Hacker & Moore's Essentials of Obstetrics and Gynecology: With STUDENT CONSULT Online Access,5th ed.

Chapter 10

Obstetric Hemorrhage and Puerperal Sepsis

Matthew Kim, Robert H. Hayashi, Joseph C. Gambone

The most common causes of maternal death are hemorrhage, embolism, hypertensive disease, and infection. In this chapter, the problems of obstetric hemorrhage and infection are considered. These conditions are associated not only with potential maternal and fetal mortality but also with significant morbidity and prolonged hospitalization.

image Antepartum Hemorrhage

It is critical for the well-being of both the mother and the fetus that the patient who presents with third-trimester bleeding be evaluated and managed emergently. The differential diagnosis of third-trimester bleeding is listed in Box 10-1.


BOX 10-1 Causes of Antepartum Bleeding.


Placenta previa

Abruptio placentaePreterm labor


Uterine rupture

Fetal (chorionic) vessel rupture

Cervical or vaginal lacerations

Cervical or vaginal lesions, including cancer

Congenital bleeding disorder

Unknown (by exclusion of the above)



If a patient is bleeding profusely, a team approach to the assessment and management should be instituted to establish hemodynamic stability. This team should include an obstetrician, an anesthesiologist, and nurses who are knowledgeable about the management of the critically ill patient. At least two large-bore peripheral intravenous lines should be placed because this allows for the most rapid replacement of fluid and blood volume. A central venous pressure line, or preferably a Swan-Ganz catheter, is helpful in the management of hypovolemic shock.

The vital signs and amount of bleeding should be checked immediately, as should the patient’s mental status. Medical history should be checked for known bleeding disorders or liver disease, which predisposes to coagulopathy. A pelvic examination should not be performed until placenta previa has been excluded by ultrasonography. Once placenta previa has been excluded, a sterile speculum examination can be safely done to rule out genital tears or lesions (e.g., cervical cancer) that may be responsible for the bleeding. If none are identified, a digital examination may be performed to determine whether cervical dilation is present.

A complete blood count should be obtained and compared with previous evaluations to help assess the amount of blood loss, although acute blood loss may not be reflected in the hemoglobin level until homeostasis has been reestablished. An assessment of the patient’s coagulation profile should be done by obtaining a platelet count, serum fibrinogen level, prothrombin time, and partial thromboplastin time. Additionally the patient should be typed and crossmatched for at least 4 units of blood (packed cells). A rapid but subjective method to test for coagulopathy is to partially fill a red-topped tube with blood. If a clot does not form, or once formed does not stay clotted, the patient most likely has disseminated intravascular coagulopathy (DIC).

An important and accurate method for determining the cause of third-trimester bleeding is ultrasonography. This evaluation should include not only the location and extent of the placenta but also an assessment of gestational age, an estimate of fetal weight, a determination of the fetal presentation, and a screening for fetal anomalies. Uterine activity and the fetal heart rate should be assessed with a monitored strip to rule out labor and establish fetal well-being.

image Abnormal Placentation

The incidence of placenta previa, the most common type of abnormal placentation, is 0.5%. Bleeding from a placenta previa accounts for about 20% of all cases of antepartum hemorrhage.Seventy percent of patients with placenta previa present with painless vaginal bleeding in the third trimester, 20% have contractions associated with bleeding, and 10% have the diagnosis made incidentally by ultrasonography or at term.


Factors that have been associated with a higher incidence of placenta previa include (1) multiparity, (2) increasing maternal age, (3) prior placenta previa, and (4) multiple gestation. Patients with a placenta previa have a 4% to 8% risk for having placenta previa in a subsequent pregnancy.


Placenta previa is classified according to the relationship of the placenta to the internal cervical os (Figure 10-1). Complete placenta previa implies that the placenta totally covers the cervical os. A complete placenta previa may be central, anterior, or posterior, depending on where the center of the placenta is located relative to the os. Partial placenta previa implies that the placenta partially covers the internal cervical os. A marginal placenta previa is one in which the edge of the placenta extends to the margin of the internal cervical os.


FIGURE 10-1 Types of placenta previa.


The classic presentation of placenta previa is painless vaginal bleeding in a previously normal pregnancy. The mean gestational age at onset of bleeding is 30 weeks, with one third presenting before 30 weeks. Placenta previa is almost exclusively diagnosed today by ultrasonography. Between 4% and 6% of patients have some degree of placenta previa on ultrasonic examination before 20 weeks’ gestation. With the development of the lower uterine segment, a relative upward placental migration occurs, with 90% of these resolving by the third trimester. Complete placenta previa is the least likely to resolve, with only 10% of cases resolving by the third trimester. When placenta previa is diagnosed in the second trimester, a repeat sonogram is indicated at 30 to 32 weeks for follow-up evaluation.

Transabdominal ultrasonography has an accuracy of 95% for placenta previa detection. If the placenta is implanted posteriorly and the fetal vertex is low, the lower margin of the placenta may be obscured and the diagnosis of placenta previa missed. Transvaginal ultrasonography can accurately diagnose placenta previa in virtually all cases.


Once the diagnosis of placenta previa is established, management decisions depend on the gestational age of the fetus and the extent of the vaginal bleeding. With a preterm pregnancy, the goal is to attempt to obtain fetal maturation without compromising the mother’s health. If bleeding is excessive, delivery must be accomplished by cesarean birth regardless of gestational age. When the bleeding episode is not profuse or repetitive, the patient is managed expectantly in the hospital on bed rest. With expectant management, 70% of patients will have recurrent vaginal bleeding before completion of 36 weeks’ gestation and will require delivery. If the patient reaches 36 weeks, fetal lung maturity should be determined by amniocentesis and the patient delivered by cesarean birth if the fetal lungs are mature. Elective delivery is preferable because spontaneous labor places the mother at greater risk for hemorrhage and the fetus at risk for hypovolemia and anemia.


A patient with a low-lying placenta, when the placental margin is within 2 cm of the endocervical os, may present in the same way as a patient with placenta previa. It may be difficult to distinguish a low-lying placenta from a marginal placenta previa, but a transvaginal ultrasound is typically diagnostic. Although vaginal delivery is not contraindicated, the same level of monitoring should be maintained for maternal hemodynamic stability and fetal well-being.


The maternal mortality from placenta previa has dropped precipitously during the past 60 years from 30% to less than 1%. This has primarily been the result of the liberal use of cesarean delivery and careful expectant management. The rare maternal death is generally associated with complications of cesarean or uncontrolled hemorrhage from the placental site. The lower uterine segment does not contract well, especially after a lower uterine incision from cesarean delivery. DIC may also result if a massive hemorrhage or an associated abruption occurs.

The risk for antepartum or intrapartum hemorrhage, or both, is a constant threat to the patient with placenta previa. Bleeding may be exacerbated by an associated placenta accreta or uterine atony. Placenta accreta implies an abnormal attachment of the placenta through the uterine myometrium as a result of defective decidual formation (absent Nitabuch’s layer). This abnormal attachment may be superficial (accreta), or the placental villi may invade partially through the myometrium (increta) or extend to the uterine serosa (percreta). Two thirds of patients with this complication require hysterectomy. Patients with a history of uterine surgery are at greatest risk for developing an accreta. In fact, those with a prior cesarean delivery have a 25% risk.

Placenta previa predisposes to preterm delivery, which poses the greatest risk to the fetus. As a result of advances in obstetric and neonatal care, the perinatal mortality rate (PMR) for premature infants has declined over the past decade. The incidence of malpresentation with placenta previa is 30%, presumably owing to the mass effect and distortion of the lower uterine segment.

image Abruptio Placentae

Abruptio placentae, or premature separation of the normally implanted placenta, complicates 0.5% to 1.5% of all pregnancies (1 in 120 births). Abruption severe enough to result in fetal death occurs in 1 per 500 deliveries.


Factors associated with an increased incidence of abruption are noted in Box 10-2. The most common of these risk factors is maternal hypertension, either chronic or as a result of preeclampsia. The risk for recurrent abruption is high: 10% after one abruption and 25% after two.


BOX 10-2 Risk Factors for Abruptio Placentae.

Maternal hypertension

Placental abruption in a prior pregnancy


Polyhydramnios with rapid decompression

Premature rupture of membranes

Short umbilical cord

Tobacco use

Folate deficiency



Placental separation is initiated by hemorrhage into the decidua basalis with formation of a decidual hematoma. The resulting separation of the decidua from the basal plate predisposes to further separation and bleeding as well as to compression and destruction of placental tissue. The inciting cause of placental separation is unknown. It may be due to an inherent weakness or anomaly in the spiral arterioles. Blood may either dissect upward toward the fundus, resulting in a concealed hemorrhage, or extend downward toward the cervix, resulting in an external or revealed hemorrhage.


Clinically, the diagnosis of a placental abruption is entertained if a patient presents with painful vaginal bleeding in association with uterine tenderness, hyperactivity, and increased tone. The signs and symptoms of placental abruption are, however, variable. The most common finding is vaginal bleeding, seen in 80% of cases. Abdominal pain and uterine tenderness are present in 66% of cases, fetal distress in 60%, uterine hyperactivity and increased uterine tone in 34%, and fetal demise in 15%.

The diagnosis of placental abruption is made clinically. Ultrasonography may detect only 2% of abruptions. Because placental abruption may coexist with a placenta previa, the reason for doing an initial ultrasonic examination is to exclude the previa.

Management of the patient with an abruption includes careful maternal hemodynamic and fetal monitoring, serial evaluation of the hematocrit and coagulation profile, and delivery. Intensive monitoring of both the mother and the fetus is essential because rapid deterioration of the condition of either one can occur. Blood products for replacement should always be available, and a large-bore (16- to 18-gauge) intravenous line must be secured. Red blood cells should be given liberally if indicated. In the setting of placental abruption, the use of tocolytics or uterine relaxants is not advised. Uterine tone must be maintained to control bleeding following delivery, or at least to control the bleeding sufficiently to allow a safe hysterectomy to be performed, if necessary.


Abruption places the fetus at significant risk for hypoxia and, ultimately, death. The perinatal mortality rate due to placental abruption is 35%, and the condition accounts for 15% of third-trimester stillbirths. Fifteen percent of live-born infants have significant neurologic impairment.

Placental abruption is the most common cause of DIC in pregnancy. This results from release into the maternal circulation of thromboplastin from the disrupted placenta and subplacental decidua, causing a consumptive coagulopathy. Clinically significant DIC complicates 20% of cases and is most commonly seen when the abruption is massive or fetal death has occurred. Hypovolemic shock and acute renal failure due to massive hemorrhage may be seen with a severe abruption if hypovolemia is left uncorrected. Sheehan’s syndrome (amenorrhea as a result of maternal postpartum pituitary necrosis) may be a delayed complication resulting from coagulation within the portal system of the pituitary stalk.

image Uterine Rupture

Uterine rupture implies complete separation of the uterine musculature through all of its layers, ultimately with all or a part of the fetus being extruded from the uterine cavity. The overall incidence is 0.5%.

Uterine rupture may be spontaneous, traumatic, or associated with a prior uterine scar, and it may occur during or before labor or at the time of delivery. A prior uterine scar is associated with 40% of cases. With a prior lower-segment transverse incision, the risk for rupture is less than 1%, whereas the risk with a high vertical (classical) scar is 4% to 7%. Sixty percent of uterine ruptures occur in previously unscarred uteri.


The signs and symptoms of uterine rupture are highly variable. Typically, rupture is characterized by the sudden onset of intense abdominal pain and some vaginal bleeding. Impending rupture may be heralded by hyperventilation, restlessness, agitation, and tachycardia. After the rupture has occurred, the patient may be free of pain momentarily and then complain of diffuse pain thereafter. The most consistent clinical finding is an abnormal fetal heart rate pattern. The patient may or may not have vaginal bleeding, and if it occurs, it can range from spotting to severe hemorrhage. The presenting part may be found to have retracted on pelvic examination, and fetal parts may be more easily palpated abdominally. Abnormal contouring of the abdomen may be seen. Fetal distress develops commonly, and fetal death or long-term neurologic sequelae may occur in 10% of cases.

A high index of suspicion is required, and immediate laparotomy is essential. In most cases, total abdominal hysterectomy is the treatment of choice, although débridement of the rupture site and primary closure may be considered in women of low parity who desire more children.


Delay in management places both mother and child at significant risk. The major risk to the mother is hemorrhage and shock. Although the associated maternal mortality rate is now less than 1%, if the mother is left untreated, she will almost certainly die. For the fetus, rapid intervention will minimize morbidity and mortality. The associated fetal mortality rate is still about 30%.

image Fetal Bleeding

Rupture of a fetal umbilical vessel complicates 0.1% to 0.8% of pregnancies. The diagnosis of fetal bleeding is made by performing an Apt test. After obtaining blood from the vagina and putting it into a red-topped test tube, tap water is added. The water will lyse blood cells and release hemoglobin. Adding 1 mL of KOH results in a brown discoloration when the hemoglobin is maternal. If the blood is fetal in origin, the color of the fluid will remain red because the fetal hemoglobin will not be denatured by the KOH. Fetal bleeding often results when the cord insertion is velamentous, implying that the vessels of the cord insert between the amnion and chorion away from the placenta. The incidence of velamentous cord insertion varies from 1% in singleton pregnancies to 10% in twins and 50% in triplets. If the unprotected vessels pass over the cervical os, this is termed a vasa previa. The incidence of vasa previa is 1 in 5000 pregnancies. Velamentously inserted vessels need not pass over the os to rupture, although the risk for rupture is greatest with a vasa previa. Rupture of a fetal vessel necessitates immediate abdominal delivery. Vasa previa alone carries a perinatal mortality rate of 50%, which increases to 75% if the membranes rupture.

image Postpartum Hemorrhage

Postpartum hemorrhage is defined as blood loss in excess of 500 mL at the time of vaginal delivery. There is normally a greater blood loss following cesarean delivery; therefore, blood loss in excess of 1000 mL is considered a postpartum hemorrhage in these patients. The excessive blood loss usually occurs in the immediate postpartum period but can occur slowly over the first 24 hours. Delayed postpartum hemorrhage can occasionally occur, with the excessive bleeding commencing more than 24 hours after delivery. This is usually a result of subinvolution of the uterus and disruption of the placental site “scab” several weeks postpartum, or of the retention of placental fragments that separate several days after delivery. Postpartum hemorrhage occurs in about 4% of deliveries.


Most of the blood loss occurs from the myometrial spiral arterioles and decidual veins that previously supplied and drained the intervillous spaces of the placenta. As the contractions of the partially empty uterus cause placental separation, bleeding occurs and continues until the uterine musculature contracts around the blood vessels and acts as a physiologic-anatomic ligature. Failure of the uterus to contract after placental separation (uterine atony) leads to excessive placental site bleeding. Other causes of postpartum hemorrhage are listed in Box 10-3.


BOX 10-3 Causes of Postpartum Hemorrhage.

Uterine atony

Genital tract trauma

Retained placental tissue

Low placental implantation

Uterine inversion

Coagulation disorders

Abruptio placentae

Amniotic fluid embolism

Retained dead fetus

Inherited coagulopathy



Most postpartum hemorrhages (75% to 80%) are due to uterine atony. The factors predisposing to postpartum uterine atony are listed in Box 10-4.


BOX 10-4 Factors Predisposing to Postpartum Uterine Atony.

Overdistention of the uterus

Multiple gestations


Fetal macrosomia

Prolonged labor

Oxytocic augmentation of labor

Grand multiparity (a parity of five or more)

Precipitous labor (lasting <3 hr)

Magnesium sulfate treatment of preeclampsia


Halogenated anesthetics

Uterine leiomyomas



Trauma during delivery is the second most common cause of postpartum hemorrhage. During vaginal delivery, lacerations of the cervix and vagina may occur spontaneously, but they are more common following the use of forceps or a vacuum extractor. The vascular beds in the genital tract are engorged during pregnancy, and bleeding can be profuse. Lacerations are particularly prone to occur over the perineal body, in the periurethral area, and over the ischial spines along the posterolateral aspects of the vagina. The cervix may lacerate at the two lateral angles while rapidly dilating in the first stage of labor. Uterine rupture may occasionally occur. At the time of delivery by low transverse cesarean, an inadvertent lateral extension of the incision can damage the ascending branches of the uterine arteries; an extension inferiorly can damage the cervical branches of the uterine artery.


In about half of patients with delayed postpartum hemorrhage, placental fragments are present when uterine curettage is performed with a large curette. Bleeding occurs as the uterus is unable to maintain a contraction and involute normally around a retained placental tissue mass.


Low implantation of the placenta can predispose to postpartum hemorrhage because the relative content of musculature in the uterine wall decreases in the lower uterine segment, which may result in insufficient control of placental site bleeding. Verifying a completely evacuated lower genital tract, a fully drained bladder and use of uterotonic agents such as pitocin, methylergonovine, or prostaglandin agents is usually sufficient. If bleeding continues, surgical management must be considered.


Peripartum coagulation disorders are high-risk factors for postpartum hemorrhage but fortunately are quite rare.

Patients with thrombotic thrombocytopenia have a rare syndrome of unknown etiology characterized by thrombocytopenic purpura, microangiopathic hemolytic anemia, transient and fluctuating neurologic signs, renal dysfunction, and a febrile course. In pregnancy, the disease is usually fatal. An amniotic fluid embolus is also rare and is associated with an 80% mortality rate. This syndrome is characterized by a fulminating consumption coagulopathy, intense bronchospasm, and vasomotor collapse. It is triggered by an intravascular infusion of a significant amount of amniotic fluid during a tumultuous or rapid labor in the presence of ruptured membranes. During the process of placental abruption, a small amount of amniotic fluid may leak into the vascular system, and the thromboplastin in the amniotic fluid may trigger a consumption coagulopathy. Patients with idiopathic thrombocytopenic purpura have platelets with abnormal function or a shortened life span. This causes thrombocytopenia and a tendency to bleed. Circulating antiplatelet antibodies of the IgG type may occasionally cross the placenta and result in fetal and neonatal thrombocytopenia as well. Von Willebrand’s disease is an inherited coagulopathy characterized by a prolonged bleeding time due to factor VIII deficiency. During pregnancy, these patients are likely to have a decreased bleeding diathesis because pregnancy elevates factor VIII levels. In the postpartum period, they are susceptible to delayed bleeding as factor VIII levels fall.


Uterine inversion is the “turning inside-out” of the uterus in the third stage of labor. It is quite rare, occurring in only about 1 in 20,000 pregnancies. Just after the second stage, the uterus is somewhat atonic, the cervix open, and the placenta attached. Improper management of the third stage of labor can cause an iatrogenic uterine inversion. If the inexperienced physician exerts fundal pressure while pulling on the umbilical cord before complete placental separation (particularly with a fundal implantation of the placenta), uterine inversion may occur. As the fundus of the uterus moves through the vagina, the inversion exerts traction on peritoneal structures, which can elicit a profound vasovagal response. The resulting vasodilation increases bleeding and the risk for hypovolemic shock. If the placenta is completely or partially separated, the uterine atony may cause profuse bleeding, which compounds the vasovagal shock.

image Obstetric Shock

Hypotension without significant external bleeding may occasionally develop in an obstetric patient. This condition is called obstetric shock. The causes of obstetric shock include concealed hemorrhage, uterine inversion, and amniotic fluid embolism.

An improperly sutured episiotomy can lead to a concealed postpartum hemorrhage. If the first suture at the vaginal apex of the episiotomy incision does not incorporate the cut and retracted arterioles, they can continue to bleed, creating a hematoma that can dissect cephalad into the retroperitoneal space. This may cause shock without external evidence of blood loss. A soft tissue hematoma, usually of the vulva, may occur following delivery in the absence of any laceration or episiotomy and may also contribute to occult blood loss.


Identification of the cause of postpartum hemorrhage requires a systematic approach. The fundus of the uterus should be palpated through the abdominal wall to determine the presence or absence of uterine atony. Next, a quick but thorough inspection of the vagina and cervix should be performed to ascertain whether any lacerations may be compounding the bleeding problem. Any uterine inversion or pelvic hematoma should be excluded during the pelvic examination. If the cause of bleeding has not been identified, manual exploration of the uterine cavity should be performed, under general anesthesia if necessary. With fingertips together, a gloved hand is slipped through the open cervix, and the hand is inserted into the uterus. The endometrial surface is palpated carefully to identify any retained products of conception, uterine wall lacerations, or partial uterine inversion. If no cause for the bleeding is found, a coagulopathy must be considered.

image Management of Postpartum Hemorrhage and Obstetric Shock

The first steps toward good management are the identification of patients at risk for postpartum hemorrhage and the institution of prophylactic measures during labor to minimize the possibility of maternal mortality. Patients with any predisposing factors for postpartum hemorrhage, including a history of postpartum hemorrhage, should be screened for anemia and atypical antibodies to ensure that an adequate supply of type-specific blood is available. An intravenous infusion through a large-bore needle or catheter should be started before delivery, and blood should be held in the laboratory for possible crossmatching.

During the diagnostic workup of an established hemorrhage, the patient’s vital signs must be monitored closely. Multiple units of packed red blood cells must be typed and crossmatched and intravenous crystalloids (such as normal saline or lactated Ringer’s solution) infused to restore intravascular volume. Resuscitation with normal saline usually requires a volume of 3 times the estimated blood loss.


If uterine atony is determined to be the cause of the postpartum hemorrhage, a rapid continuous intravenous infusion of dilute oxytocin (40 to 80 U in 1 L of normal saline) should be given to increase uterine tone. If the uterus remains atonic and the placental site bleeding continues during the oxytocic infusion, ergonovine maleate or methylergonovine, 0.2 mg, may be given intramuscularly. The ergot drugs are contraindicated in patients with hypertension because the pressor effect of the drug may increase blood pressure to dangerous levels.

Analogues of prostaglandin F given intramuscularly are quite effective in controlling postpartum hemorrhage caused by uterine atony. The 15-methyl analogue (Hemabate) has a more potent uterotonic effect and longer duration of action than the parent compound. The expected time of onset of the uterotonic effect when the 15-methyl analogue (0.25 mg) is given intramuscularly is 20 minutes, whereas when injected into the myometrium, it may take up to 4 minutes.

Failing these pharmacologic treatments, a bimanual compression and massage of the uterine corpus may control the bleeding and cause the uterus to contract. Although packing the uterine cavity is not widely practiced, it may occasionally control postpartum hemorrhage and obviate the need for surgical intervention. Alternatively, a large-volume balloon catheter has been developed that performs a similar function while maintaining a channel into the uterine cavity, allowing further bleeding to be monitored.

If uterine bleeding persists in an otherwise stable patient, she can be transported to the angiocatheterization laboratory, where radiologists can place an angiocatheter into the uterine arteries for injection of thrombogenic materials to control blood flow and hemorrhaging.

Operative intervention is a last resort. If the patient has completed her childbearing, a supracervical or total abdominal hysterectomy is definitive therapy for intractable postpartum hemorrhage caused by uterine atony. When reproductive potential is important to the patient, ligation of the uterine arteries adjacent to the uterus will lower the pulse pressure. This procedure is more successful in controlling placental site hemorrhage and is easier to perform than bilateral hypogastric artery ligation.


When postpartum hemorrhage is related to genital tract trauma, surgical intervention is necessary. When repairing genital tract lacerations, the first suture must be placed well above the apex of the laceration to incorporate any retracted bleeding arterioles into the ligature. Repair of vaginal lacerations requires good light and good exposure, and the tissues should be approximated without dead space. A running lock suture technique provides the best hemostasis (Figure 10-2). Cervical lacerations need not be sutured unless they are actively bleeding. Large, expanding hematomas of the genital tract require surgical evacuation of clots and a search for bleeding vessels that can be ligated, then packed for hemostasis. Stable hematomas can be observed and treated conservatively. A retroperitoneal hematoma generally begins in the pelvis. If the bleeding cannot be controlled from a vaginal approach, a laparotomy and bilateral hypogastric artery ligation may be necessary.


FIGURE 10-2 Suturing a cervical laceration. The first suture must be placed above the apex of the laceration.

The intraoperative laceration of the ascending branch of the uterine artery during delivery through a low transverse incision can be easily controlled by the placement of a large suture ligature through the myometrium and broad ligament below the level of the laceration. A uterine rupture usually necessitates subtotal or total abdominal hysterectomy, although small defects may be repaired.


When the placenta cannot be delivered in the usual manner, manual removal is necessary (Figure 10-3). This should be performed urgently if bleeding is profuse. Otherwise, it is reasonable to delay 30 minutes to await spontaneous separation. General anesthesia may be required. Following manual removal of the placenta or placental remnants, the uterus should be scraped with a large curette.


FIGURE 10-3 Manual removal of the placenta. The abdominal hand provides counterpressure on the uterine fundus against the shearing force of the fingers in the uterus.


The management of a uterine inversion requires quick thinking. The patient rapidly goes into shock, and immediate intravascular volume expansion with intravenous crystalloids is required. An anesthesiologist should be present. When the patient’s condition is stable, the partially separated placenta should be completely removed and an attempt made to replace the uterus by placing a cupped hand around the fundus and elevating it in the long axis of the vagina. If this is unsuccessful, a further attempt should be made using IV nitroglycerin (100 μg) or general anesthesia to relax the uterine muscle. Once replaced, a dilute oxytocin infusion should be started to cause the uterus to contract before removing the intrauterine hand. Rarely, the uterus cannot be replaced from below, and a surgical procedure may be required. At laparotomy, a vertical incision should be made through the posterior portion of the cervix to incise the constriction ring and allow the fundus to be replaced into the peritoneal cavity. Suturing of the cervical incision completes this procedure.


The principal objectives of treatment for amniotic fluid embolism are to support the respiratory system, correct the shock, and replace the coagulation factors. This type of embolism requires immediate cardiopulmonary resuscitation, usually with mechanical ventilation; rapid volume expansion with an electrolyte solution; positive inotropic cardiac support; placement of a bladder catheter to monitor urine output; correction of the red cell deficit by transfusion with packed red blood cells; and reversal of the coagulopathy with the use of platelets, fibrinogen, and other blood components.


When postpartum hemorrhage is associated with coagulopathy, the specific defect should be corrected by the infusion of blood products, as outlined in Table 10-1 and Box 10-5Patients with thrombocytopenia require platelet concentrate infusions; those with von Willebrand’s disease require factor VIII concentrate or cryoprecipitate.


Blood Product

Volume (mL) in 1 Unit

Effect of Transfusion

Platelet concentrate


Increases platelet count by about 20,000 to 25,000



Supplies fibrinogen, factor VIII, and factor XIII (3-10 times more concentrated than the equivalent volume of fresh plasma)

Fresh-frozen plasma


Supplies all factors except platelets (1 g of fibrinogen)

Packed red blood cells


Raises hematocrit 3%-4%

 Quantity obtained from 1 Unit (500 mL) of fresh whole blood.


BOX 10-5 Laboratory Evaluation of Disseminated Intravascular Coagulation.

• Platelet count (normal range = 150-450 × 109/L): 1 unit of platelets will raise the platelet count by 5-10 × 109/L

• Plasma fibrinogen (normal range = 175-600 mg/dL): fresh frozen plasma (FFP): 1 unit = 1 g of fibrinogen; 4 units of FFP will raise the plasma fibrinogen by 5-10 mg/dL

• Cryoprecipitate: 1 bag = 0.25 g of fibrinogen; 16 bags raise the plasma fibrinogen by 5-10 mg/dL

• Fibrin split products: normal range = <0.05 μg/mL (D-dimer method)


A packed red cell infusion is given to a patient who has bled enough to drop the circulating red cell population sufficiently to compromise the delivery of oxygen to the tissues. Therefore, institution of blood transfusion is best judged by symptoms of oxygen deprivation rather than by some empirical hemoglobin level. No important physiologic impairment has been noted at hemoglobin levels as low as 6 to 8 g/dL (hematocrit, 18% to 24%). In general, a 1-U transfusion of packed red blood cells will increase the hemoglobin level by 1 g/dL (and the hematocrit by 3% to 4%).

Massive blood replacement (when total blood volume is replaced in a 24-hour period) may be associated with thrombocytopenia, prolonged prothrombin time (PT), and hypofibrinogenemia.Thrombocytopenia is the most common abnormality, so platelet transfusion following determination of a low platelet count is not an uncommon scenario. Fresh frozen plasma may be transfused for prolonged PT or hypofibrinogenemia.

image Puerperal Sepsis

Puerperal sepsis still accounts for significant postpartum maternal morbidity and mortality. Patients with a puerperal genital tract infection are susceptible to the development of septic shock, pelvic thrombophlebitis, and pelvic abscess.

Following a vaginal delivery, about 6% to 7% of women demonstrate febrile morbidity, defined as a temperature of 100.4°F (38°C) or higher that occurs for more than 2 consecutive days (exclusive of the first postpartum day) during the first 10 postpartum days. Following primary cesarean delivery, the incidence of febrile morbidity is about twice that following vaginal delivery. Most of these fevers are caused by endometritis.


The pathophysiology of puerperal sepsis is closely related to the various microbial inhabitants of the vagina and cervix. The vaginal flora during gestation resembles the nonpregnant state, although there is a trend toward isolating more Mycoplasma genitalis and anaerobic streptococci in the last trimester. Potentially pathogenic organisms can be cultured from the vagina in about 80% of pregnant women. These organisms include enterococci, hemolytic and nonhemolytic streptococci, anaerobic streptococci, enteric bacilli, pseudodiphtheria bacteria, and Neisseria species other than N. gonorrhoeae.Excessive overgrowth of these organisms during pregnancy is inhibited by the acidity of the vagina (pH 4 to 5), primarily as a result of the production of lactic acid by the lactobacilli.

The uterine cavity is normally free of bacteria during pregnancy. After parturition, the pH of the vagina changes from acidic to alkaline because of the neutralizing effect of the alkaline amniotic fluid, blood, and lochia, as well as the decreased population of lactobacilli. This change in pH favors an increased growth of aerobic organisms. About 48 hours postpartum, progressive necrosis of the endometrial and placental remnants produces a favorable intrauterine environment for the multiplication of anaerobic bacteria.

About 70% of puerperal infections are caused by anaerobic organisms. Most of these are anaerobic cocci (Peptostreptococcus, Peptococcus, and Streptococcus), although mixed infections with Bacteroides fragilis are encountered in up to one third of cases. Of the aerobic organisms, Escherichia coli is the most common pathogen, followed by enterococci. Puerperal infection from clostridia is rare.

Intrauterine staphylococcal infection is rare. This organism is frequently responsible for infection of perineal wounds and abdominal incisions. Trichomonas vaginalis and Candida albicans are frequent inhabitants of the vagina, but no connection with puerperal sepsis has been established. Mycoplasma organisms have been shown to contribute to puerperal endometritis.


Predisposing factors for the development of a puerperal genital tract infection are shown in Box 10-6.


BOX 10-6 Factors Predisposing to the Development of Puerperal Genital Tract Infection.

Poor nutrition and hygiene


Premature rupture of the membranes (PROM)

Prolonged rupture of the membranes

Prolonged labor

Frequent vaginal examinations during labor

Cesarean delivery

Forceps or vacuum delivery

Cervical or vaginal lacerations

Manual removal of the placenta

Retained placental fragments or fetal membranes


After delivery, the placental site vessels are clotted off, and there is an exudation of lymph-like fluid along with massive numbers of neutrophils and other white cells to form the lochia. Vaginal microorganisms readily enter the uterine cavity and may become pathogenic at the placental site, depending on such variables as the size of the inoculum, the local pH, and the presence or absence of devitalized tissue. The latter may include tissue incorporated in the suture line of a cesarean incision.

The normal body defense mechanisms usually prevent any progressive infection, but a breakdown of these defenses allows the bacteria to invade the myometrium. Further invasion into the lymphatics of the parametrium can cause lymphangitis, pelvic cellulitis, and the possibility of widespread infection from septic emboli.

Endomyoparametritis is a potentially life-threatening condition. It commonly begins with retention of secundines (placental and amniochorionic membrane fragments) that block the normal lochial flow, allowing accumulation of intrauterine lochia, which in turn changes the local pH and acts as a culture medium for bacterial growth. Unless normal lochial flow is established, bacterial invasion progresses.


Puerperal infection presents with a rising fever and increasing uterine tenderness on postpartum day 2 or 3. With the development of parametritis (pelvic cellulitis), the temperature elevation will be sustained, and signs of pelvic peritonitis may develop. Erratic temperature fluctuations and severe chills suggest bacteremia and dissemination of septic emboli, with the particular likelihood of spread to the lungs.

When the usual relative pelvic venous stasis is combined with a large inoculum of pathogenic anaerobic bacteria, a pelvic vein thrombophlebitis is likely to develop, usually on the right side of the pelvis. The clinical picture of pelvic thrombophlebitis is characterized by a persistent spiking fever for 7 to 10 days after delivery, despite antibiotic therapy.


Evaluation of a febrile postpartum patient should include a careful history and physical examination. Extrapelvic causes of fever, such as breast engorgement, mastitis, aspiration pneumonia, atelectasis, pyelonephritis, thrombophlebitis, or wound infection, should be excluded.

Although a pelvic examination is generally not helpful in diagnosing pelvic thrombophlebitis, it may allow the palpation of tender, thrombosed, and edematous ovarian, parauterine, or iliac veins. An abdominal pelvic computed tomography scan or ultrasonogram may be helpful. This diagnosis is usually made by exclusion, however, and by the prompt regression of fever following commencement of heparin anticoagulant therapy.

Before the institution of antibiotic therapy for puerperal endometritis, aerobic and anaerobic cultures should be obtained from the blood, endocervix, and uterine cavity, and a catheterized urine specimen obtained for culture.


A febrile puerperal patient with cessation of lochial flow should undergo a pelvic examination and removal of any secundines that may be occluding the cervical os.

The antibiotic treatment of puerperal infection usually follows two major principles. First, early antibiotic treatment should be instituted to confine and then eliminate the infectious process. Second, the antibiotics should provide anaerobic coverage because these organisms are involved in 70% of puerperal infections. Antibiotics should be continued for at least 48 hours after the patient becomes afebrile. Anaerobic organisms especially require prolonged chemotherapy for elimination.

Broad-spectrum antibiotics, such as ampicillin and the cephalosporins, are effective first-line drugs for mild and moderate cases of puerperal infection. When the infection is moderate to severe, a penicillin-aminoglycoside combination has traditionally been used as first-line therapy. The major pelvic pathogen resistant to this combination is Bacteroides fragilis, which is usually sensitive to clindamycin. the use of clindamycin with either an aminoglycoside or ampicillin will provide the best first-line coverage.

When pelvic thrombophlebitis or thromboembolism is suspected or clinically diagnosed, unfractionated heparin therapy should be instituted to increase the clotting time (Lee-White method) or activated prothrombin time to 2 to 3 times normal. Only 2 to 3 weeks of anticoagulant therapy are needed for uncomplicated pelvic thrombophlebitis. Patients with femoral thrombophlebitis require 4 to 6 weeks of heparin therapy followed by the administration of oral anticoagulants for a few months.

If the patient does not respond to heparin therapy and the clinical course is one of unrelenting fever and pelvic tenderness, a diagnosis of pelvic abscess must be considered. Diagnosis is made by pelvic examination and confirmed by pelvic ultrasonography or computed tomography scan. The finding of a tender, pelvic parametrial mass suggests an abscess. Ultrasonography will confirm that the mass is fluid-filled rather than solid. The presence of a pelvic abscess requires surgical drainage.


Adesiyun A.G. Septic postpartum uterine inversion. Singapore Med J. 2007;48:943-945.

American College of Obstetricians and Gynecologists: Postpartum hemorrhage. ACOG Practice Bulletin No. 76. Obstet Gynecol. 2006;108:1039-1047.

Castagnola D.E., Hoffman M.K., Carlson J., Flynn C. Necrotizing cervical and uterine infection in the postpartum period caused by group A streptococcus. Obstet Gynecol. 2008;111:533-535.

Dane B, Dane C: Maternal death after uterine rupture in an unscarred uterus: A case report. J Emerg Med 2008 Mar 31 (Epub ahead of print).

Maharaj D. Puerpueral pyrexia: A review. Parts 1 and 2. Obstet Gynecol Surv. 2007;62:393-406.

Papathanasiou K., Tolikas A., Dovas D., et al. Ligation of internal iliac artery for severe obstetric and pelvic hemorrhage: 10 Year experience with 11 cases is a university hospital. J Obstet Gynaecol Res. 2008;28:183-184.