Peripartum hysterectomy refers to surgical removal of the pregnant or recently pregnant uterus. Most procedures follow delivery and are prompted by pregnancy or delivery complications. However, the term also includes hysterectomy with the pregnancy in situ, which is much less frequently performed. Radical hysterectomy early in the second trimester for cervical cancer is one example.

This chapter focuses on peripartum hysterectomy including cesarean hysterectomy and postpartum hysterectomy. These procedures are indispensable for management of intractable obstetric hemorrhage unresponsive to other treatment. In the puerperium, advanced uterine infection with necrosis is another indication.

Peripartum hysterectomy frequently is lifesaving and should be within the capabilities of all obstetric consultants. That said, hysterectomy in these circumstances can be a formidable operation, particularly when performed for a life-threatening emergency. Skills necessary for its performance are best acquired from an experienced mentor.

Pre-Porro Era

Prior to the 19th century, cesarean delivery was uniformly fatal. Thus, peripartum hysterectomy was developed in the late 18th century to improve survival rates. Until that time, bleeding and infections were untreatable complications. Anesthetics were limited, and surgical antisepsis was virtually nonexistent. Moreover, most surgeons lacked the expertise to perform such massive pelvic surgery. Unfortunately, many graduating obstetrician-gynecologists also lack the technical skills needed to perform this operation, and thus further mentoring is essential.

The evolution of peripartum hysterectomy can be divided into epochs before and after the description of hysterectomy technique by Eduardo Porro in 1876. Until the early 20th century, pelvic deformities from nutritional deficiencies or infectious diseases were common. Examples are rachitis (rickets) from vitamin D deficiency and tuberculosis. Women in that era had no reliable contraception, and mothers with a very small or distorted pelvis often labored to exhaustion and died along with their fetus. Some attendants used fetal destructive operations to attempt to save the mothers, who often had extensive pelvic soft tissue injuries if they survived.

Joseph Cavallini of Florence developed the concepts that enabled the development of obstetric hysterectomy. In 1768, he disproved the prevailing idea that the uterus was an essential organ for life by removing the uterus successfully in pregnant and nonpregnant animals (Durfee, 1969). Investigators in Germany and England in the early 1800s concluded that abdominal delivery in animals was less dangerous if they removed the uterus after delivery (Young, 1944). These experiments prepared the way for safe cesarean delivery and for obstetric hysterectomy.

The earliest documented human peripartum hysterectomy was performed in 1868 by Horatio Robinson Storer of Boston (Bixby, 1869). Storer was among the first specialists in diseases of women, and he was confronted by a woman whose labor was obstructed by a large uterine tumor. The baby was already dead, but the tumor prevented any fetal destructive procedures. Using chloroform for anesthesia, Storer performed cesarean delivery, and because of life-threatening hemorrhage, he removed the uterus with its “fibrocystic tumor the size of a baby’s head.” Ligatures around the cervix and the tumor pedicle controlled bleeding. The cervical stump was seared with a hot iron and returned to the pelvis, after which the abdominal wound was closed with silver wire. Unfortunately, the woman died 3 days later.

Eduardo Porro Era

In 1876, Porro published the first case report to describe a woman who survived hysterectomy after cesarean delivery. For this reason, some clinicians still refer to cesarean hysterectomy as the Porro operation. Porro had previously performed an emergency cesarean delivery in a woman with obstructed labor from a rachitic pelvis. Although she had died, he was determined to prove the safety of cesarean delivery and repeatedly rehearsed cesarean hysterectomy in an animal laboratory.

In his report, Porro carefully followed 19th-century surgical principles and techniques. The amphitheater was heated to a precise 18°C. The surgeons washed their hands with dilute carbolic acid. Chloroform anesthesia was induced, and a 12-cm vertical midline infraumbilical incision was made. A vertical uterine incision allowed delivery of a 3300-g female by version and extraction. Attempts at manual and suture control of bleeding after removal of the placenta were unsuccessful. He wrote: “It was providential that we had made all the preparations necessary for hysterectomy; otherwise the patient would surely have died.” An assistant elevated the uterus out of the abdominal wound. Porro slipped the wire noose of a Cintrat constrictor over the uterus and adnexa. The loop was tightened, and the uterus and adnexa were excised with a scalpel. He drained the cul-de-sac, treated the cervical stump with perchloride of iron, and exteriorized the Cintrat constrictor onto the abdominal wall. The abdominal wound was closed around these with silver wire mass sutures. The operation lasted 26 minutes.

The exteriorized constriction device was left on the cervical stump under dressings for 4 days. The wound sutures were removed on postoperative day 7. The ischemic cervical pedicle detached 7 days later. The woman left the hospital at 6 weeks.

Post-Porro Era

In 1880, Robert P. Harris reviewed the world literature on cesarean hysterectomy. He collected 50 cases from seven countries and reported a maternal mortality rate of 58 percent and a fetal survival rate of 86 percent. That same year, Isaac Taylor of New York performed the first Porro operation in the United States (Durfee, 1969). The woman survived the surgical procedure but died of a pulmonary embolism 3 weeks later. Soon after, Richardson (1878) reported the first maternal survival after a Porro operation in the United States. In 1884, Clement Godson reviewed a total of 134 cases. At that time, the most common indication for surgery still was a contracted pelvis due to rickets. He reported a maternal mortality rate of 45 percent. Patients usually succumbed to hemorrhage, peritonitis, and septicemia.

European surgeons modified the Porro procedure in an attempt to reduce operative risks and to improve outcomes. Tait (1890) of England described the Tait-Porro operation that enjoyed increasing popularity. Thereafter and into the 1900s, as surgical and anesthetic techniques improved, indications for the procedure were expanded beyond absolute emergencies.

The extensive experiences with cesarean hysterectomy at Charity Hospital in New Orleans began in 1938 and were reported by Barclay (1970) and Mickal and associates (1969). Subsequent ongoing Charity Hospital experiences were described periodically thereafter with steadily improving outcomes (Bey, 1993; Gonsoulin, 1991; Plauché, 1981, 1983).

Most obstetricians reserved peripartum hysterectomy for life-threatening emergencies. Davis (1951) of Chicago, however, championed elective cesarean hysterectomy and cited a 20-percent incidence of hysterectomy in a series of 736 cesarean deliveries. Brenner and colleagues (1970) concluded that complications outweighed benefits if hysterectomy was performed for sterilization and that tubal ligation after either cesarean or vaginal delivery was safer.

Currently, enthusiasm for nonemergency peripartum hysterectomy varies by region across the United States. Most programs reserve peripartum hysterectomy for major obstetric emergencies, but some perform obstetric hysterectomy when indications for cesarean delivery coexist with gynecologic indications for hysterectomy (Shellhaas, 2009). Currently, there is no codified list of indications for peripartum hysterectomy (Dahlke, 2015).

The cited rates for peripartum hysterectomy vary widely and range from 4 to 25 per 10,000 births. By way of example, from the Maternal-Fetal Medicine Units Network, the frequency of cesarean hysterectomy in nearly 185,000 deliveries was 10.1 per 10,000 births (Shellhaas, 2009). The rate at Parkland Hospital during a 22-year period was 17 per 10,000 births (Hernandez, 2013). Evidence indicates that the frequency of obstetric hysterectomy has risen during the past 20 years (Kramer, 2013). One source is Nationwide Inpatient Sample, which totaled more than 56 million births in the United States from 1994 through 2007 (Bateman, 2012). From this database, the rate of peripartum hysterectomy increased 15 percent during these 14 years (Fig. 26-1). This rising rate likely reflects the increasing rates of cesarean delivery and associated complications in a subsequent pregnancy (Bateman, 2012; Bodelon, 2009; Flood, 2009; Orbach, 2011; Owolabi, 2013).

Obstetric hysterectomy classification considers the timing, indication, extent, and circumstances associated with each case. Hysterectomy can be classified as antepartum, peripartum, or postpartum (Table 26-1). Although peripartum hysterectomy follows cesarean delivery in 75 to 80 percent of cases, 20 to 25 percent are performed in women who have been delivered vaginally (Jakobsson, 2015). These operations can further be classified as supracervical, total, or radical and can involve removal of one or both adnexa. Of peripartum hysterectomies performed in this country, half to two thirds are total and the remaining cases are supracervical (Rossi, 2010; Shellhaas, 2009). Finally, each operation may be identified as an emergent, an indicated nonemergent, or an elective case. Only when all these factors are considered can valid conclusions be drawn regarding comparative risks and outcomes.

An imposing list of factors that predispose to peripartum hysterectomy is found in Table 26-2. Many of these are maternal characteristics and antepartum features that can be identified and highlighted prior to delivery. One example is the placenta accrete syndromes, that is, placenta accreta, increta, and percreta, which are also known by the term morbidly adherent placenta. However, a significant number of factors are not evident or do not evolve until the time of labor and delivery. Thus, intrapartum surveillance for these risks is encouraged.

Some indications for obstetric hysterectomy are listed in Table 26-3. These are separated into procedures done for emergent indications and those performed for indicated nonemergent and elective operations.

Emergent Indications

As noted, the absolute rate of emergent peripartum hysterectomies is increasing. However, the proportions of indications for the operation have changed during the past 20 years. For example, arrest of hemorrhage remains the most common reason for obstetric hysterectomy. And, although the causes of hemorrhage have not changed over many decades, their individual contributions to cesarean hysterectomy rates now differ. Three examples are shown in Figure 26-1. Here, the rates of peripartum hysterectomy are increased for accrete syndromes and uterine atony, but not for placenta previa. During these times, morbidly adherent placenta has become much more common coincidental with the cesarean delivery rate. Cesarean delivery currently accounts for approximately one third of all births in the United States (Chap. 25, p. 404). Thus, the relationship between cesarean delivery and a morbidly adherent placenta, coupled with the rise in cesarean delivery rates, helps explain the new contribution from the accrete syndromes.

The evolution of these contributory indications is instructive. In 1964, Bowman and associates reported that 55 percent of cesarean hysterectomies at Charity Hospital were done for postpartum hemorrhage, and two thirds of these were for uterine atony. Almost 30 years later, uterine atony was the indication for only 21 percent of emergent peripartum hysterectomies performed at Brigham and Women’s Hospital (Zelop, 1993). At the same time, 64 percent of operations were done to control bleeding from a morbidly adherent placenta. This indication continues to be the most common in the United States, as shown in Figure 26-1. This has also been reported for several European countries (D’Arpe, 2015; Jakobsson, 2015).

At least before 1980, a commonly cited reason for postpartum hysterectomy was a disrupted prior hysterotomy scar. Although these indications included women with an overtly ruptured uterus, it also was contemporaneously acceptable to perform obstetric hysterectomy for a “defective uterine scar” in women with one or more prior cesarean deliveries. Indeed, in the earlier Charity Hospital series, almost half of hysterectomies were done for this indication (Plauché, 1981). Similarly, in a series of planned hysterectomies from private California hospitals, 28 percent were performed because of a defective uterine scar (McNulty, 1984). Today, although an occasional peripartum hysterectomy must be done for an irreparably damaged uterus from dehiscence of a prior uterine incision(s), this is no longer an accepted indication. Indeed, experience has accrued that thin scars or those with minor separations may be safely repaired rather than performing hysterectomy.

Severe uterine infections that follow cesarean delivery occasionally will require hysterectomy for removal of necrotic tissue. These usually manifest as uterine incisional necrosis and may be associated with peritonitis. This complication is further described in Chapter 32 (p. 512). Particularly virulent are infections caused by group A beta-hemolytic Streptococcus pyogenes (Fig. 26-2).

Nonemergent Indications

Peripartum hysterectomy is more controversial when performed for reasons other than obstetric emergencies. These indicated but nonemergent procedures are performed when women requiring cesarean delivery also have a gynecologic disorder that would usually be managed by hysterectomy. These have been variously termed nonemergent, planned, or anticipated cesarean hysterectomies (Plauché, 1992b; Zelop, 1993). The rationale is to perform both cesarean delivery and hysterectomy with a single procedure.

Of indications, uterine leiomyomas and cervical intraepithelial neoplasia are common (Plauché, 1992b). In the previously mentioned report, McNulty (1984) reviewed a 10-year experience with 80 planned cesarean hysterectomies performed in eight California private hospitals. The most prevalent diagnoses in this series were uterine leiomyomas—25 percent; defective scars after previous cesarean deliveries—28 percent; other benign gynecologic conditions—25 percent; and pelvic malignancy—5 percent. These are similar to the British experience reported by Sturdee and Rushton (1986). In another report contemporaneous to those times, abnormal gynecologic bleeding was the indication in 34 percent and chronic pelvic pain in 32 percent (Yancey, 1993).

Elective Cesarean Hysterectomy

Completely elective obstetric hysterectomies have never been endorsed by most. This was true even in the era of relative permissiveness for the operation (McNulty, 1984; Mickal, 1969; Plauché, 1986). In earlier series, although sterilization was the ostensible indication for peripartum hysterectomy, none of these authors encouraged primary abdominal delivery to create opportunities for surgical sterilization by hysterectomy.

Preoperative Preparation

Prior to hysterectomy, evaluation includes careful assessment of the woman for complicating conditions, hemodynamic stability, and coagulation status. The availability of appropriate blood replacement products is paramount, and transfusion usually begins before surgery in unstable patients. This is described in detail in Chapter 29 (p. 468). In patients with antepartum indications and confirmed plans for cesarean hysterectomy, perioperative considerations include patient consent, laboratory testing, antibiotic prophylaxis, venous thromboembolism prevention, and anesthesia selection. These topics are covered extensively in Chapter 18 (p. 291).

The surgical team is assembled, and hemodynamic support and anesthetic technique are discussed. Details of the operation are reviewed with the team to ensure maximum efficiency. Instruments and sutures are chosen before the operation begins. Additionally, requests for special instrumentation should be addressed preoperatively to prevent potential patient compromise and intraoperative delays.

A single suture, such as no. 1 or 0 chromic or polyglactin 910 (Vicryl) suture anchored to a general closure needle, usually suffices for most phases of the operation. Selection of a single type and gauge of suture expedites the work of both surgeon and circulating nurse. Some surgeons favor chromic suture, which tends to cut less than Vicryl through edematous tissue or engorged veins. However, the longer half-life of Vicryl may provide longer support of the vaginal apex during the healing phase. No robust data support superiority of one suture material compared with the other.

General Considerations

Surgery can be performed through either a vertical or low transverse abdominal wall incision. The steps for these incisions and their specific advantages are outlined in Chapter 4 (p. 49). Adequate visualization of the operative field is essential, and the pregnant uterus often restricts general examination of the abdomen early in the case.

The fastest laparotomy method is the subumbilical midline vertical incision. It offers suitable exposure to explore the entire abdomen, provides access to the pelvic sidewalls, and is easily extended cephalad if greater operating space is needed. Also, this midline incision encounters fewer vessels and thus may be preferred for patients with comorbid disseminated intravascular coagulopathy.

Several anatomic and physiologic changes in gravidas increase the likelihood of intraoperative problems during hysterectomy. First, the uterus is markedly enlarged and engorged, and pelvic blood vessels are wide-caliber and tortuous. Collateral vessels are similarly expanded. For this reason, rapid control of all vascular supply to the uterus is essential during hysterectomy. To assist, extra clamps are often necessary to avoid retrograde blood loss from severed blood vessels. As a reminder, when placing a clamp, the distal two thirds of most clamp jaws hold tissues and vessels with maximum security. Thus, positioning a clamp across too large of a tissue span increases the risk of tissue slippage and bleeding from escaped vessels. Another prudent practice is to manipulate clamps on vascular pedicles as little as possible. This also applies to clamps across pelvic tissues, which are typically edematous and more friable in gravidas than in nonpregnant women.

When clamped tissue pedicles are ligated, sutures may cut or tear through friable pedicles unless carefully secured. Thus, abrupt ligature cinching is avoided, but this is balanced against delayed cinching in which vessels can escape ligation and bleed. Also, pedicles shrink in size as edema subsides after delivery. Thus, single circumferential sutures may also allow vessels to escape control. Accordingly, some prefer double ligation of vascular pedicles. Moreover, use of a transfixing suture as the second distal ligature improves hemostatic control as edema subsides (Fig. 1-10, p. 13). Notably, hematomas can occasionally result if transfixing sutures are used alone without an initial circumferential proximal tie. Last, all pedicles should be inspected several times for hemostasis before the procedure’s end.

Clamps should be supported, not manipulated, during suture placement. Traction or twisting of clamps on vascular pedicles may strip pedicles away from adjacent tissues. This opens blood vessels that are difficult to clamp and suture without injury to nearby structures, particularly the ureter, bladder, and adnexa.

The bulky uterus is usually difficult to manipulate during obstetric surgery. Moreover, it often cannot be safely manipulated by means of clamps across friable adnexal structures at the cornua. In these cases, a tumor tenaculum placed on the uterine fundus is an efficient traction device. A finger in the apex of a vertical hysterotomy incision can also provide effective traction and manipulation. This maneuver, combined with handheld retractors—Deaver or Richardson—provides adequate exposure during the early stages of the operation.

The bladder wall may have become edematous and friable during labor. This may render it vulnerable to injury by blunt dissection. Handheld retractors that are swept from side to side under tension can also injure the bladder. Thus, assistants are encouraged to release traction before repositioning bladder retractors. A laparotomy sponge placed between the retractor blade and the bladder can also help avoid damage.

In some women with one or more prior cesarean deliveries, dense adhesions may fill the vesicouterine space between the bladder and the lower uterine segment. In these cases, if cesarean hysterectomy is planned, the bladder reflection is ideally carried down to the level of the cervix before the uterine incision is made. This time spent prior to hysterotomy and initiation of hysterectomy can help avoid cystotomy and substantial blood loss, which often accrues during tedious dissection in a blood-filled field. Metzenbaum scissors with the curve turned downward, away from the bladder, may best accomplish this meticulous sharp dissection. When blunt dissection is required, a Kittner dissector sponge or rolled umbilical tape in the tip of a Kelly clamp can be used with small, focused movements. It is important to avoid gross rolling or pushing motions with large sponges.

Bladder injuries are sometimes difficult to identify, but can lead to subsequent vesicovaginal fistula formation if undetected and unrepaired. One method used to confirm bladder integrity uses sterile colored or opaque solutions to fill the bladder and then search for small or hidden perforations of the bladder wall. Many find that infant formula from presterilized plastic nursery bottles is ideal. The circulating nurse attaches a milk-filled, 60-mL syringe to the urethral catheter and fills the bladder repeatedly to instill 200 to 300 mL of formula. As the bladder fills, milk is easily seen flowing from small defects in the bladder wall. Milk can be used repeatedly during a case because it does not stain tissues. If not available, methylene blue is a suitable dye to mix with saline for bladder filling. Indigo carmine was another popular choice, but current nationwide shortages limit its use now. Once identified, bladder defects are repaired promptly using a two-layer closure with absorbable suture as described and illustrated in Chapter 28 (p. 456).

The bladder and the ureter may be damaged or displaced by trauma, hematomas, or pelvic tumors. Recall that during pregnancy the ureter is dilated and sluggish. Its proximity should be observed at all phases of the operation. If there is any question of its safety, dissection to delineate its course is recommended. If necessary, an intentional cystotomy can be used to open the dome of the bladder. Following intravenous injection of methylene blue or indigo carmine dye, blue fluid jets should appear from both ureteral orifices within minutes. Notably, in the volume-depleted patient, jets may be delayed. If ureteral integrity remains in question, retrograde ureteral catheters can be inserted into the ureteral orifices under direct visualization to ensure ureteral integrity and outline their course. Any identified ureteral injury should be repaired and supported by standard methods along with expert intraoperative consultation (Chap. 28, p. 457). Also, in cases with sonographic findings that strongly suggest bladder invasion by the placenta, we often position the patient in low lithotomy position in booted support stirrups. This allows access to the perineum for stent placement and cystoscopy as needed.