Epidemiology

The incidence of placenta previa is 4.0 per 1000 pregnancies.^23,24 The exact cause is unclear, but prior uterine trauma (e.g., scar from prior cesarean delivery) is a common element. The placenta may implant in the scarred area, which typically includes the lower uterine segment. Conditions associated with placenta previa include multiparity, advanced maternal age, smoking history, male fetus, previous cesarean delivery or other uterine surgery, and previous placenta previa.²³ The presence of placenta previa increases the likelihood that the patient will require a peripartum hysterectomy.²⁵

Diagnosis

Transvaginal ultrasonography has become the gold standard for diagnosis of placenta previa; the distance from the placental edge to the internal os is measured and predicts the likelihood of antepartum hemorrhage and need for cesarean delivery.^26,27 Advances in ultrasonography have made the double setup examination (i.e., vaginal examination with all personnel ready for immediate cesarean delivery) nearly obsolete in modern obstetric practice. Magnetic resonance imaging (MRI) is also useful for the diagnosis of placenta previa, but its use is not practical in most cases of antepartum hemorrhage.

The classic clinical sign of placenta previa is painless vaginal bleeding during the second or third trimester. The first episode of bleeding typically occurs preterm and is not related to any particular inciting event. The lack of abdominal pain and/or absence of abnormal uterine tone helps distinguish this event from placental abruption. The absence of these factors does not exclude abruption, however, and patients with placenta previa are at risk for coexisting placental abruption.²⁴

Obstetric Management

Obstetric management is based on the severity of vaginal bleeding and the maturity and status of the fetus. Active labor, persistent bleeding, a mature fetus (≥ 36 weeks’ gestational age), or nonreassuring fetal status should prompt delivery.²⁸ The fetus is at risk from two distinct pathophysiologic processes: (1) progressive or sudden placental separation that causes uteroplacental insufficiency and (2) preterm delivery and its sequelae. The first episode of bleeding characteristically stops spontaneously and rarely causes maternal shock or fetal compromise. Expectant management in the hospital has been shown to prolong pregnancy by an average of 4 weeks after the initial bleeding episode.²⁸ Maternal vital signs are assessed frequently, and the hemoglobin concentration is checked at regular intervals. Fetal evaluation involves frequent performance of a nonstress test or biophysical profile, ultrasonographic assessment of fetal growth, and fetal lung maturity studies as indicated. Hemorrhage may be prevented by limitations on physical activity and avoidance of vaginal examinations and coitus, although the evidence supporting these measures is limited.

Outpatient management has resulted in good outcome in carefully selected patients.²⁹ Outpatient management is reserved for stable patients without bleeding in the previous 48 hours who have both telephone access and the ability to be transported quickly to the hospital. Expectant management requires immediate access to a medical center with 24-hour obstetric and anesthesia coverage and a neonatal intensive care unit.²⁸

In most cases of placenta previa diagnosed between 24 and 34 weeks’ gestation, a corticosteroid (e.g., betamethasone) is administered to accelerate fetal lung maturity.²⁸ A significant number of patients with placenta previa have preterm labor, which may provoke bleeding. Obstetricians may administer tocolytic therapy to decrease preterm contractions, with the goal to stabilize antepartum bleeding. Ritodrine has been shown to prolong pregnancy in women with placenta previa, but no studies have confirmed any decrease in the frequency or severity of vaginal bleeding.^28,30 Obstetricians must balance the potential cardiovascular consequences of tocolytic therapy in the event of maternal hemorrhage against the consequences of preterm delivery. Tocolytic therapy is not recommended for patients with uncontrolled hemorrhage or those in whom placental abruption is suspected. Although expectant management reduces the risk for prematurity, it does not eliminate it, and prematurity remains the most common cause of neonatal mortality and morbidity, especially if bleeding begins before 20 weeks’ gestation.³¹

Fetuses of women with placenta previa may be at risk for other complications, including asymmetric fetal growth restriction (also known as intrauterine growth restriction).³² Several factors may account for the association between placenta previa and fetal growth restriction. First, the lower uterine segment may be less vascular than normal sites of placental implantation. Second, the placenta often is adherent to an area of fibrosis tissue. Third, patients with placenta previa have a higher incidence of first-trimester bleeding, which may promote a partial placental separation, reducing the surface area for placental exchange. Fourth, although the blood loss from placenta previa is almost entirely maternal, trauma to the placenta with vaginal examination or coitus may result in some fetal blood loss, which could retard fetal growth.³² Some studies have reported a higher incidence of congenital anomalies in the fetuses of women with placenta previa.²⁰

Anesthetic Management

All patients admitted with antepartum vaginal bleeding should be evaluated by an anesthesia provider on arrival. Special consideration should be given to the airway examination, intravascular volume assessment, and history of previous cesarean delivery or other procedures that create a uterine scar. Volume resuscitation should be initiated using a non–dextrose-containing balanced salt solution (e.g., lactated Ringer’s, normal saline). Women with placenta previa may remain hospitalized for some time prior to delivery, and at least one intravenous catheter should be maintained if bleeding is recurrent or imminent delivery is anticipated. For women without recurrent bleeding, consideration may be given to either deferring venous access entirely or placing a peripherally-inserted central catheter (PICC) for drug administration so that other peripheral veins are preserved for large-bore intravenous catheter access in the event of hemorrhagic emergency. Hemoglobin concentration measurement may be indicated after a bleeding episode. A blood type and screen, and for women who are actively bleeding, a blood type and crossmatch, should be maintained. The American Association of Blood Banks (AABB) recommends repeating such tests every 3 days in pregnant women owing to the small but finite risk for developing a new alloantibody during pregnancy.³³ This recommendation is written into the U.S. Code of Federal Regulations.³⁴ The use of lower-extremity sequential compression devices may decrease the risk for venous thromboembolism in patients on bed rest. Pharmacologic prophylaxis is not commonly used because of the risk for bleeding.

Epidemiology

Placental abruption complicates 0.4% to 1.0% of preg­nancies, and the incidence is increasing, particularly among African-American women in the United States.^21,42–44 The causes are not well understood, but several condi­tions are known risk factors for abruption (Box 38-1).^42,43 Patients hospitalized for both acute and chronic respiratory diseases are at risk for placental abruption for unclear reasons.⁴⁵

Diagnosis

The classic presentation of abruption consists of vaginal bleeding, uterine tenderness, and increased uterine activity; however, not all of these symptoms are always present. In cases of concealed abruption, vaginal bleeding may be absent and gross underestimation of maternal hypovolemia can occur. Bleeding may be painless.⁴¹ In some cases, abruption may manifest as idiopathic preterm labor. Patients may have a variety of nonreassuring fetal heart rate (FHR) patterns, including bradycardia, late or variable decelerations, and/or loss of variability.⁴¹ The diagnosis of placental abruption is primarily clinical, but in a subset of cases, ultrasonography may help confirm it. Ultrasonography is highly specific for placental abruption (96%), but it is not very sensitive (24%).⁴⁶ It is also useful for determining placental location, which can exclude placenta previa as a cause of vaginal bleeding.⁴¹ The ultrasonographic examination can ascertain whether retroplacental or subchorionic hematoma is present. Normal findings do not exclude the diagnosis of placental abruption.

Pathophysiology

Complications of placental abruption include hemorrhagic shock, coagulopathy, and fetal compromise or demise. One third of coagulopathies in pregnancy are attributable to abruption, and coagulopathy is associated with fetal demise.⁴⁷ Placental tissue displays tissue factor and other procoagulant substances on cell membranes, and it is surmised that when bleeding at the decidual-placental interface (i.e., abruption) occurs, these thromboplastic substances are released into the central circulation, resulting in consumptive coagulopathy and disseminated intravascular coagulation (DIC).⁴⁸

Although some cases of abruption occur acutely (e.g., in the setting of trauma), many abruptions complicate chronic, long-standing placental abnormalities. Investigators have noted strong associations between abruption, fetal growth restriction, and preeclampsia, and all three conditions share similar risk factors.⁴⁹ Naeye et al.⁵⁰ prospectively studied more than 53,000 deliveries and found that decidual necrosis at the placental margin and large placental infarcts were the most common abnormalities among patients who suffered placental abruption and fetal demise. Infants who died had 14% less placental weight, 8% less body weight, and 3% shorter body length than surviving control infants of the same gestational age. In addition, histologic evidence of shallow trophoblastic invasion of the spiral arteries supports the conclusion that “ischemic placental disease” may underlie chronic placental hypoxia, leading to preeclampsia, fetal growth restriction, and abruption.⁴⁹

The major risks for the fetus are hypoxia and prematurity. Fetal oxygenation depends on adequate maternal oxygen-carrying capacity, uteroplacental blood flow, and transplacental exchange. Separation of all or part of the placenta reduces gas exchange surface area and can lead to fetal death. The risk for intrauterine fetal demise increases as the detachment area increases, particularly when the location of bleeding is retroplacental rather than subchorionic.^41,51,52 Inadequate transplacental oxygen exchange is exacerbated by maternal hypotension, which decreases uteroplacental blood flow. Ananth and Wilcox²¹ reviewed outcomes for 7.5 million pregnancies in the United States; the perinatal mortality rate associated with placental abruption was 12%. The high mortality rate is due in large part to the fact that infants of mothers with placental abruption are five times more likely to be delivered preterm.²¹

Obstetric Management

If the diagnosis of abruption is suspected, the practi­tioner should insert a large-bore intravenous catheter and obtain blood for assessment of hematocrit, coagulation status, and type and crossmatch. When assessing volume status, the clinician must remain aware of the possibility of hemorrhage concealed behind the placenta. Placement of a urethral catheter to monitor urine output may help the physician assess the adequacy of renal perfusion. The definitive treatment is delivery of the infant and placenta, but the degree of maternal and fetal compromise and estimated gestational age determine the timing and route of delivery.⁴¹ If the fetus is at or near term and both maternal and fetal status are reassuring, vaginal delivery may be appropriate. If the patient is preterm, the extent of abruption is minimal, and the mother and fetus show no signs of compromise, the patient may be hospitalized and the pregnancy allowed to continue to optimize fetal maturation. The obstetrician may administer a corticosteroid to promote fetal lung maturity. If the mother develops hemodynamic instability or coagulopathy, or fetal status becomes nonreassuring, urgent cesarean delivery may become necessary. Vaginal delivery is preferred for patients with intrauterine fetal demise.⁴¹

Anesthetic Management

The anesthesia provider should consider the severity of the abruption and the urgency of delivery in planning anesthetic management.

Epidemiology

Rupture of the gravid uterus can be disastrous for both the mother and the fetus. Fortunately, it does not occur often. Previous uterine surgery (e.g., cesarean delivery or myomectomy) increases the risk, but the incidence of true uterine rupture after cesarean delivery is still low, occurring at a rate of less than 1%.^54,55 Uterine rupture is rare in the primigravid woman or the woman with an unscarred uterus, but it does occur.⁵⁵ Box 38-2 lists additional conditions that have been associated with uterine rupture.^55,56 Very rarely, uterine rupture occurs without explanation.⁵⁷

Rupture of a previous uterine scar may occur in the absence of labor. In a review of records from a large multistate hospital system, nearly half of all true uterine ruptures occurred in the absence of a history of cesarean delivery, and 22% of ruptures occurred in the absence of labor.⁵⁸ Lydon-Rochelle et al.⁵⁹ undertook a population-based retrospective analysis of more than 20,000 women who had undergone one previous cesarean delivery. The risk for rupture among nonlaboring women was 1.6 per 1000. Among women in spontaneous labor the risk increased approximately threefold to 5.2 per 1000; among women undergoing induction of labor the risk increased nearly fivefold to 7.7 per 1000; and among women undergoing prostaglandin induction the risk increased almost 16-fold to 24.5 per 1000. Additional risk factors for uterine rupture during a trial of labor after cesarean (TOLAC) include post-term gestation (≥ 42 weeks), birth weight greater than 4000 g, maternal age older than 35 years, and maternal height greater than 164 cm.⁵⁴

Because of variation in nomenclature and severity, accurate determination of maternal and fetal morbidity secondary to uterine rupture is difficult. The most common variety of uterine scar disruption is separation or dehiscence, some cases of which are asymptomatic. Uterine scar dehiscence is defined as a uterine wall defect that does not result in excessive hemorrhage or FHR abnormalities and does not require emergency cesarean delivery or postpartum laparotomy. In contrast, uterine rupture, less common than dehiscence, refers to a uterine wall defect with maternal hemorrhage and/or fetal compromise sufficient to require emergency cesarean delivery or postpartum laparotomy.

The rupture of a classical uterine incision scar (a vertical incision involving the muscular uterine fundus) is associated with greater morbidity and mortality than rupture of a low transverse uterine incision scar because the anterior uterine wall is highly vascular and may include the area of placental implantation. Lateral extension of the rupture can involve the major uterine vessels and is typically associated with massive bleeding. Maternal death secondary to uterine rupture is rare, although there were three deaths attributed to uterine rupture in the 2006 to 2008 triennial report from the United Kingdom.⁵ In Sweden between 1983 and 2001, Kaczmarczyk et al.⁵⁴ estimated that the neonatal mortality rate associated with uterine rupture was approximately 5%.

Diagnosis

The variable presentation of uterine rupture may cause diagnostic difficulty. Abdominal pain and an abnormal FHR pattern are the two most common presenting signs of uterine rupture,⁶⁰ but neither is 100% sensitive. One retrospective study reported the occurrence of abdominal pain in 17% of patients; an FHR abnormality was the first sign of uterine rupture in 87% of patients (see Chapter 19).⁶¹ Other presenting signs include vaginal bleeding, uterine hypertonia, cessation of labor, maternal hypotension, loss of the fetal station, decrease in cervical dilation, or a change in fetal presentation. Breakthrough pain during neuraxial labor analgesia may also indicate uterine rupture.⁶²

Obstetric Management

Treatment options for uterine rupture include repair of the uterus, arterial ligation, and hysterectomy. Uterine repair is appropriate for most cases of separation of a prior low transverse uterine scar and for some cases of rupture of a classical incision. However, the risk for rupture in a future pregnancy remains. A disadvantage of arterial ligation is that it may not control the bleeding and may delay definitive treatment. Hysterectomy may be required for some cases of uterine rupture.⁵⁶

Anesthetic Management

Patient evaluation and resuscitation are initiated while the patient is being prepared for emergency laparotomy. If rupture has occurred antepartum, fetal compromise is likely. General anesthesia is often necessary, except in some stable patients with preexisting epidural labor analgesia. Aggressive volume replacement is essential, and transfusion may be necessary. Urine output should be monitored. Invasive hemodynamic monitoring may be appropriate if there is uncertainty about the intravascular volume status.

Epidemiology

Vasa previa occurs rarely (1 in 2500 to 1 in 5000 deliveries).²⁸ Because it involves the loss of fetal blood, vasa previa is associated with a high fetal mortality rate (nearly 60% if vasa previa is unrecognized).⁶³ The blood volume of the fetus at term is approximately 80 to 100 mL/kg. Therefore, the amount of blood that can be lost without fetal death is small. In addition, the vulnerable fetal vessels may be compressed by the fetal presenting part, resulting in fetal hypoxia and death. Risk factors for vasa previa include the presence of placenta previa or low-lying placenta in the second trimester, placental accessory lobes, in vitro fertilization, and multiple gestation.²⁸

Diagnosis

Ultrasonography can be used to visualize the velamentous insertion of the vessels, and transvaginal color Doppler imaging can confirm the diagnosis.^28,63 Vasa previa should be suspected whenever bleeding occurs with rupture of membranes, particularly if the rupture is accompanied by FHR decelerations or fetal bradycardia. Hemorrhage can also occur without rupture of membranes, making the diagnosis more difficult. Rarely, vasa previa can be diagnosed via digital cervical examination or amnioscopy. The diagnosis of vasa previa can be confirmed through examination of the shed blood for evidence of fetal hemoglobin (e.g., Kleihauer-Betke test); however, when bleeding occurs, the emergency nature of vasa previa usually precludes such diagnostic confirmation.

Obstetric Management

Prenatal diagnosis confers a neonatal survival benefit. Oyelese et al.⁶³ conducted a retrospective study of 155 pregnancies complicated by vasa previa. Neonatal mortality was 3% when the vasa previa was diagnosed antenatally and 56% when it was not. The authors recommended ultrasonographic examination with transvaginal color Doppler in patients at risk for vasa previa. The management of vasa previa is directed solely toward ensuring fetal survival. Some authors advocate hospitalization of the patient between 30 and 32 weeks’ gestation to ensure prompt delivery if rupture of membranes should occur; consideration should be given to the administration of a corticosteroid to promote fetal lung maturity.²⁸ Timing of delivery reflects a balance between the risks of preterm delivery and the risk for vessel rupture if the pregnancy is allowed to continue. Robinson and Grobman⁶⁴ compared delivery timing strategies for women with vasa previa and calculated that the best fetal outcomes occurred with elective delivery between 34 and 35 weeks’ gestation. They further determined that confirmation of fetal lung maturity via amniocentesis was not necessary.⁶⁴

Ruptured vasa previa is a true obstetric emergency that requires immediate delivery of the fetus, almost always by the abdominal route. Neonatal resuscitation requires immediate attention to neonatal volume replacement with colloid, balanced salt solutions, and blood.

Anesthetic Management

The choice of anesthetic technique depends on the urgency of the cesarean delivery. In many cases, general anesthesia is necessary for prompt delivery.

Epidemiology

Uterine atony is the most common cause of severe postpartum hemorrhage, accounting for approximately 80% of cases.^10,11 In addition to normal hemostatic mechanisms, postpartum hemostasis involves the release of endogenous uterotonic agents—primarily oxytocin and prostaglandins—that contract the uterus and constrict uterine vessels. Uterine atony represents a failure of this process. In addition, parturients with obstetric hemorrhage may have uterine arteries that are relatively unresponsive to vasoconstrictor substances.⁷⁶ Box 38-3 lists conditions associated with uterine atony.

Diagnosis

A soft, poorly contractile uterus and vaginal bleeding are the most common findings in patients with uterine atony. The absence of vaginal bleeding does not exclude this disorder because the atonic, engorged uterus may contain more than 1000 mL of blood. Unrecognized bleeding may manifest initially as tachycardia; worsening hypovolemia eventually leads to hypotension (see Table 38-1).

Obstetric and Anesthetic Management

Anesthetic Management

Choice of anesthetic technique for the repair of genital lacerations and evacuation of pelvic hematomas depends on the affected area, surgical requirements, volume/hemodynamic status of the patient, and urgency of the procedure. Local infiltration and a small dose of intravenous opioid suffice for drainage of some vulvar hematomas; however, repair of extensive lacerations and drainage of vaginal hematomas require significant levels of analgesia or anesthesia. Pudendal nerve block may not be technically feasible because of anatomic distortion or severe pain from the hematoma. Low doses of ketamine (10-mg boluses, not exceeding a total dose of 0.5 mg/kg) may suffice to produce sedation and analgesia with minimal risk for altering airway reflexes. Spinal or epidural anesthesia may be necessary, although the clinician should exercise caution initiating (or extending) a neuraxial block in a hypovolemic patient. In some cases, general anesthesia with tracheal intubation may be necessary. Exploratory laparotomy for a retroperitoneal hematoma typically requires the administration of general anesthesia.

Obstetric Management

Treatment of retained placenta during the early postpartum period often involves manual removal and inspection of the placenta. Curettage may be required. After removal of the placenta, uterine tone should be augmented with oxytocin and the patient should be observed for evidence of recurrent hemorrhage.

Anesthetic Management

Choice of anesthetic technique depends on the degree of hemorrhage. In some cases, the administration of small amounts of sedatives and analgesics is adequate to allow examination and manual placental extraction by a skilled obstetrician. Neuraxial anesthesia may be considered in patients who are not bleeding severely and who are hemodynamically stable. This may be accomplished with either administration of additional local anesthetic through an existing labor epidural catheter or initiation of spinal anesthesia. General anesthesia sometimes becomes necessary, particularly in patients who are hemodynamically unstable.

In some cases, the obstetrician requires uterine relaxation to facilitate manual removal of the placenta. Historically, anesthesia providers have performed rapid-sequence induction of general anesthesia, followed by the administration of a high dose of volatile halogenated agent to relax the uterus. Equipotent doses of halothane, sevoflurane, and desflurane depress uterine contractility equally and in a dose-dependent manner. In a study of isolated human uterine muscle, an equi-anesthetic concentration of isoflurane was a less effective uterine relaxant than the other volatile agents.¹²³ Uterine contractility is decreased by 50% with administration of approximately 1.5 minimum alveolar concentration (MAC) of a volatile anesthetic agent.¹²³ The induction of general anesthesia and administration of a volatile halogenated agent results in rapid onset of uterine relaxation, and discontinuation of the volatile agent results in rapid offset when uterine relaxation is no longer necessary. However, induction of general anesthesia in a parturient entails risk for failed ventilation, failed tracheal intubation, and/or aspiration of gastric contents.

Alternatively, nitroglycerin may be administered for uterine relaxation. Nitroglycerin provides a rapid onset of reliable smooth muscle relaxation and a short plasma half-life (2 to 3 minutes).^124,125 Nitroglycerin has been administered for various obstetric emergencies without clinically significant side effects.¹²⁵ Peng et al.¹²⁶ described successful removal of retained placenta in 15 parturients after administration of intravenous nitroglycerin 500 µg. DeSimone et al.¹²⁷ used a substantially smaller dose of nitroglycerin (50 to 100 µg) with similar results; all patients were managed successfully without the need for induction of general anesthesia. Nitroglycerin may also be administered sublingually via spray or tablet. A double-blind, randomized, controlled study compared sublingual nitroglycerin with placebo for management of retained placenta¹²⁸; the placenta was delivered successfully within 5 minutes in all 12 of the parturients who received nitroglycerin, compared with only 1 of the 12 who received placebo. Nitroglycerin most likely produces uterine smooth muscle relaxation by releasing nitric oxide; it may require the presence of placental tissue to be effective.¹²⁹

Epidemiology

Uterine inversion, or the turning inside-out of all or part of the uterus, is a rare but potentially disastrous event. It is associated with severe postpartum hemorrhage, and hemodynamic instability may be worsened by concurrent vagal reflex–mediated bradycardia. Inversions may be acute or chronic, but only acute peripartum inversions involve the obstetric anesthesia provider. The reported incidence of this disorder varies widely; recent reports suggest an incidence of approximately 1 in 2500 deliveries.¹³⁰

Risk factors for uterine inversion include uterine atony, a short umbilical cord, uterine anomalies, and overly aggressive management of the third stage of labor, including maneuvers such as inappropriate fundal pressure or excessive umbilical cord traction.¹³⁰ Uterotonic therapy can convert a partial inversion to a complete inversion. An abnormally implanted placenta (i.e., placenta accreta) may be first recognized when uterine inversion occurs.

Diagnosis

Many cases of uterine inversion are obvious because of hemorrhage and a mass in the vagina, but others may not be readily apparent. Inversion should be suspected in all cases of postpartum hemorrhage. Ultrasonographic examination may show characteristic findings, such as an echolucent zone within an echogenic mass filling the uterine cavity on transverse view.¹³¹ Historically, obstetricians have stated that the shock is out of proportion to the blood loss, but an underestimation of obstetric hemorrhage is more likely. An incomplete inversion not protruding through the introitus is more likely to result in missed or delayed diagnosis.¹¹⁶

Obstetric Management

Immediate replacement of the uterus, even before removal of the placenta, is the treatment goal, but it may be difficult to achieve. The appropriate technique for correcting an inversion has been described.¹³⁰ All uterotonic drugs should be discontinued immediately. The obstetrician should attempt to right the inversion by applying pressure through the vagina to the uterine fundus; ring forceps may be used on the cervix to apply countertraction. Early diagnosis and prompt correction may reduce the morbidity and mortality associated with uterine inversion.

Anesthetic Management

Often, uterine tone precludes replacement of the uterus, and uterine relaxation is necessary for successful uterine reduction.¹³⁰ The use of nitroglycerin to facilitate relaxation and replacement of the uterus has been reported.^132,133 Fairly large intravenous doses (200 to 250 µg) may be required, and the anesthesiologist typically will need to support the circulation with intravenous fluids and vasopressors. Administration of general anesthesia with a volatile halogenated agent may become necessary, not only for uterine relaxation but also to prepare for laparotomy should it become necessary to correct the inversion. Once the uterus has been replaced, a firm, well-contracted uterus is desired.¹³⁰ Oxytocin should be infused, and additional uterotonic drugs may be needed.

Epidemiology

Evidence suggests that the incidence of placenta accreta is rising, primarily because of the increasing cesarean delivery rate. Between 1994 and 2007 the rate of peripartum hysterectomy for abnormal placentation increased by 20% in the United States, an increase entirely explained by adjustment for the rising prevalence of previous cesarean delivery among childbearing women.¹⁵ The percentage of U.S. cesarean deliveries increased from 20.7% in 1997 to 32.9% in 2009, owing to increases in both primary and repeat cesarean deliveries and a decrease in the TOLAC rate.^135–137 Fortunately, the cesarean delivery rate remained stable between 2009 and 2012.¹³⁸

Previous cesarean delivery or other uterine surgery increases the risk for both placenta previa and accreta. The combination of placenta previa with previous cesarean delivery synergistically increases the risk for coexisting placenta accreta, particularly if the placenta is anterior and overlies the uterine scar. In a prospective multicenter observational study, Silver et al.²⁵ determined the relationship between placenta accreta, placenta previa, and previous cesarean delivery. Placenta previa with no prior uterine surgery conferred a 3% risk for placenta accreta. In women with placenta previa and one previous cesarean delivery, the risk for placenta accreta was 11%. In patients with placenta previa and a history of two previous cesarean deliv­eries, the incidence of placenta accreta was increased to 40%. The incidence of placenta accreta was over 60% in women with placenta previa and a history of three or more previous cesarean deliveries (see Table 38-2). Another study noted a relationship between the extent of uterine wall invasion and the number of previous cesarean deliveries.¹³⁹

Diagnosis

In some cases of placenta accreta the condition is first suspected at vaginal delivery, when the obstetrician notes difficulty in separating the placenta. The definitive diagnosis is then made at laparotomy. Antenatal diagnosis of placenta accreta facilitates effective planning. Antenatal diagnosis is associated with less maternal and neonatal morbidity, including decreased blood loss at delivery and transfusion of fewer units of blood products.¹⁴⁰ Ultrasonography is a useful screening tool in patients with placenta previa and/or previous cesarean delivery and is the primary imaging modality used for making the diagnosis of placenta accreta. However, among women at risk for placenta accreta, ultrasonography has imperfect sensitivity and specificity.¹³⁴ Some evidence suggests that MRI may help confirm the diagnosis in at-risk patients with inconclusive ultrasonographic examinations.¹⁴¹

Obstetric Management

The ACOG advises that providers working at small hospitals without adequate blood bank supplies consider transferring patients with placenta accreta to a tertiary care facility.¹³⁴ Patients treated at institutions with a 24-hour in-house obstetrician and anesthesiologist, immediate availability of a gynecologic oncologist, a fully stocked blood bank, and interventional radiology services suffer less morbidity than those treated at hospitals without these services.¹⁴² Planned delivery with all of the necessary multidisciplinary collaborators present compared with emergency delivery is associated with less maternal morbidity, including fewer transfusions, complications, and intensive care unit admissions. Some cases of vaginal bleeding remote from term resolve spontaneously, and expectant management may prolong the duration of pregnancy, at least into the third trimester. However, the risk for severe antenatal bleeding increases as gestational length increases.¹⁴² Timing of delivery therefore involves balancing this risk against the neonatal risks of preterm delivery. Institutions that manage women with suspected placenta accreta expectantly must have the capacity to mobilize the entire perioperative team at any time.

Most patients with known placenta accreta should undergo planned preterm cesarean delivery and hysterectomy with the placenta left in situ because attempts to remove the placenta are likely to initiate hemorrhage. Because the positive predictive value of ultrasonography may be low, the ACOG advises that it is reasonable to await spontaneous placental delivery in some cases, but manual removal of the placenta should be avoided.¹³⁴ Preoperative placement of ureteral stents may minimize urinary tract injury.¹³⁴ A midline vertical skin incision may provide optimal surgical exposure; it may be necessary to modify the uterine incision to avoid cutting through the placenta.

The preoperative insertion of internal iliac artery balloon catheters is controversial. Optimally, the balloons are inflated after delivery as a means to provide a less bloody surgical field and decrease blood loss and transfusion requirements. Retrospective cohort studies have reported conflicting data regarding effects of balloon catheter placement on estimated blood loss, transfusion requirements, and duration of the surgical procedure.^143–146 The largest of these studies compared 59 patients who had placement of preoperative balloon catheters with 58 who did not and determined there was lower mean estimated blood loss, fewer cases with estimated blood loss greater than 2500 mL, and fewer massive transfusions in the balloon group.¹⁴⁶

Multiple complications can arise from the placement of these devices, some of them involving serious disruptions of the vasculature and lower extremity ischemia.^{144,146–148} Introduction of the arterial catheters (without inflation) can cause fetal bradycardia, necessitating emergency delivery.¹⁴⁹ If employed, therefore, internal iliac artery balloon catheters should be placed in the operating room to avoid dislodgement during transport and to allow for rapid delivery should fetal compromise occur. Authors of a recent review of the literature opined that current evidence, confined to case reports, case series, and small retrospective studies, is inadequate to guide clinical decision-making.¹⁵⁰ The Society of Maternal-Fetal Medicine (SMFM) recommends reserving the use of prophylactic intra-arterial balloon catheters for well-counseled women with a strong desire for fertility preservation, those who decline blood products, and those with unresectable placenta percreta.¹⁵¹

Two forms of conservative therapy for placenta accreta have been described. In selected patients with a partial placenta accreta, small focal areas of placental invasion may be managed by curettage and oversewing. Alternatively, it may be possible to leave the intact placenta in situ, close the uterus and abdomen, and await spontaneous placental involution.^152–154 However, conservatively managed patients often require additional therapies such as internal iliac artery balloon inflation, embolization, or methotrexate. Some patients treated in this way have subsequently had successful pregnancies¹⁵²; however, complications have been reported, including secondary hemorrhage, subsequent need for hysterectomy, and sepsis.^152,154 The ACOG considers planned peripartum hysterectomy to be the management of choice for patients with placenta accreta and cautions the obstetrician to reserve uterine conservation strategies for hemodynamically stable patients who strongly desire future fertility.¹³⁴

Anesthetic Management

Anesthetic management is similar to other cases of severe postpartum hemorrhage and peripartum hysterectomy (see later discussion). Preoperative suspicion for placental implantation abnormalities should alert the anesthesia provider to the potential for massive blood loss. One group of investigators reported that estimated blood loss exceeded 2000 mL in 66% of cases of placenta accreta, 5000 mL in 15%, 10,000 mL in 6.5%, and 20,000 mL in 3% of cases.¹⁵⁵ Initial blood loss may be minimal but can rapidly become torrential if the placental bed is disturbed or if the surgeons encounter unavoidable placental tissue during the course of the hysterectomy.

Peripartum Hysterectomy

Peripartum hysterectomy is the definitive treatment for postpartum hemorrhage unresponsive to medical and other invasive therapies. The two most common indications for this procedure are uterine atony and placenta accreta.¹⁵ Between 1994 and 2007 the overall rate of peripartum hysterectomy increased by 15% in the United States, because of a 130% increase in hysterectomy for atony and a 23% increase in hysterectomy for placental implantation abnormalities.¹⁵ The increase in hysterectomy for placental abnormalities is almost entirely explained by an increase in the cesarean delivery rate.¹⁵ Parturients with a history of previous cesarean delivery are more than five times as likely to require a peripartum hysterectomy than those without this history, and risk for hysterectomy rises progressively with an increasing number of previous cesarean deliveries.^15,25 The incidence of peripartum hysterectomy due to uterine rupture has declined in the United States because the rate of TOLAC has declined.¹⁵ Elective peripartum hysterectomy may also be undertaken for concurrent gynecologic abnormalities, especially malignancies.

Peripartum hysterectomy is a technically challenging operation; the uterus is enlarged, exposure may be difficult, the vessels are engorged, and the pregnant uterus receives a rich collateral blood supply. The presence of dense adhesions from previous surgeries can further complicate the surgery. A 2010 systematic review of emergency postpartum hysterectomy for hemorrhage revealed a perioperative morbidity rate of 56% and a mortality rate of 2.6%.¹⁶⁸ Compared with nonobstetric hysterectomy, patients undergoing obstetric hysterectomy are more likely to suffer postoperative hemorrhage and require blood transfusion, have intraoperative urinary tract injury, and experience perioperative complications such as wound infection, venous thromboembolism, and cardiovascular and other medical complications.¹⁶⁹ Mortality is more than 25 times higher in peripartum than nonperipartum hysterectomy.¹⁶⁹

Transfusion is required in 44% or more of patients.¹⁶⁸ Emergency peripartum hysterectomy is associated with increased blood loss, worse coagulopathy, and increased transfusion rates compared with elective peripartum hysterectomy.¹⁷⁰ A multicenter review showed that the mean blood loss for emergency obstetric hysterectomy was 2526 mL, with a mean transfusion requirement of 6.6 units of PRBCs; in elective procedures, the mean blood loss was 1319 mL and the average replacement was 1.6 units of PRBCs (Table 38-4).¹⁷⁰

Owing to the challenging technical aspects of the procedure, the obstetrician may elect to perform a subtotal hysterectomy, wherein the cervix is left in situ. Subtotal approaches are associated with fewer urinary tract and other operative injuries and a shorter length of stay,^168,169 but greater mean transfusion requirements and more frequent rates of reoperation than total hysterectomy.¹⁶⁹ Subtotal hysterectomy is not appropriate for patients with bleeding from the cervix, lower uterine segment, or both (e.g., implantation of the placenta on the lower uterine segment).

Manual compression of the aorta can be a lifesaving procedure in the event of catastrophic obstetric hemorrhage.¹⁷¹ Effective aortic compression against a vertebral body in the upper abdomen should decrease blood flow to the pelvis, thereby allowing hemodynamic and hemostatic resuscitation and surgical control.¹⁷² An aortic cross-clamp requires vascular surgery expertise and retroperitoneal dissection but may be necessary to achieve hemostasis. Mild cardiac and renal dysfunction have been noted in nonobstetric patients if the aortic cross-clamp time exceeds 50 minutes¹⁷³; if a prolonged clamp time is required, the anesthesiologist should prepare for lactic acidosis and hemodynamic instability at the time the clamp is released. Advanced surgical techniques to control friable, engorged blood vessels include felt or Teflon pledgets to buttress sutures, the rapid application of straight clamps, and the application of high-pressure surgical sealants.^172,174

Risk of Anemia

Blood oxygen content and oxygen delivery to the tissues are a function of hemoglobin concentration; however, compensatory physiologic responses offset the negative effect of anemia on oxygen transport, especially if euvolemia is maintained with crystalloid or colloid intravascular volume expansion after moderate hemorrhage. Tachycardia and increased stroke volume combine to increase cardiac output, and blood viscosity and systemic vascular resistance decrease, augmenting blood flow to the tissues. Additionally, tissue oxygen extraction increases. Experience with patients who refuse transfusion indicate that these compensatory mechanisms are usually adequate to compensate for moderate blood loss; no increase in mortality is observed as long as hemoglobin concentration remains greater than 5.0 g/dL.¹⁹⁹ Weiskopf et al.²⁰⁰ studied the effects of acute normovolemic hemodilution on oxygen delivery and extraction. As hemoglobin concentration fell, systemic vascular resistance decreased and heart rate, stroke volume, and cardiac index increased. Oxygen transport rate and mixed venous oxyhemoglobin saturation did not decrease until hemoglobin concentration levels reached 5.0 g/dL. Plasma lactate did not accumulate, leading the authors to conclude that transport of oxygen was not compromised during anemia of this magnitude. These measurements were made in healthy, nonpregnant patients and volunteers who were at rest; extrapolation of these results to sick or pregnant patients may not be warranted. Anemia carries increased risks in patients with cardiovascular disease.²⁰¹

Postpartum anemia is associated with fatigue. However, fatigue is transient, and within 1 week fatigue and health-related quality of life scores are similar between postpartum patients who were anemic at discharge and those who were not.²⁰² Transfusion of 1 or 2 units of PRBCs to moderately anemic parturients has no effect on length of hospital stay.²⁰³ Jonsson et al.²⁰⁴ investigated the correlation between postoperative hematocrit and wound tissue oxygenation and wound collagen deposition. There was no correlation until the hematocrit fell below 15%, most likely because compensatory responses to anemia mitigate the effects of low oxygen content. The effect of moderate anemia on breast-feeding is unknown.

Risks of Transfusion

Blood product transfusion is associated with known risks (see Table 55-2). Transfusion-associated circulatory overload (TACO) occurs in 1% to 6% of transfused patients.^205–207 TACO sometimes develops in young patients and may follow the transfusion of as little as 1 unit of PRBCs.²⁰⁸ The likelihood of TACO increases with larger amounts of plasma transfused²⁰⁵ and may occur during the correction of the coagulopathy that accompanies severe hemorrhage and massive transfusion.

Transfusion-related acute lung injury (TRALI) is defined as a new acute lung injury (ALI) that occurs within 6 hours of transfusion in a patient without an alternative risk factor for ALI. It is accompanied by hypoxemia and radiographic evidence of pulmonary edema in the absence of circulatory overload (Box 38-4).²⁰⁹ TRALI results when human leukocyte antigen (HLA) class I and II neutrophil or possibly monocyte antibodies in donor plasma prime and activate recipient white blood cells (WBCs) to cause an ALI. This WBC activation leads to increased pulmonary microvascular permeability and interstitial and alveolar edema and extravasated neutrophils in the alveolar spaces.²¹⁰ Multiparous female donors are more likely to carry the offending antibodies.²¹¹ In 2006, U.S. blood collection agencies instituted male-only donor plasma transfusion policies, and the TRALI rate has fallen to approximately one third of its prior level (approximately 1 case per 12,000 transfused units).²¹¹ Multicomponent apheresis collection techniques also decreased the risk for TRALI.²¹⁰

Allogeneic RBC administration also risks transfusion-related immunomodulation (TRIM). The mechanisms for development of TRIM are incompletely understood, but subsequent to RBC transfusion, the host develops immune tolerance and a period of generalized immune suppression. Consequences of immune tolerance and suppression include an increased incidence of nosocomial infection, postoperative infection, and cancer recurrence.²¹² In addition, microchimerism, whereby donor cells/DNA persist in the host for several years, occurs and predisposes the recipient to autoimmune illnesses.²¹² There is some suggestion that TRIM increases the lifetime risk for nonsolid cell malignancies.²¹³ These consequences are more likely to manifest as time passes and therefore may be more likely to affect young obstetric patients than older patients. Leukoreduction may have a modest role in mitigating TRIM.²¹⁰

The risk for viral transmission because of allogeneic blood transfusion continues to decrease with thorough donor screening and use of nucleic acid amplification testing (nucleic acid technology [NAT]).²¹⁴ The residual risk for transmission after both NAT and serologic testing of donor blood is approximately 1 in 2 million for both human immunodeficiency virus (HIV) and hepatitis C virus (HCV).^198,214,215 NAT is not routinely used to test for hepatitis B virus (HBV). Instead, donor screening for hepatitis B surface antigen (HBsAg) and anti-core (HBc) antibody have greatly reduced the risk for transmission of HBV; the current risk for transmission with a blood transfusion is 1 in 350,000.^198,214 The transfusion transmission of Creutzfeldt-Jakob disease has been reported. Because of the long incubation period of this prion, symptoms may not be evident for several years after transfusion.²¹⁶ Levels of infectivity are believed to be very low.²¹⁴ The first case of West Nile virus transmitted via a blood transfusion was identified in 2002; since 2003, routine blood screening has been implemented, virtually eliminating this risk.²¹⁴ Cytomegalovirus (CMV) is carried in the monocytes of asymptomatic donors and may be transmitted to uninfected recipients of blood transfusions.²¹⁷ Most of the subsequent infections are asymptomatic or mild, but CMV infection of an immunocompromised patient and/or fetus or neonate can lead to serious sequelae. The transmission rate may be as high as 30% without preventive techniques. The risk for transmission of CMV is reduced to 1.3% if seronegative blood is transferred²¹⁸ and to 2.5% with the use of leukodepleted PRBCs.²¹⁹ In 2000, a Canadian consensus conference concluded that these two methods of pre­venting post-transfusion CMV infection can be used interchangeably.²¹⁷

Bacterial contamination occasionally occurs. Because platelets may undergo conformational changes at temperatures below 18° C, they are stored at 20° C to 24° C. This warmer storage temperature (compared with that used for PRBCs) increases the risk for bacterial proliferation. In 2004, the AABB mandated testing of all platelets for bacterial contamination. Use of culture-negative platelets has resulted in a reduction in the risk of septic transfusion to 1 in 75,000.²²⁰

Hemolytic transfusion reaction is a rare complication; it occurs most commonly as a result of accidental administration of ABO-incompatible blood.¹⁹⁸ Acute intravascular hemolysis typically results in fever, chills, nausea, flushing, and chest and flank pain. These symptoms are masked by general anesthesia. Signs that may manifest during general anesthesia include hypotension, tachycardia, DIC, and hemoglobinuria.²²¹ Immediate supportive care consists of discontinuation of the transfusion, treatment of hypotension and hyperkalemia, administration of a diuretic, and alkalinization of the urine. Assays for urine and plasma hemoglobin concentration and antibody screening confirm the diagnosis. A second crossmatch must be performed.

The biochemical and additional changes that occur during blood storage can lead to complications in the recipient, particularly when blood products are infused rapidly, as during massive transfusion for severe hemorrhage. The anticoagulant used for blood collection and storage contains citrate, which binds ionized calcium. Citrate is rapidly metabolized in the liver and typically does not lead to significant hypocalcemia. In patients who are hypothermic, have liver disease, or require rapid infusion of multiple units of blood products, however, citrate may accumulate and cause a decrease in ionized calcium. The concentration of citrate is seven times higher in fresh frozen plasma (FFP) and platelets than in PRBCs.²²² Hypocalcemia results in reduced cardiac contractility, hypotension, and elevated central venous pressure.²²³

Plasma potassium concentration increases in stored blood. Transfused potassium usually moves intracellularly or is excreted in the urine. However, rapid infusion of multiple units of blood can lead to hyperkalemia, particularly in the hypothermic acidotic patient.²²³ Blood maintained at 4° C also can contribute to hypothermia, especially if the patient is anesthetized in a cold operating room. The decreased pH of stored blood is caused by the addition of citrate-phosphate-dextrose and the accumulation of lactic and pyruvic acids as a result of RBC metabolism and glycolysis. Despite the lower pH, transfusion of large amounts of stored blood rarely causes acidosis as long as tissue perfusion remains normal.²²³

Blood Products

Allogeneic blood can be transfused as whole blood or component products. Whole blood would be an ideal choice for maintaining intravascular volume in the setting of massive hemorrhage. Alexander et al.²⁵⁷ performed a population-based observational study of 1540 obstetric patients who required transfusion. The authors compared outcomes among patients who received only whole blood (43%), only PRBCs (39%), or multicomponent therapy (19%). Whole blood was associated with lower rates of acute tubular necrosis than the two other transfusion practices. However, few donor units are kept as whole blood in the modern blood bank. The high demand for blood components such as platelets, FFP, and cryoprecipitate necessitates that more than 90% of donor blood be fractionated into blood components. Blood component therapy provides the patient with only those products that are required and helps extend the shelf-life of each unit of donor blood because derivatives from one unit of blood can be used to treat several patients. Characteristics of commonly administered blood products are summarized in Table 38-6. During massive resuscitation, care must be taken to avoid hypothermia, acidosis, and hypocalcemia, because these conditions contribute to coagulopathy.

PRBC units are prepared by removing plasma from whole blood and replacing it with additives to improve red cell survival. These units are packaged with preservatives and anticoagulant (citrate, phosphate, dextrose, adenine) and have a 42-day shelf-life. Each unit has volume of approximately 300 mL with a hematocrit of 70%. Transfusion of 1 unit of PRBC increases the hemoglobin concentration by approximately 1 mg/dL in the absence of ongoing bleeding.

A unit of fresh frozen plasma has a volume of approximately 250 mL and contains coagulation factors. Transfusion of FFP is indicated when replacement of coagulation factors is necessary to achieve hemostasis, as may occur during massive transfusion or in the presence of DIC. Administration of FFP may be considered for correction of microvascular bleeding if the prothrombin time (PT) is more than 1.5 times normal, the international normalized ratio (INR) is greater than 2.0, or the activated partial thromboplastin time (aPTT) is more than two times normal.²²¹ FFP should not be used to treat hypovolemia or as a protein supplement. One unit of FFP per 20 kg of body weight (or 10 to 15 mL/kg body weight) is an appropriate initial dose.²²¹ The prophylactic use of FFP is not effective for decreasing blood loss in patients at risk for massive blood loss.²²²

Cryoprecipitate is prepared from thawed FFP and contains fibrinogen, factor VIII, von Willebrand factor, fibronectin, and factor XIII. In the setting of postpartum hemorrhage, cryoprecipitate is used to replace fibrinogen, which is rapidly consumed during obstetric hemorrhage. Normal pregnancy is a hypercoagulable state,^258,259 and coagulation activity peaks at the time of parturition^260,261 because of an increase in circulating tissue factor concentration and enhancement of the tissue factor–dependent coagulation pathway.²⁶² It is postulated that tissue factor and other procoagulant substances are released from the placental implantation site at the time of placental separation, augmenting thrombin formation and serving an important hemostatic function after delivery.⁴⁸ In the setting of continued bleeding from the placental bed, these thromboplastic substances may continue to enter the circulation and, in severe cases, may lead to a consumptive coagulopathy and DIC.⁴⁸

The consumption of fibrinogen appears to play a central role in the pathophysiology of peripartum hemorrhage.²⁶³ Charbit et al.¹⁸⁸ prospectively identified 128 patients who had postpartum hemorrhage and classified the hemorrhage as severe (decrease in hemoglobin concentration of ≥ 4 g/dL, transfusion of ≥ 4 units PRBCs, or invasive hemostatic intervention required) or nonsevere.¹⁸⁸ At the time postpartum hemorrhage was diagnosed, patients who subsequently developed severe hemorrhage had lower fibrinogen, prothrombin, factor V, and antithrombin levels compared with patients without severe hemorrhage. These early differences were most marked for fibrinogen. A fibrinogen concentration less than 200 mg/dL at the time hemorrhage was diagnosed had a 100% positive predictive value for severe hemorrhage; a fibrinogen concentration greater than 400 mg/dL had a 79% negative predictive value for severe hemorrhage. The coagulation changes were consistent with a consumptive coagulopathy because they were accompanied by increases in thrombin-antithrombin complexes and D-dimer levels, both markers of excessive coagulation. Furthermore, because the fall in fibrinogen concentration was twice the fall in hemoglobin concentration, it was believed that dilution did not account for the difference in fibrinogen levels between the two groups. Other investigators have confirmed that decreases in fibrinogen correlate better than other hemostatic measures with the severity of hemorrhage.^264,265

The alarmingly rapid consumption of coagulation factors, especially fibrinogen, during obstetric hemorrhage has prompted experts to recommend early monitoring for, and aggressive treatment of, hypofibrinogenemia with rapid central laboratory or point-of-care testing.^266–268 During active hemorrhage, clinicians should attempt to maintain the fibrinogen concentration higher than 150 to 200 mg/dL.²⁶⁸ Each dose of cryoprecipitate supplied by most blood banks contains 5 to 10 single-donor units of cryoprecipitate, and each unit of cryoprecipitate contains approximately twice the fibrinogen of 1 unit of FFP. Therefore, the most efficient method to replace fibrinogen during obstetric hemorrhage may be to administer cryoprecipitate.²⁶⁹ In a multicenter prospective cohort study of trauma victims (n = 1175), administration of cryoprecipitate compared with large doses of FFP was associated with a decreased risk for multiorgan failure.²⁷⁰ In a small case series of obstetric hemorrhage associated with hypofibrinogenemia, fibrinogen concentrate was used to restore fibrinogen levels.²⁷¹ This product is currently approved for the treatment of acute bleeding in individuals with congenital hypofibrin­ogenemia. Further study is required to clarify its role in the treatment of the acquired hypofi­brinogenemia associated with obstetric hemorrhage.^263,270

Thrombocytopenia may develop after massive transfusion secondary to dilution or in association with obstetric comorbidities such as HELLP syndrome. Platelet transfusion may be necessary if hemorrhage is accompanied by a platelet count less than 50,000/mm³.²²¹ In nonbleeding patients, a transfusion trigger of 20,000/mm³ has traditionally been suggested, although many clinicians prefer to administer platelets before the platelet count decreases to this value.²⁷² Platelet dysfunction associated with bleeding may also necessitate platelet administration.²²¹ One unit of donor platelets increases the platelet count by 5000 to 10,000/mm³ in the average adult. The blood bank typically provides pooled random-donor platelets or single-donor apheresis platelets obtained from an ABO- and Rh-compatible donor, although ABO compatibility is not essential. One unit of apheresis platelets is equivalent to 4 to 6 units of pooled platelets.

The optimal FFP : PRBC transfusion ratio remains a topic of research and debate. Borgman et al.²⁷³ sought to characterize the relationship between this ratio and survival in a retrospective review of combat victims in Iraq between 2003 and 2005. The investigators identified 246 patients who required massive transfusion and separated them into low (median ratio 1 : 8), medium (median ratio 1 : 2.5), and high (median ratio 1 : 1.4) FFP : PRBC ratio groups. A high FFP : PRBC ratio was independently associated with increased odds of survival after correcting for confounders (odds ratio, 8.6; 95%, CI 2.1 to 35.2). After publication of these data, some experts recommended a 1 : 1 FFP : PRBC ratio during massive hemorrhage. However, many authors,²⁷⁴ including Borgman et al.,²⁷³ pointed out the well-known limitations of retrospective data, which include the potential existence of unidentified confounders. These authors have articulated the fact that deaths in the low-ratio group occurred much earlier than deaths in the high-ratio group, raising the possibility that those in the low-ratio group had more severe injuries and died before FFP could be thawed and administered (e.g., survivor bias). Indeed, if the analysis is adjusted for survivor bias, the survival benefit associated with high FFP : PRBC ratio disappears.^274,275 Furthermore, a study in civilian trauma victims identified a FFP : PRBC ratio of 1 : 2 to 1 : 3 as optimal for survival.²⁷⁶ However, the extrapolation of data from young male trauma victims to bleeding parturients seems fraught with potential for error.

A 2013 publication described the retrospective review of records from 142 women who had postpartum hemorrhage and required transfusion within 6 hours of delivery.²⁷⁷ Patients were divided into two groups based on their response to sulprostone therapy: those in whom bleeding was controlled with sulprostone alone and those who required advanced interventional procedures. Propensity score analysis revealed that a high FFP : PRBC ratio (> 1 : 2) was associated with fewer requirements for interventional procedures. Retrospective studies such as these can only demonstrate an association between the FFP : PRBC ratio and outcome and cannot infer cause and effect. Investigators have uniformly called for high-quality randomized trials to adequately define the optimal FFP : PRBC ratio during resuscitation for massive hemorrhage.^{273,275–278}

Many blood banks have in place massive transfusion protocols whereby blood products are delivered to the operating room in fixed ratios, sometimes in a cooler or refrigerator.²⁷⁹ Such protocols allow the blood bank to more quickly provide component products²⁸⁰ and may encourage clinicians to more effectively prevent and/or treat coagulopathy. A massive transfusion protocol for postpartum hemorrhage has been described (Figure 38-5).²⁸¹

Recombinant Activated Factor VII

A number of reports have described the administration of recombinant activated factor VII (rFVIIa) for hemorrhage and coagulopathy unresponsive to conventional blood product resuscitation.²⁸² Factor VIIa binds not only to tissue factor but also with low affinity to the thrombin-activated platelet. This low-affinity binding to the activated platelet allows direct activation of factor X to Xa on the surface of the activated platelet, bypassing the normal need for factors VIIIa and IXa. Activation of factor X leads to a small thrombin burst. This thrombin, in turn, activates more platelets. Additionally, rFVIIa enhances platelet aggregation and adhesion. Recombinant activated factor VII is currently approved in the United States for the treatment and prophylaxis of bleeding in patients with hemophilia A and B with inhibitors to factors VIII and IX, acquired hemophilia, and congenital factor VII deficiency. However, its use has been reported off-label for multiple clinical scenarios, including the following: platelet disorders; severe liver disease; major trauma; cardiac, prostate, and liver surgery; stroke; and postpartum hemorrhage.^236,283 Greater numbers of patients are being treated with rFVIIa each year, and the majority of treated patients are receiving the drug for off-label indications.²⁸⁴

No randomized controlled trials investigating the use of rFVIIa in the setting of obstetric hemorrhage have been published; most recommendations are therefore based on case report or case series data. Franchini et al.²⁸⁵ compiled 9 reports of 272 obstetric patients who received rFVIIa. A 10th report described a registry of 105 patients who received rFVIIa for postpartum hemorrhage in Australia and New Zealand.²⁸⁶ Of the 377 patients included in these 10 reports, 82% were judged by their practitioners to have had a positive treatment response (bleeding decreased or stopped) after one or more doses of rFVIIa. These data are difficult to interpret. Variation in underlying pathology existed; 49% of patients underwent cesarean delivery, and 51% underwent vaginal delivery. Nineteen percent had vaginal or uterine lacerations, 25% had placental abnormalities, and 8% had retained placenta. Additionally, the doses were not standardized and there were no criteria for the administration of rFVIIa, nor were there criteria for what constituted a positive response. Most patients received simultaneous therapy with other traditional therapies such as FFP and cryoprecipitate administration. These circumstances make judgments regarding response to rFVIIa unreliable and uncontrolled; thus, a high like­lihood for bias exists. Several authors have called for randomized placebo-controlled trials of rFVIIa use for postpartum hemorrhage, but such a trial will be difficult to undertake given the low incidence and unpredictability of massive obstetric hemorrhage.

There is concern that rFVIIa may increase the likelihood of thromboembolic events. The true incidence is not known. Arterial events (e.g., myocardial infarction, cerebral thrombosis) are more likely than venous events^287,288 and are dose²⁸⁷ and age dependent.²⁸⁸ Among the 377 reported cases of rFVIIa administered for postpartum hemorrhage, eight patients developed venous thromboembolism, one suffered myocardial infarction, and one had a thrombotic cerebrovascular accident, suggesting a thrombotic complication frequency of 2.7% in the postpartum period. However, severe obstetric hemorrhage and peripartum hysterectomy are known to increase the risk for thrombotic complications; thus, this frequency cannot support or refute any causal relationship with rFVIIa. Leighton et al.²⁸⁹ advise against rFVIIa administration in the setting of amniotic fluid embolism because tissue factor may play a role in its pathophysiology and thrombotic complications may be increased.

Given the unknown efficacy and safety of rFVIIa, as well as its high cost, rFVIIa is not recommended for routine use in obstetric practice. In 2005, an American panel, based on type II evidence (obtained from well-designed, nonrandomized trials or cohort or case-control analytic trials), judged the use of rFVIIa for treatment of postpartum hemorrhage as appropriate only when bleeding continued despite clotting factor replacement.²⁹⁰ The ASA guidelines recommend consideration of rFVIIa therapy if “traditional well-tested options for treating microvascular bleeding (i.e., coagulopathy) have been exhausted.”²²¹ In addition, maintenance of normothermia and correction of acidosis are necessary for optimal rFVIIa activity; optimization of platelet count, fibrinogen, and serum calcium are also advisable as rFVIIa requires these substances to be effective.

Antifibrinolytic Therapy

It has been suggested that antifibrinolytic therapies such as tranexamic acid may be useful in treating postpartum hemorrhage–associated coagulopathy.²⁹¹ A meta-analysis of six randomized trials comparing tranexamic acid with placebo administered shortly after birth demonstrated that tranexamic acid was associated with a nonsignificant reduction in blood loss of 33 mL.²⁹² In five of the six trials, tranexamic acid was administered prophylactically. The sixth trial enrolled women who had lost at least 800 mL after vaginal delivery; tranexamic acid reduced the total blood loss over the subsequent 6 hours by 50 mL (170 mL versus 221 mL, P = .04).²⁹³ A large international placebo-controlled trial of tranexamic acid therapy in the setting of postpartum hemorrhage is ongoing.²⁹⁴ Tranexamic acid may be most beneficial for women who demonstrate hyperfibrinolysis based on hemostatic monitoring such as thromboelastography.^263,268