In 1986, the Centers for Disease Control and Prevention (CDC) initiated a program of national surveillance of pregnancy-related deaths. It is serially updated and can be accessed at: www.cdc.gov/reproductivehealth/maternalinfanthealth/pmss.html. Using this database, Creanga and colleagues (2015) reported an increase in the maternal mortality rate from 7.2 per 100,000 in 1987 to 17.8 per 100,000 in 2009. It is also clear from this report that hemorrhage remains a significant cause of pregnancy-related deaths, accounting for 11.4 percent of such deaths from 2006 to 2010. This is despite widespread recognition of the consequences of obstetric hemorrhage and the availability of modern blood-banking techniques. Notably, the United States is one of the few countries worldwide that has not reported a decline in maternal mortality rates but instead has shown a significant rise. One worrisome trend was the marked racial disparity in pregnancy-related mortality rates. In data from 2006 to 2010, white women suffered 12 deaths per 100,000 births, whereas the rate for black gravidas was 36.4 deaths per 100,000 births (Creanga, 2015). Also disturbing is that approximately 90 percent of maternal hemorrhage-related deaths have been considered potentially preventable (Berg, 2005; D’Alton, 2014).

In addition to mortality, postpartum hemorrhage can lead to severe postpartum morbidity. In a multicenter surveillance study of postpartum hemorrhage from Brazil, Rocha Filho and coworkers (2015) reported that the hemorrhage-related incidence of severe maternal morbidity approximated 3 events per 1000 live births. Significant or severe postpartum hemorrhage is estimated to complicate 2 to 4 percent of vaginal deliveries and 6 percent of cesarean births (Combs, 1991a,b; Klapholz, 1990).

Although postpartum hemorrhage has always been one of the leading causes of maternal mortality and morbidity, no universally accepted definition for this complication is recognized. Pritchard and associates (1962) quantitatively measured the blood loss in women who had a vaginal delivery, had a repeat cesarean delivery, or underwent repeat cesarean delivery with hysterectomy (Table 29-1). In a review of postpartum blood loss, Gahres and colleagues (1962) reported a mean loss following vaginal delivery of 450 mL for all series. Traditionally, postpartum hemorrhage has been defined as blood loss exceeding 500 mL following vaginal delivery and exceeding 1000 mL following cesarean delivery.

Several problems muddy this definition. First, clinical estimates of blood loss are notoriously inaccurate, and blood loss is often underestimated and underappreciated. Another major problem is that blood loss often approaches or exceeds these amounts without clinical consequences.

Another criterion used to define postpartum hemorrhage is a drop in the hematocrit of 10 percent or more. The major flaw with this definition is that the hematocrit may not reflect current blood volume status in the acutely bleeding patient (American College of Obstetricians and Gynecologists, 2015; Combs, 1991a,b). Namely, hematocrit values typically lag true losses, and values may reflect only the degree of hemorrhage.

Rather than estimating blood loss or relying on laboratory parameters, Gilstrap and Ramin (1994) suggested that postpartum hemorrhage could best be gauged by clinical evaluation of the gravida’s hemodynamic status. They suggested that term hemorrhage could best be thought of as that amount of bleeding resulting in hemodynamic instability or that would result in such if left unabated. Used in this context, hemorrhage is “significant or serious” postpartum bleeding.

The precise amount of bleeding that meets this criterion varies depending on several factors. Paramount is the degree of circulating blood volume in a given woman. The equation to calculate blood volume is shown in Table 29-2. For example, a blood loss of 1000 to 1500 mL may be of little consequence following an otherwise uncomplicated twin delivery. In contrast, a similar loss might cause significant hypotension in a woman with severe preeclampsia. Hernandez and associates (2012) prospectively studied 1443 women who received a blood transfusion for hypovolemia from obstetric hemorrhage. These authors found that median blood loss was 3529 mL and that 93 percent of these parturients had losses ≥3000 mL. They concluded that the amount of blood loss sufficient to produce signs and symptoms of hypovolemia—that is, hemodynamic instability—was large and required blood infusion to restore circulation.

Kramer and colleagues (2013) reviewed 8.5 million hospital deliveries in the U.S. Nationwide Inpatient Sample from 1999 to 2008 for risk factors for severe postpartum hemorrhage. Bleeding was defined as severe if hemorrhage was accompanied by blood transfusion, hysterectomy, or surgical repair of the uterus. They found a rise in severe postpartum hemorrhage rates from 1.9 to 4.4 cases per 1000 deliveries during this time period. Risk factors for severe hemorrhage included multifetal pregnancy, leiomyomas, preeclampsia, placenta previa or abruption, cervical laceration, operative vaginal delivery, uterine rupture, and cesarean delivery. Others have identified similar and additional risk factors Table 29-3 (Combs, 1991a,b; Gilstrap, 1987). Importantly, a familiarity with this list can prompt preventive steps when significant or multiple at-risk characteristics are present.

The California Maternal Quality Care Collaborative (CMQCC) has established risk criteria for peripartum hemorrhage (Lyndon, 2015). In a study of more than 10,000 women in a single hospital, Dilla and coworkers (2013) used three defined risk groups. They evaluated whether these three categories correlated with the risk of significant postpartum hemorrhage, which was defined as bleeding requiring transfusion. These investigators reported peripartum hemorrhage in 0.8 percent of the low-risk group, 2.0 percent of the medium-risk group, and 7.3 percent of the high-risk group. For the CMQCC, characteristics that described high-risk gravidas included: placenta previa, morbidly adherent placenta, hematocrit less than 30 percent, platelets less than 100,000/μL, active bleeding at admission, and known coagulopathy.

In a cohort study by the Canadian Institute for Health Information of more than 2 million deliveries from 2003 to 2010, Mehrabadi (2014) reported that postpartum hemorrhage rates increased from 5.1 percent in 2003 to 6.2 percent in 2010. Severe postpartum hemorrhage was defined as postpartum hemorrhage plus blood transfusion, hysterectomy, or other procedures to control bleeding. The rise in bleeding rates was mostly attributed to an increase in uterine atony rates. Lacerations, abnormal placental attachments, and coagulopathy accounted for most of the remaining cases.

Physiologic Changes

To prevent maternal mortality and morbidity from postpartum hemorrhage, clinicians should be familiar with common preventable errors. These include failures to recognize the degree of blood loss, to appreciate the clinical signs of hypovolemia, and to promptly initiate lifesaving interventions and restoration of blood volume (Clark, 2012; D’Alton, 2014).

Fortunately, the hypervolemia of pregnancy in most cases allows the parturient to accommodate normal to even excessive blood loss at vaginal delivery without a drop in postpartum hematocrit. In one study, the mean postpartum hematocrit decline was 2.6 ± 4.3 percent, and a third of women either had no decline or had an actual rise (Combs, 1991b). Women undergoing cesarean delivery experienced a mean drop in hematocrit of 4.0 ± 4.2 percent after equilibration, but approximately 20 percent had no decline (Combs, 1991a).

Even with moderate additional blood loss, venous compensatory mechanisms allow most women to tolerate such hemorrhage without hemodynamic compromise. With continued hemorrhage, however, these mechanisms may be overwhelmed and lead to hypotension, decreased tissue perfusion, oliguria, cellular hypoxia, and death. Indeed, hemorrhage is the most common cause of shock in obstetrics and gynecology.

As blood volume deficits exceeds a critical level, compensatory mechanisms usually are inadequate to maintain cardiac output and blood pressure. At this point, additional small blood losses result in rapid clinical deterioration. Despite an initial increase in total oxygen extraction by maternal tissue, maldistribution of blood flow results in local tissue hypoxia and metabolic acidosis. This produces a vicious cycle of vasoconstriction, organ ischemia, and cellular death. Consequently, capillary membrane integrity and additional intravascular volume are lost (Slater, 1973). Increased platelet aggregation is also found in hypovolemic shock. The subsequent, release of several vasoactive mediators causes small vessel occlusion and further impairment of microcirculatory perfusion.

Patient Resuscitation

In an extensive “toolkit” developed by the California Maternal Quality Care Collaborative (CMQCC), obstetric hemorrhage is divided into stages 1 through 3. Stage 0 is prevention and recognition of hemorrhage (Lyndon, 2015). Stage 3 is defined as an estimated blood loss of >1500 mL or the administration of greater than two units of packed red blood cells (pRBCs), persistent unstable vital signs, or the suspicion of disseminated intravascular coagulopathy (DIC). For stage 3, this group recommends activation of a massive transfusion protocol. This seems a bit overcautious, and elements of this protocol are discussed in further detail in Chapter 7 (p. 98). Surgical approaches to control hemorrhage sometimes are also necessary in this latter stage.

Not surprisingly, opinions vary regarding the most appropriate ratio of plasma, pRBCs, and platelets in the presence of severe postpartum bleeding. As discussed in Chapter 7, generally, 1:1:1 blood product replacement is suggested by many. This ratio refers to one unit of pRBCs to one unit of fresh frozen plasma (FFP) and one unit of platelets. In some instances, plasma and platelets are given even if those laboratory values are normal with the goal of preventing coagulopathy from developing.

From the foregoing, it can be appreciated that fresh whole blood would be preferable in most cases of postpartum hemorrhage. Whole blood provides needed intravascular volume; it increases the fibrinogen level and hematocrit; and it reduces the risk for acute kidney injury. For example, Alexander and colleagues (2009) summarized outcomes from 1540 women who received a blood transfusion. Of these, 659 were given whole blood, 593 received packed red cells only, and 288 received a combination of blood component therapy. Those transfused with whole blood had a lower incidence of acute kidney injury, intensive care unit admission, and death compared with the other two groups.

Although fresh whole blood is ideal for the management of serious hemorrhage, in practice it is rarely available. Thus, crystalloid solution and pRBCs are the mainstays of volume replacement. When blood loss is massive, however, such replacement may result in a depletion of platelets and soluble clotting factors leading to a dilutional coagulopathy (Cunningham, 2015). This impairs hemostasis and further contributes to blood loss. As shown in Table 29-4, stored whole blood is deficient in factors V, VIII, XI, and platelets, and almost all soluble clotting factors are absent from pRBCs. Thus, severe hemorrhage treated only with pRBC replacement may also lead to diminished fibrinogen levels and prolonged prothrombin (PT) and partial thromboplastin times (PTT). In some cases, frank DIC accompanies shock, and this may confuse the distinction between dilutional and consumptive coagulopathy (Abdul-Kadir, 2014). Fortunately, in most situations encountered in obstetrics, treatment of both is the same (Cunningham, 2015).

Various algorithms have been proposed to guide the replacement of platelets and clotting factors according to the volume of blood loss, especially when using only pRBCs. That said, understandably, patient variability is great. Such component replacement is rarely needed with the acute replacement of five units of pRBCs or less. When blood loss exceeds this amount, consideration often prompts laboratory evaluation of platelet count and fibrinogen levels. Replacement is guided by laboratory values and clinical evidence of bleeding from a coagulopathy. In the woman who is bleeding, the platelet count ideally is kept above 50,000/μL by the infusion of platelet concentrates. A fibrinogen level <150 mg/dL or a sufficiently prolonged PT or PTT in a bleeding patient is an indication for fresh-frozen plasma to be given in doses of 10 to 15 mL/kg. Importantly, approximately 30 minutes are required for frozen plasma to thaw. Cryoprecipitate (15 mL per unit) may also be used as a source of fibrinogen. This product usually offers no advantage compared with FFP for general clotting factor replacement. An uncommon exception may be a woman in whom volume overload is a concern during resuscitation. Following platelet or factor replacement, levels are followed closely until the patient is stable.

Proper management of postpartum bleeding requires a thorough search for the specific cause of the hemorrhage. Individual diagnoses require specific management, both in terms of technique and in the rapidity with which such techniques are applied. For example, proper treatment of intermittent uterine atony may at times involve a prolonged period of uterine massage, observation, and continued attempts at pharmacologic reversal. On the other hand, postpartum hemorrhage from placenta percreta will generally require prompt hysterectomy. Errors can be made when the clinician manages “postpartum hemorrhage” without attempts to determine specific etiology.

Fortunately, the causes of severe postpartum hemorrhage are few. In virtually all cases, one or more of four factors are involved: uterine atony; retained placenta, including the placental accrete syndromes; or upper or lower genital tract lacerations, including cesarean delivery. The fourth is coagulopathy that manifests bleeding from the first three causes.

Uterine Atony

Risk Factors

The most frequent cause of serious postpartum hemorrhage is failure of the uterus to contract. Risk factors for uterine atony include advanced parity, the use of oxytocin for labor augmentation, chorioamnionitis, arrest disorders of labor, and uterine overdistention, as with macrosomia or multifetal gestation. In the series described by Clark and coworkers (1984), 80 percent of women undergoing hysterectomy for intractable atony had one or more of these risk factors. Recognition of these characteristics should alert the clinician to the possibility of this complication.

Bleeding from atony may be rapid and leave little time for indecision. Therefore, every obstetric unit should establish a protocol or checklist for atony treatment. Uterine massage is classically taught as an essential component of third-stage labor to prevent atony and postpartum hemorrhage. However, the benefit of this procedure in the review by Hofmeyr and associates (2013) was inconclusive. Notably, the authors concluded that this was not reason enough to change current practice.

Uterotonic Agents

Despite uterine massage, if atony develops following removal of the placenta, firm compression is exerted on the uterus by bimanual compression (Fig. 29-1). Even if atony does not promptly resolve, such compression can often bring temporary effective control of hemorrhage. This provides time to administer uterotonic agents, obtain blood products, and summon assistance. Simultaneous with such compression, a dilute solution of 20 or 40 units of oxytocin in 1000 mL of crystalloid is infused rapidly. Because rapid crystalloid infusion is the first step in management of such women, the intravenous line can be opened fully. The direct intravenous bolus injection of undiluted oxytocin may cause a paradoxical hypotensive response and thus is not recommended (Secher, 1978).

If these initial attempts to reverse uterine atony are unsuccessful, additional pharmacologic maneuvers are indicated (Table 29-5). Methylergonovine (Methergine), 0.2 mg, may be given intramuscularly every 2 to 4 hours for up to five doses. It should not be given intravenously. Side effects include nausea, vomiting, and headaches. Severe hypertension can develop in women with preeclampsia or gestational hypertension.

Various prostaglandin preparations can also effectively reverse uterine atony. One commonly used is carboprost tromethamine (Hemabate), which is a synthetic 15-methyl analogue of prostaglandin F_2α. A single dose of 0.25 mg is usually effective. In selected cases, however, repeat dosing at intervals of 15 to 90 minutes can be given. The decision for additional injections and the dosing interval should be dictated by clinical events. The total dose should not exceed eight doses (Pfizer, 2014). It is administered as an intramuscular injection. It can be given into the myometrium, but is never given intravenously.

The use of intrarectal or sublingual misoprostol (Cytotec) has also been described (Tunçalp, 2012). This is prostaglandin E₁ and available as 200-μg tablets. Doses of 800 to 1000 μg are given per rectum.

These agents may be added sequentially, and thus administration can overlap as needed to control bleeding. That said, none of these latter pharmacologic agents has proven to be superior to oxytocin for the treatment of uterine atony.

The use of uterine packing for atony following failure of pharmacologic therapy is controversial but may provide hemostasis in some women. Uterine packing may also be useful as a temporary method to decrease bleeding to a degree to allow for adequate blood replacement prior to attempting surgical control of hemorrhage.

At times, atony may be intermittent, requiring repeated physical or pharmacologic intervention over a period of several hours. Under such circumstances, care is taken to assess ongoing blood losses and to ensure adequate fluid and blood product replacement and hemodynamic stability. If, following the above maneuvers, the patient with atony continues to bleed significantly, surgical intervention is indicated.

Lacerations

Lower genital tract lacerations and uterine rupture are discussed in detail in Chapters 20 (p. 320) and 30 (p. 482).

Retained Placenta

After birth of the newborn, separation and expulsion of the placenta will typically follow. If spontaneous delivery is delayed, manual extraction may be indicated. In one study of 335 women with postpartum hemorrhage, third-stage labor lasting longer than 18 minutes was associated with a significant risk of postpartum hemorrhage. A third stage lasting more than 30 minutes was associated with a sixfold higher risk for hemorrhage (Magann, 2005).

Manual Curettage

Following its delivery, the placenta is carefully inspected for integrity of the cotyledons. This is especially true if a succenturiate lobe was identified antepartum during routine sonographic evaluation. With this placental anomaly, one or more small accessory lobes develop in the membranes at a distance from the main placenta.

If there is a suggestion of retained fragments, then manual exploration of the uterine cavity is undertaken in the delivery room. This is typically well tolerated by women with regional analgesia in place. However, for others, supplemental analgesia can be provided by anesthesia staff. As discussed in Chapter 19 (p. 309), suitable agents can include intravenous ketamine or inhaled nitrous oxide.

For curettage, an operator can wrap a 4 × 4-inch gauze around one or more fingers. This can generate greater friction against retained placental fragments and membranes to aid removal. During exploration, the membranes typically create a flat slippery uterine cavity surface. In contrast, placental fragments render the decidual surface irregular, raised, and slightly spongy.

To begin, the gauze-covered hand is introduced through the cervix into the uterine cavity and is guided to the fundus. Here, manual pressure against the decidual surface begins at the fundus and moves toward the cervix. After this first pass, the hand reaches again for the fundus but is moved a bit counterclockwise and positioned at a point adjacent to the first pass. A sweep toward the cervix is again completed. This process is repeated until the circumferential area of the cavity has been explored. If slippery or spongy tissue is encountered, it is firmly grasped and teased away from the cavity wall.

Surgical Curettage

If retained fragments cannot be evacuated in this manner, sharp curettage is undertaken. For this, established regional analgesia is sufficient, and curettage can be completed in the labor room. Alternatively, if there is greater concern for a morbidly adherent placenta or if a greater level of analgesia is required, then transfer to the operating room prior to curettage is reasonable. Before and during curettage, dilute oxytocin can be infused simultaneously.

With appropriate anesthesia established and the patient in standard lithotomy position, sharp curettage can be completed. The steps mirror those for sharp curettage for abortion, described in Chapter 9 (p. 136), but differ in that a large-loop postpartum curette is selected. The large loop easily passes through the dilated cervix and provides a sizable surface area to clear the cavity quickly. Moreover, its wide, blunt tip minimizes the risk of fundal perforation.

To begin, the uterine curette is inserted though the cervix and is advanced to the fundus, following the long axis of the corpus. On reaching the fundus, the sharp surface of the curette is positioned to contact the adjacent decidua. Pressure is exerted against the wall as the curette is pulled toward the lower uterine segment. The curette is then redirected to the fundus and positioned immediately adjacent to the path of the first curettage pass. In this fashion, the entire uterine cavity is sequentially and circumferentially curetted. The collected specimen is sent for pathologic evaluation.

Despite such maneuvers, if pieces of placenta remain adhered and are associated with hemorrhage, morbidly adherent placenta is diagnosed clinically. Exploratory laparotomy and hysterectomy are typically indicated, as discussed in Chapter 27 (p. 446).

Uterine Packing

Packing Efficacy

Two major techniques can be employed for intrauterine tamponade to treat postpartum hemorrhage. These are balloon tamponade or uterine packing with gauze. For either, anesthesia requirements are typically less than for dilatation and curettage. Thus, an epidural previously placed for labor or intravenous sedation is suitable in most cases. These methods can be usually completed in a labor room.

Although once popular, uterine packing curiously fell from routine practice during the past 30 to 50 years. Three specific objections to the practice of uterine packing can be identified. The first objection is that the procedure is unphysiologic, which cannot be denied, but this objection can be set aside for lack of specific scientific relevance. The other two objections are the potential for concealed hemorrhage and infection.