95% CI, 1.5 to 10.4), largely attributed to thrombosis in major organs.⁵⁸ An unmeasured increase in severity of disease in the treatment group compared with the control group could explain these dismal results because the population of patients who did not survive to the time of operation was excluded from the analysis. In this case-controlled study of case reports,⁵⁸ 50% of parturients treated with rVIIa survived, whereas in the 2010 UKOSS dataset,⁵ 93% of the parturients treated with rVIIa survived (n = 14). However, the UKOSS data did not report neurologic outcomes or thromboembolic complications.⁵ The differences in outcomes between these two studies may be explained in part by differing definitions of AFE and differences in case ascertainment. Only five cases in the case-controlled study⁵⁸ were peer-reviewed publications; the remainder were obtained from abstracts presented at national meetings and from registry data.

Maternal and Perinatal Outcomes

The maternal mortality ratio associated with AFE has been reported between 0.5 and 1.7 deaths per 100,000 live births.⁸ AFE accounted for 7.5% of pregnancy-related deaths in the United States between 1998 and 2005.¹ The case-fatality rate from the UKOSS data was 20% (95% CI, 11 to 32),⁵ significantly lower than the reported fatality rate of 61% in the 1988 to 1994 analysis of the U.S. AFE registry.¹² In the UKOSS dataset, patients who died of AFE were more likely to be from ethnic minority groups than were survivors, even after adjusting for confounders such as age, socioeconomic status, obesity, and parity (adjusted odds ratio [aOR], 11.8; 95% CI, 1.4 to 99.5).⁵ Similar racial/ethnic disparities were seen in the United States.¹ Improvements in the management of critically ill parturients may have contributed to the decline in the case-fatality rate observed in the two studies, but the disparity in outcomes by race or ethnicity suggests further systems improvements are necessary.

Cardiac arrest complicated 87% of AFE cases reported to the U.S. registry but only 40% of cases identified by UKOSS.^5,12 This difference may reflect more comprehensive ascertainment in the United Kingdom that captures less severe cases; alternatively, improvements in early recognition and management of AFE over the past 15 years may explain improved outcomes recently reported from the United Kingdom compared with those previously published from the U.S. registry.

Neurologic outcomes for survivors are poor. The 1995 analysis of the U.S. registry reported a 15% overall rate of intact neurologic survival, with only 8% neurologically intact after cardiac arrest.^5,12 Neonates have a high overall survival rate, with approximately 80% of infants surviving,^5,12 but also have low rates of intact neurologic survival, with only 39% of survivors neurologically intact.¹² Intact neurologic survival for infants is related to the cardiac arrest-to-delivery interval; delays of greater than 15 minutes are associated with worse outcomes.¹²

Recurrent AFE has not been reported.

Incidence

Venous thromboembolic events (VTE) in pregnancy refer to deep vein thrombosis (DVT) and pulmonary thromboembolism (PTE). The incidence of pregnancy-related thromboembolic events is 1.0 to 1.7 events per 1,000 pregnancies.^61–63 The range of reported incidences may reflect differing study designs and diagnostic criteria for thromboembolism, in addition to biologic differences among populations. Analysis of administrative data from a nationally representative sample of United States hospital admissions that included 9,058,162 pregnancy admissions and 73,834 postpartum admissions demonstrated 1.36 cases of DVT and 0.36 cases of PTE per 1,000 deliveries.⁶¹ Fifteen to 24 percent of pregnant women with an untreated DVT develop a PTE.^61,64,65

There is a fivefold greater odds of thromboembolic events during pregnancy (odds ratio [OR], 4.6; 95% CI, 2.7 to 7.8), and a 60-times greater odds in the postpartum period (95% CI, 26.5 to 135.9) than in nonpregnant patients.⁶⁶ Study results conflict as to whether there is an increased risk by trimester. A meta-analysis of 12 studies that evaluated the period of risk for DVT in pregnancy found that 21.9% of antepartum DVTs develop in the first trimester (95% CI, 17.4 to 27.3), 33.7% in the second trimester (95% CI, 28.1 to 39.8), and 47.6% in the third trimester (95% CI, 39.2% to 56.2%).⁶⁷