CHAPTER 19: Ectopic Pregnancy

Following fertilization and fallopian tube transit, the blastocyst normally implants in the endometrial lining of the uterine cavity. Implantation elsewhere is considered ectopic and accounts for 0.5 to 1.5 percent of all first-trimester pregnancies in the United States (Hoover, 2011; Stulberg, 2014). This small proportion disparately accounts for 3 percent of all pregnancy-related deaths (Creanga, 2017). Fortunately, urine and serum beta-human chorionic gonadotropin (β-hCG) assays and transvaginal sonography allow earlier diagnosis. As a result, both maternal survival rates and conservation of reproductive capacity are improved.

Classification

Nearly 95 percent of ectopic pregnancies are implanted in the various segments of the fallopian tube. These segments are shown in Chapter 2 (Fig. 2-14). The ampulla (70 percent) is the most frequent site, followed by isthmic (12 percent), fimbrial (11 percent), and interstitial tubal pregnancies (2 percent) (Bouyer, 2002). The remaining 5 percent of nontubal ectopic pregnancies implant in the ovary, peritoneal cavity, cervix, or prior cesarean scar. Occasionally, a multifetal pregnancy contains one conceptus with normal uterine implantation that coexists with one implanted ectopically. The natural incidence of these heterotopic pregnancies approximates 1 per 30,000 pregnancies (Reece, 1983). However, with assisted reproductive technologies (ART), their incidence is 9 in 10,000 pregnancies (Perkins, 2015). Rarely, twin tubal pregnancy with both embryos in the same tube or with one in each tube has been reported (Eze, 2012; Goswami, 2015).

Regardless of location, D-negative women with an ectopic pregnancy who are not sensitized to D-antigen are given IgG anti-D immunoglobulin (American College of Obstetricians and Gynecologists, 2017). In first-trimester pregnancies, a 50-μg or 300-μg dose is appropriate, whereas a standard 300-μg dose is used for later gestations (Chap. 15, Prevention of Anti-D Alloimmunization).

Risks

Abnormal fallopian tube anatomy underlies many cases of tubal ectopic pregnancy. Surgeries for a prior tubal pregnancy, for fertility restoration, or for sterilization confer the highest risk. After one previous ectopic pregnancy, the chance of another is increased fivefold (Bhattacharya, 2012). Prior sexually transmitted disease or other tubal infection, which can distort normal tubal anatomy, is another factor. Specifically, one episode of salpingitis can be followed by a subsequent ectopic pregnancy in up to 9 percent of women (Westrom, 1992). Peritubal adhesions subsequent to salpingitis, appendicitis, or endometriosis can also increase chances. Salpingitis isthmica nodosa, which is a condition in which epithelium-lined diverticula extend into a hypertrophied muscularis layer, is another (Bolaji, 2015). Finally, congenital fallopian tube anomalies, especially those secondary to in utero diethylstilbestrol exposure, can predispose (Hoover, 2011).

Infertility, as well as the use of ART to overcome it, is linked to substantively increased risks for ectopic pregnancy (Clayton, 2006). With ART, the ectopic pregnancy rate in the United States between 2001 and 2011 was 1.6 percent (Perkins, 2015). And “atypical” implantations—cornual, abdominal, cervical, ovarian, and heterotopic pregnancy—are more frequent. Smoking is another known association, although the underlying mechanism is unclear (Hyland, 2015). Last, with any form of contraception, the absolute number of ectopic pregnancies is decreased because pregnancy occurs less often. However, with some contraceptive method failures, the relative number of ectopic pregnancies is increased. Examples include tubal sterilization, copper and progestin-releasing intrauterine devices (IUDs), and progestin-only contraceptives (Chap. 38).

Evolution and Potential Outcomes

With tubal pregnancy, because the fallopian tube lacks a submucosal layer, the fertilized ovum promptly burrows through the epithelium. The zygote comes to lie near or within the muscularis, which is invaded by rapidly proliferating trophoblast. The embryo or fetus in an ectopic pregnancy is often absent or stunted.

Outcomes of ectopic pregnancy include tubal rupture, tubal abortion, or pregnancy failure with resolution. With rupture, the invading expanding conceptus and associated hemorrhage can tear rents in the fallopian tube (Fig. 19-1). Tubal ectopic pregnancies usually burst spontaneously but may occasionally rupture following coitus or bimanual examination.

Alternatively, the pregnancy may pass out the distal fallopian tube. Tubal abortion frequency depends in part on the initial implantation site, and distal implantations are favored. Subsequently, hemorrhage may cease and symptoms eventually disappear. But bleeding can persist as long as products remain in the tube. Blood slowly trickles from the tubal fimbria into the peritoneal cavity and typically pools in the rectouterine cul-de-sac. If the fimbriated extremity is occluded, the fallopian tube may gradually become distended by blood, forming a hematosalpinx. Uncommonly, an aborted fetus will implant on a peritoneal surface and become an abdominal pregnancy, which is discussed in Abdominal Pregnancy.

Last, an unknown number of ectopic pregnancies spontaneously fail and are reabsorbed. This may be documented now more regularly with the advent of sensitive β-hCG assays.

Distinctions between “acute” ectopic pregnancy just described and “chronic” ectopic pregnancy can be drawn. The more common acute ectopic pregnancies are those with a high serum β-hCG level and rapid growth, leading to a timely diagnosis. These carry a higher risk of tubal rupture (Barnhart, 2003c). With chronic ectopic pregnancy, abnormal trophoblast dies early, and thus negative or low, static serum β-hCG levels are found (Brennan, 2000). Chronic ectopic pregnancies typically rupture late, if at all, but commonly form a complex pelvic mass, which often is the reason prompting diagnostic surgery (Cole, 1982; Uğur, 1996).

Clinical Manifestations

Earlier patient presentation and more precise diagnostic technology typically allow identification before rupture. In these cases, symptoms and signs of ectopic pregnancy are often subtle or even absent. The woman does not suspect tubal pregnancy and assumes that she has a normal early pregnancy or is having a miscarriage.

With later diagnosis, the classic triad is delayed menstruation, pain, and vaginal bleeding or spotting. With tubal rupture, lower abdominal and pelvic pain is usually severe and frequently described as sharp, stabbing, or tearing. Abdominal palpation elicits tenderness. Bimanual pelvic examination, especially cervical motion, causes exquisite pain. The posterior vaginal fornix may bulge from blood in the rectouterine cul-de-sac, or a tender, boggy mass may be felt beside the uterus. The uterus can also be slightly enlarged due to hormonal stimulation. Symptoms of diaphragmatic irritation, characterized by neck or shoulder pain, especially on inspiration, develop in perhaps half of women with sizable hemoperitoneum.

Some degree of vaginal spotting or bleeding is reported by 60 to 80 percent of women with tubal pregnancy. Although profuse vaginal bleeding suggests an incomplete abortion, such bleeding occasionally is seen with tubal gestations. Moreover, tubal pregnancy can lead to significant intraabdominal hemorrhage. Responses to moderate bleeding include no change in vital signs, a slight rise in blood pressure, or a vasovagal response with bradycardia and hypotension. Blood pressure will fall and pulse will rise only if bleeding continues and hypovolemia becomes significant. Vasomotor disturbances develop, ranging from vertigo to syncope.

Even after substantive hemorrhage, hemoglobin or hematocrit readings may at first show only a slight reduction. Hence, after an acute hemorrhage, a trending decline in hemoglobin or hematocrit levels over several hours is a more valuable index of blood loss than is the initial level. In approximately half of women with a ruptured ectopic pregnancy, varying degrees of leukocytosis up to 30,000/μL may be documented.

Decidua is endometrium that is hormonally prepared for pregnancy, and the degree to which the endometrium is converted with ectopic pregnancy is variable. Thus, in addition to bleeding, women with ectopic tubal pregnancy may pass a decidual cast. This is the entire sloughed endometrium that takes the form of the endometrial cavity (Fig. 19-2). Importantly, decidual sloughing may also occur with uterine abortion. Thus, tissue is carefully evaluated visually by the provider and then histologically for evidence of a conceptus. If no clear gestational sac is seen or if no villi are identified histologically within the cast, then the possibility of ectopic pregnancy must still be considered.

Multimodality Diagnosis

The differential diagnosis for abdominal pain coexistent with pregnancy is extensive. Pain may derive from uterine conditions such as miscarriage, infection, degenerating or enlarging leiomyomas, or round-ligament pain. Adnexal disease may include ectopic pregnancy; hemorrhagic, ruptured, or torsed ovarian masses; salpingitis; or tuboovarian abscess. Last, appendicitis, cystitis, renal stone, and gastroenteritis are more common nongynecological sources of lower abdominal pain in early pregnancy.

Several algorithms have been proposed to identify ectopic pregnancy. Most include these key components: physical findings, transvaginal sonography (TVS), serum β-hCG level measurement—both the initial and the subsequent pattern of rise or decline, and diagnostic surgery, which includes dilation and curettage (D&C), laparoscopy, and occasionally, laparotomy (Fig. 19-3). Algorithm use applies only to hemodynamically stable women, and those with presumed rupture undergo prompt surgical therapy. For a suspected unruptured ectopic pregnancy, all diagnostic strategies involve trade-offs. Strategies that maximize detection of ectopic pregnancy may result in termination of a normal intrauterine pregnancy (IUP). Conversely, those that reduce the potential for normal pregnancy interruption will delay ectopic pregnancy diagnosis. Patient desires for the index pregnancy are also discussed and may influence these trade-offs.

Beta-Human Chorionic Gonadotropin

Rapid and accurate determination of pregnancy is essential to identify an ectopic pregnancy. Current pregnancy tests use enzyme-linked immunosorbent assays (ELISAs) for the beta subunit of hCG. With these assays, lower limits of detection are 20 to 25 mIU/mL for urine and ≤5 mIU/mL for serum (Greene, 2015).

With bleeding or pain and a positive pregnancy test result, an initial TVS is typically performed to identify gestation location. If a yolk sac, embryo, or fetus is identified within the uterus or the adnexa, then a diagnosis can be made. In many cases, however, TVS is nondiagnostic, and tubal pregnancy is still a possibility. In these cases in which neither intrauterine nor extrauterine pregnancy is identified, the term pregnancy of unknown location (PUL) is used until additional clinical information allows determination of pregnancy location.

Levels above the Discriminatory Zone

Several investigators have described discriminatory β-hCG levels above which failure to visualize a uterine pregnancy indicates that the pregnancy either is not alive or is ectopic (Barnhart, 1994). Some institutions set their discriminatory threshold at ≥1500 mIU/mL, whereas others use ≥2000 mIU/mL. Connolly and associates (2013) suggested an even higher threshold. They noted that with live uterine pregnancies, a gestational sac was seen 99 percent of the time with a discriminatory level of >3510 mIU/mL.

If the initial β-hCG level exceeds the set discriminatory level and no evidence for an IUP is seen with TVS, then ectopic pregnancy is a concern. The diagnosis is narrowed in most cases to a failing IUP, a recent complete abortion, or an ectopic pregnancy. Early multifetal gestation also remains a possibility. Without clear evidence for ectopic pregnancy, serial β-hCG level assessment is reasonable, and a level is checked 48 hours later. This averts unnecessary methotrexate administration and avoids harming an early normal multifetal pregnancy. With greater concern for an ectopic gestation, D&C is another option to distinguish an ectopic from a failing IUP. Importantly, patient factors greatly influence these decisions.

Levels below the Discriminatory Zone

If the initial β-hCG level is below the set discriminatory value, pregnancy location is often not technically discernible with TVS. With these PULs, serial β-hCG level assays are done to identify patterns that indicate either a growing or failing IUP. Levels that rise or fall outside these expected parameters increase the concern for ectopic pregnancy. Thus, appropriately selected women with a possible ectopic pregnancy, but whose initial β-hCG level is below the discriminatory threshold, are seen 2 days later for further evaluation. Trends in levels aid diagnosis.

With early normal progressing IUPs, Barnhart and coworkers (2004b) reported a 53-percent 48-hour minimum rise with a 24-hour minimum rise of 24 percent. Seeber and associates (2006) found an even more conservative minimal 35-percent 48-hour rise in normal IUPs. With multifetal gestation, this same anticipated rate of rise is expected (Chung, 2006). Despite these guidelines, Silva and colleagues (2006) caution that a third of women with an ectopic pregnancy will have a 53-percent rise at 48 hours. They further reported that no single pattern characterizes ectopic pregnancy and that approximately half of ectopic pregnancies will show decreasing β-hCG levels, whereas the other half will have increasing levels. Also, despite a declining β-hCG level, a resolving ectopic pregnancy may rupture.

With a failing IUP, patterned rates of β-hCG level decline can also be anticipated. Following spontaneous abortion, rates decline by 21 to 35 percent at 48 hours and 68 to 84 percent at 7 days. Of note, these ranges reflect that β-hCG percentages drop faster if the initial β-hCG level is higher (Barnhart, 2004a). With resolving PULs, Butts and coworkers (2013) found greater rates of decline that ranged from 35 to 50 percent at 48 hours and 66 to 87 percent at 7 days for starting hCG values between 250 and 5000 mIU/mL.

In pregnancies without these expected rises or falls in β-hCG levels, distinction between a nonliving IUP and an ectopic pregnancy may be aided by additional β-hCG levels (Zee, 2014). Again delay is balanced against the risk from rupture. D&C is an option and provides a quicker diagnosis balanced against normal pregnancy interruption. Before curettage, a second TVS examination may be indicated and may display new informative findings.

Serum Progesterone

A single serum progesterone measurement may clarify the diagnosis in a few cases (Stovall, 1989, 1992). A value exceeding 25 ng/mL excludes ectopic pregnancy with 92-percent sensitivity (Lipscomb, 1999a; Pisarska, 1998). Conversely, values <5 ng/mL are found in only 0.3 percent of normal progressing IUPs (Mol, 1998; Verhaegen, 2012). Thus, values <5 ng/mL suggest either a nonliving IUP or an ectopic pregnancy. Because in most ectopic pregnancies, progesterone levels range between 10 and 25 ng/mL, the clinical utility of this practice is limited. One caveat is that pregnancy achieved with assisted reproductive technology may be associated with higher than usual progesterone levels (Perkins, 2000).

Transvaginal Sonography

Endometrial Findings

In a woman in whom ectopic pregnancy is suspected, TVS is performed to look for findings indicative of uterine or ectopic pregnancy. During endometrial cavity evaluation, an intrauterine gestational sac is usually visible between 4½ and 5 weeks. The yolk sac appears between 5 and 6 weeks, and a fetal pole with cardiac activity is first detected at 5½ to 6 weeks (Fig. 9-3). With transabdominal sonography, these structures are visualized slightly later.

In contrast, with ectopic pregnancy, a trilaminar endometrial pattern can be diagnostic (Fig. 19-4). Its specificity is 94 percent, but with a sensitivity of only 38 percent (Hammoud, 2005). In addition, Moschos and Twickler (2008b) determined in women with a pregnancy of unknown location at presentation that no normal IUPs had a stripe thickness <8 mm.

Anechoic fluid collections, which might normally suggest an early intrauterine gestational sac, may also be seen with ectopic pregnancy. These include pseudogestational sac and decidual cyst. First, a pseudosac is a fluid collection between the endometrial layers and conforms to the cavity shape (see Fig. 19-4). If a pseudosac is noted, the risk of ectopic pregnancy is increased (Hill, 1990; Nyberg, 1987). Second, a decidual cyst is identified as an anechoic area lying within the endometrium but remote from the canal and often at the endometrial-myometrial border. Ackerman and colleagues (1993b) suggested that this finding represents early decidual breakdown and precedes decidual cast formation.

These two findings contrast with the intradecidual sign seen with uterine pregnancies. With this, an early gestational sac is seen as an anechoic sac eccentrically located within one of the endometrial stripe layers (Dashefsky, 1988). The American College of Obstetricians and Gynecologists (2016) advises caution in diagnosing an IUP in the absence of a definite yolk sac or embryo.

Adnexal Findings

The sonographic diagnosis of ectopic pregnancy rests on visualization of an adnexal mass separate from the ovary (Fig. 19-5). If fallopian tubes and ovaries are visualized and an extrauterine yolk sac, embryo, or fetus is identified, then an ectopic pregnancy is confirmed. In other cases, a hyperechoic halo or tubal ring surrounding an anechoic sac is seen (Nadim, 2017). Alternatively, an inhomogeneous adnexal mass is usually caused by hemorrhage within the ectopic sac. Overall, approximately 60 percent of ectopic pregnancies are seen as an inhomogeneous mass adjacent to the ovary; 20 percent appear as a hyperechoic ring; and 13 percent have an obvious gestational sac with a fetal pole (Condous, 2005). Importantly, not all adnexal masses represent an ectopic pregnancy, and integration of sonographic findings with other clinical information is necessary.

Placental blood flow within the periphery of the complex adnexal mass—the ring of fire—can be seen with transvaginal color Doppler imaging. Although this can aid diagnosis, this finding can also be seen with a corpus luteum cyst, and differentiation can be challenging.

Hemoperitoneum

In affected women, blood in the peritoneal cavity is most often identified using sonography, but assessment can also be made by culdocentesis (Fig. 19-6). Sonographically, anechoic or hypoechoic fluid initially collects in the dependent retrouterine cul-de-sac, and then additionally surrounds the uterus as it fills the pelvis. As much as 50 mL of blood can be seen in the cul-de-sac using TVS, and transabdominal imaging then is used to assess the hemoperitoneum extent. Importantly, however, a small amount of peritoneal fluid is physiologically normal. With significant intraabdominal hemorrhage, blood will track up the pericolic gutters to fill Morison pouch near the liver. Free fluid in this pouch typically is not seen until accumulated volumes reach 400 to 700 mL (Branney, 1995; Rodgerson, 2001; Rose, 2004). Diagnostically, peritoneal fluid in conjunction with an adnexal mass is highly predictive of ectopic pregnancy (Nyberg, 1991). Ascites from ovarian or other cancer is a notable mimic.

Culdocentesis is a simple technique used commonly in the past. The cervix is pulled outward and upward toward the symphysis with a tenaculum, and a long 18-gauge needle is inserted through the posterior vaginal fornix into the retrouterine cul-de-sac. If present, fluid can be aspirated. However, a failure to do so is interpreted only as unsatisfactory entry into the cul-de-sac. Fluid containing fragments of old clots or bloody fluid that does not clot suggests hemoperitoneum. In contrast, if the blood sample clots, it may have been obtained from an adjacent blood vessel or from a briskly bleeding ectopic pregnancy. Several studies have challenged its usefulness, and culdocentesis has been largely replaced by TVS (Glezerman, 1992; Vermesh, 1990).

Endometrial Sampling

Several endometrial changes accompany ectopic pregnancy, and all lack coexistent trophoblast. Decidual reaction is found in 42 percent of samples, secretory endometrium in 22 percent, and proliferative endometrium in 12 percent (Lopez, 1994). Some recommend that the absence of trophoblastic tissue be confirmed by D&C before methotrexate treatment is given (Chung, 2011; Shaunik, 2011). Investigators found that the presumptive diagnosis of ectopic pregnancy is inaccurate in nearly 40 percent of cases without histological exclusion of a spontaneous pregnancy loss. Nevertheless, the risks of D&C are weighed against the limited maternal risks of methotrexate.

Endometrial biopsy with a Pipelle catheter was studied as an alternative to D&C and found inferior (Barnhart, 2003b; Ries, 2000). By comparison, frozen section of curettage fragments to identify products of conception is accurate in more than 90 percent of cases (Barak, 2005; Li, 2014b).

Laparoscopy

Direct visualization of the fallopian tubes and pelvis by laparoscopy offers a reliable diagnosis in most cases of suspected ectopic pregnancy. This also permits a ready transition to definitive operative therapy, which is discussed in Surgical Management.

Medical Management

Regimen Options

Medical therapy traditionally involves the antimetabolite methotrexate (MTX). This drug is a folic acid antagonist. It tightly binds to dihydrofolate reductase, blocking the reduction of dihydrofolate to tetrahydrofolate, which is the active form of folic acid. As a result, de novo purine and pyrimidine synthesis is halted, which leads to arrested DNA, RNA, and protein synthesis. Thus, MTX is highly effective against rapidly proliferating tissue such as trophoblast. Overall, ectopic tubal pregnancy resolution rates approximate 90 percent with its use. The drawbacks, however, are that bone marrow, gastrointestinal mucosa, and respiratory epithelium can also be harmed. It is directly toxic to hepatocytes and is renally excreted. MTX is also a potent teratogen, and MTX embryopathy is notable for craniofacial and skeletal abnormalities and fetal-growth restriction (Nurmohamed, 2011). In addition, MTX is excreted into breast milk and may accumulate in neonatal tissues and interfere with neonatal cellular metabolism (American Academy of Pediatrics, 2001; Briggs, 2015). Based on all these findings, a list of contraindications and pretherapy laboratory testing is found in Table 19-1.

Of precautions, MTX is bound primarily to albumin, and its displacement by other medications such as phenytoin, tetracyclines, salicylates, and sulfonamides can increase MTX serum drug levels. Moreover, renal clearance of MTX may be impaired by nonsteroidal antiinflammatory drugs including aspirin, probenecid, or penicillins (Stika, 2012). Last, vitamins containing folic acid may lower MTX efficacy.

For ease and efficacy, intramuscular MTX administration is used most often for ectopic pregnancy medical resolution, and single-dose and multidose MTX protocols are available (see Table 19-1). As noted, MTX can lead to bone marrow depression. This toxicity can be blunted by early administration of leucovorin, which is folinic acid and has activity equivalent to folic acid. Thus, leucovorin, which is given within the multidose protocol, allows for some purine and pyrimidine synthesis to buffer side effects.

In comparing these two protocols, trade-offs are recognized. For example, single-dose therapy offers simplicity, less expense, and less intensive posttherapy monitoring and does not require leucovorin rescue. However, some but not all studies report a higher success rate for the multidose regimen (Alleyassin, 2006; Barnhart, 2003a; Lipscomb, 2005). At our institution, we use single-dose MTX.

Patient Selection

The best candidate for medical therapy is the woman who is asymptomatic, motivated, and compliant. With medical therapy, some classic predictors of success include a low initial serum β-hCG level, small ectopic pregnancy size, and absent fetal cardiac activity. Of these, initial serum β-hCG level is the single best prognostic indicator of successful treatment with single-dose MTX. Specifically, reported failure rates are 1.5 percent if the initial serum β-hCG concentration is <1000 mIU/mL; 5.6 percent at 1000 to 2000 mIU/mL; 3.8 percent at 2000 to 5000 mIU/mL; and 14.3 percent when levels range between 5000 and 10,000 mIU/mL (Menon, 2007). Interestingly, the initial serum β-hCG value is not a valid indicator of the number of doses needed for successful resolution (Nowak-Markwitz, 2009).

Many early trials also used “large size” as an exclusion criterion, although these data are less precise. Lipscomb and colleagues (1998) reported a 93-percent success rate with single-dose MTX when the ectopic mass was <3.5 cm. This compared with success rates between 87 and 90 percent when the mass was >3.5 cm. Last, failure rates rise if cardiac activity is seen, with an 87-percent success rate in such cases.

Treatment Side Effects

These regimens are associated with minimal laboratory changes and symptoms, although occasional toxicity may be severe. Kooi and Kock (1992) reviewed 16 studies and reported that adverse effects resolved by 3 to 4 days after MTX was discontinued. The most common were liver involvement—12 percent; stomatitis— 6 percent; and gastroenteritis—1 percent. One woman had bone marrow depression. Fortunately, MTX treatment does not diminish ovarian reserve (Boots, 2016; Uyar, 2013). Moreover, conceptions within the first 6 months after MTX treatment for this indication are not associated with elevated rates of miscarriage or fetal malformations and growth restriction (Svirsky, 2009).

Importantly, 65 to 75 percent of women initially given MTX will have increasing pain beginning several days after therapy. Thought to reflect separation of the ectopic pregnancy from the tubal wall, this “separation pain” generally is mild and relieved by analgesics. In a series of 258 MTX-treated women by Lipscomb and colleagues (1999b), 20 percent had pain that merited evaluation in a clinic or emergency room. Ultimately, 10 of these 53 underwent surgical exploration. Said another way, 20 percent of women given single-dose MTX will have significant pain, and about 20 percent of these will require laparoscopy.

Monitoring Therapy Efficacy

As shown in Table 19-1, monitoring single-dose therapy calls for serum β-hCG determinations at days 4 and 7 following initial injection on day 1. After single-dose MTX, mean serum β-hCG levels may rise or fall during the first 4 days and then gradually decline. If the level fails to drop more than 15 percent between days 4 and 7, then a second dose of MTX is required. This is necessary in 15 to 20 percent of women treated with single-dose therapy (Cohen, 2014a; Kirk, 2007).

With multidose MTX, levels are measured at 48-hour intervals until they fall more than 15 percent. Up to four doses may be given to one patient if required (Stovall, 1991).

Once appropriately dropping levels are achieved in either regimen, serum β-hCG determinations are then measured weekly until undetectable. Outpatient monitoring is preferred, but if patient safety or compliance is questioned, the woman is hospitalized during initial surveillance. Lipscomb and colleagues (1998) used single-dose MTX to successfully treat 287 women and reported that the average time to resolution—defined as a serum β-hCG level <15 mIU/mL, was 34 days. Importantly, the longest time was 109 days.

Failure is judged when the β-hCG level plateaus or rises or the tube ruptures. Importantly, tubal rupture can occur even with declining β-hCG levels. Lipscomb and associates (1998) described a 14-day mean time to rupture, but one woman had tubal rupture 32 days after single-dose MTX.

From one metaanalysis, the overall success rate for treatment with MTX is 89 percent. The success for the multidose regiment is 92.7 percent, whereas that for single-dose is 88.1 percent (Barnhart, 2003a). Despite this difference, the single dose is more frequently used because of its simplicity and convenience.

Surgical Management

Studies have compared laparotomy with laparoscopic surgery for ectopic pregnancy (Lundorff, 1991; Murphy, 1992; Vermesh, 1989). Overall, tubal patency and number of subsequent uterine pregnancies do not differ between these routes. Thus, laparoscopy is the preferred surgical treatment for ectopic pregnancy unless a woman is hemodynamically unstable. As experience has accrued, cases previously managed by laparotomy—for example, ruptured tubal pregnancies with hemoperitoneum—can safely be managed laparoscopically by those with suitable expertise (Cohen, 2013; Sagiv, 2001). That said, the lowered venous return and cardiac output associated with the pneumoperitoneum of laparoscopy must be factored into the decision to select minimally invasive surgery for hypovolemic women.

Before surgery, future fertility desires are discussed. In women desiring permanent sterilization, the unaffected tube can be ligated or removed concurrently with salpingectomy for the affected fallopian tube.

Two procedures—salpingostomy or salpingectomy—are options. Two multicenter, randomized controlled trials have compared laparoscopic outcomes between the two procedures in women with a normal contralateral fallopian tube. The European Surgery in Ectopic Pregnancy (ESEP) study randomized 231 women to salpingectomy and 215 to salpingostomy. After surgery, subsequent rates of ongoing pregnancy by natural conception did not differ significantly between groups—56 versus 61 percent, respectively (Mol, 2014). Again, in the DEMETER trial, the subsequent 2-year rate for achieving a uterine pregnancy did not differ between groups—64 versus 70 percent, respectively (Fernandez, 2013). In women with an abnormal-appearing contralateral tube, salpingostomy is a conservative option for fertility preservation.

Salpingostomy

This procedure is typically used to remove a small unruptured pregnancy. A 10- to 15-mm linear incision is made on the antimesenteric border of the fallopian tube over the pregnancy. The products usually will extrude from the incision. These can be carefully removed or flushed out using high-pressure irrigation that more thoroughly removes the trophoblastic tissue (Al-Sunaidi, 2007). Small bleeding sites are controlled with needlepoint electrocoagulation, and the incision is left unsutured to heal by secondary intention. Serum β-hCG levels are used to monitor response to both medical and surgical therapy. After linear salpingostomy, serum β-hCG levels decline rapidly over days and then more gradually, with a mean resolution time of approximately 20 days.

Seldom performed today, salpingotomy is essentially the same procedure except that the incision is closed with delayed-absorbable suture. According to Tulandi and Guralnick (1991), prognosis does not differ with or without suturing, and laparoscopic suturing adds surgical time.

Salpingectomy

Tubal resection may be used for both ruptured and unruptured ectopic pregnancies. To minimize the rare recurrence of pregnancy in the tubal stump, complete excision of the fallopian tube is advised. With one laparoscopic technique, the affected fallopian tube is lifted and held with atraumatic grasping forceps (Thompson, 2016). One of several suitable bipolar grasping devices is placed across the fallopian tube at the uterotubal junction. Once desiccated, the tube is cut. The bipolar device is then advanced across the most proximal portion of mesosalpinx. Similarly, current is applied, and the desiccated tissue cut. This process moves serially from the proximal mesosalpinx to its distal extent under the tubal ampulla. Alternatively, an endoscopic suture loop can be used to encircle and ligate the knuckle of involved fallopian tube and its underlying vascular supply within the mesosalpinx. Two consecutive suture loops are placed, and the tube distal to these ligatures is then cut free with scissors. Salpingectomy during laparotomy is shown in Chapter 39 (Nonpuerperal Tubal Sterilization).

Most tubal ectopic pregnancies are small and pliant. Accordingly, they can be held firmly by grasping forceps and drawn up into one of the accessory site cannulas. Larger tubal ectopic pregnancies may be placed in an endoscopic sac to prevent fragmentation as they are removed through the laparoscopic port site. Importantly, to remove all trophoblastic tissue, the pelvis and abdomen should be irrigated and suctioned free of blood and tissue debris. Slow and systematic movement of the patient from Trendelenburg to reverse Trendelenburg positioning during irrigation can also assist in dislodging stray tissue and fluid. These should be suctioned and removed from the peritoneal cavity.

Persistent Trophoblast

After surgery, β-hCG levels usually fall quickly and approximate 10 percent of preoperative values by day 12 (Hajenius, 1995; Vermesh, 1988). Persistent trophoblast is rare following salpingectomy, but complicates 5 to 15 percent of salpingostomies (Kayatas, 2014; Pouly, 1986; Seifer, 1993). Rates are lower for laparotomy versus laparoscopic procedures (Hajenius, 1995). Other risk factors are debatable but may include greater serum β-hCG levels and smaller ectopic size (Rabischong, 2010; Seifer, 1997). Bleeding caused by retained trophoblast is the most serious complication.

Incomplete removal of trophoblast can be identified by stable or rising β-hCG levels. Monitoring approaches are not codified. One scheme measures serum β-hCG levels on postoperative day 1, and values dropping <50 percent of the preoperative value reflect risk for persistent trophoblast (Spandorfer, 1997). Another measures weekly levels (Mol, 2008). With stable or increasing β-hCG levels, additional surgical or medical therapy is necessary. Without evidence for tubal rupture, standard therapy for this is single-dose MTX, 50 mg/m² × body surface area (BSA). Rupture and bleeding require surgical intervention.

Medical versus Surgical Therapy

Several randomized trials have compared methotrexate treatment with laparoscopic surgery. One multicenter trial compared a multidose MTX protocol with laparoscopic salpingostomy and found no differences for tubal preservation and primary treatment success (Hajenius, 1997). In this same study group, however, health-related quality-of-life factors such as pain, posttherapy depression, and decreased perception of health were significantly impaired after systemic MTX compared with laparoscopic salpingostomy (Nieuwkerk, 1998). In their randomized controlled trial, Fernandez and coworkers (2013) compared multidose medical therapy against salpingostomy and found that medical and conservative surgery provided similar 2-year rates of attaining a uterine pregnancy.

Evidence is conflicting when single-dose MTX is compared with surgical intervention. In two separate studies, single-dose MTX was overall less successful in resolving pregnancy than laparoscopic salpingostomy, although tubal patency and subsequent uterine pregnancy rates were similar between both groups (Fernandez, 1998; Sowter, 2001). Women treated with MTX had significantly better physical functioning immediately following therapy, but there were no differences in psychological functioning. Krag Moeller and associates (2009) reported the results from their randomized trial that had a median surveillance period of 8.6 years during which future pregnancy rates were evaluated. Ectopic-resolution success rates were not significantly different between those managed surgically and those treated with MTX. Moreover, cumulative spontaneous uterine pregnancy rates were not different between the MTX group (73 percent) and the surgical group (62 percent).

Based on these studies, we conclude that women who are hemodynamically stable and in whom there is a small tubal diameter, no fetal cardiac activity, and serum β-hCG concentrations <5000 mIU/mL have similar outcomes with medical or surgical management. Despite lower success rates with medical therapy for women with larger tubal size, higher serum β-hCG levels, and fetal cardiac activity, medical management can be offered to the motivated woman who understands the risks.

Expectant Management

In select cases, it is reasonable to observe very early tubal pregnancies that are associated with stable or falling serum β-hCG levels. Mavrelos and coworkers (2013) noted that almost one third of 333 tubal ectopic pregnancies measuring <3 cm and with β-hCG levels <1500 mIU/mL resolved without intervention. Cohen and associates (2014b) similarly followed 674 women with declining β-hCG levels to successful resolution. These findings have been supported by smaller randomized trials (Jurkovic, 2017; van Mello, 2013).

With expectant management, subsequent rates of tubal patency and intrauterine pregnancy are comparable with surgical or medical management. That said, compared with the established safety of medical and surgical therapy, the prolonged surveillance and risks of tubal rupture support the practice of expectant therapy only in appropriately selected and counseled women.

Diagnosis

An interstitial pregnancy is one that implants within the proximal tubal segment that lies within the muscular uterine wall (Fig. 19-7). Incorrectly, they may be called cornual pregnancies, but this term describes a conception that develops in the rudimentary horn of a uterus with a müllerian anomaly (Moawad, 2010). Risk factors are similar to others discussed for tubal ectopic pregnancy, although previous ipsilateral salpingectomy is a specific risk factor for interstitial pregnancy (Lau, 1999). Undiagnosed interstitial pregnancies usually rupture following 8 to 16 weeks of amenorrhea, which is later than for more distal pregnancies. This is due to greater distensibility of the myometrium covering the interstitial fallopian tube segment. Because of the proximity of these pregnancies to the uterine and ovarian arteries, hemorrhage can be severe and associated with mortality rates as high as 2.5 percent (Tulandi, 2004).

With transvaginal sonography and serum β-hCG assays, interstitial pregnancy can now be diagnosed early in many cases, but diagnosis can be challenging. These pregnancies sonographically can appear similar to an eccentrically implanted uterine pregnancy, especially in a uterus with a müllerian anomaly. Criteria that may aid differentiation include: an empty uterus, a gestational sac seen separate from the endometrium and >1 cm away from the most lateral edge of the uterine cavity, and a thin, <5-mm myometrial mantle surrounding the sac (Timor-Tritsch, 1992). Moreover, an echogenic line, known as the “interstitial line sign,” extending from the gestational sac to the endometrial cavity most likely represents the interstitial portion of the fallopian tube and is highly sensitive and specific (Ackerman, 1993a). In unclear cases, three-dimensional (3-D) sonography, magnetic resonance (MR) imaging, or diagnostic laparoscopy can help clarify anatomy (Parker, 2012; Tanaka, 2014). Laparoscopically, an enlarged protuberance is found lying outside the round ligament and coexistent with a normal distal fallopian tube and ovary.

Management

Surgical management with either cornual resection or cornuostomy may be performed via laparotomy or laparoscopy, depending on patient hemodynamic stability and surgeon expertise (Hoffman, 2016; Zuo, 2012). With either approach, intraoperative intramyometrial vasopressin injection may limit surgical blood loss, and β-hCG levels should be monitored postoperatively to exclude remnant trophoblast. Cornual resection removes the gestational sac and surrounding cornual myometrium by means of a wedge excision (Fig. 19-8). Alternatively, cornuostomy involves incision of the cornua and suction or instrument extraction of the pregnancy. Both instances require myometrial closure.

With early diagnosis, medical management may be considered. But because of the low incidence, consensus regarding methotrexate regimens is lacking. In their small series, Jermy and associates (2004) reported a 94-percent success with systemic MTX using a dose of 50 mg/m² × BSA. Others have described direct MTX injection into the gestational sac (Framarino-dei-Malatesta, 2014). Importantly, because these women typically have higher initial serum β-hCG levels, longer surveillance is usually needed.

The risk of uterine rupture with subsequent pregnancies following either medical or surgical management is unclear. Thus, careful observation of these women during pregnancy, along with strong consideration of elective cesarean delivery, is warranted.

Distinct from interstitial pregnancy, the term angular pregnancy describes implantation within the endometrial cavity but at one cornu and medial to the uterotubal junction and round ligament. An angular pregnancy displaces the round ligament upward and outward, whereas an interstitial tubal pregnancy does not shift it (Arleo, 2014). This distinction is important because angular pregnancies can sometimes be carried to term but with increased risk of abnormal placentation and its consequences (Jansen, 1981).

Diagnosis

This term describes implantation within the myometrium of a prior cesarean delivery scar. Its incidence approximates 1 in 2000 normal pregnancies and has increased along with the cesarean delivery rate (Ash, 2007; Rotas, 2006). The pathogenesis of cesarean scar pregnancy (CSP) has been likened to that for placenta accreta and carries similar risk for serious hemorrhage (Timor-Tritsch, 2014a,b). It is unknown if the incidence increases with multiple cesarean deliveries or if it is affected by either one- or two-layer uterine incision closure during cesarean.

Women with CSP usually present early, and pain and bleeding are common. Still, up to 40 percent of women are asymptomatic, and the diagnosis is made during routine sonographic examination (Rotas, 2006). Sonographically, differentiating between a cervicoisthmic intrauterine pregnancy and CSP can be difficult (Moschos, 2008a; Timor-Tritsch, 2016). According to Godin (1997), four sonographic criteria should be satisfied for the diagnosis and are described in Figure 19-9. Although TVS is the typical first-line imaging tool, MR imaging is useful when sonography is inconclusive (Huang, 2014; Osborn, 2012).

Management

Treatment standards are lacking, and several options are available. Expected management is an option, and live birth rates were 57 percent in one review (Maheux-Lacroix, 2017). However, hemorrhage, placenta accreta, and uterine rupture are risks. Thus, hysterectomy is an acceptable initial choice in those desiring sterilization. It is sometimes necessary with heavy uncontrolled bleeding. Fertility-preserving options include systemic or locally injected methotrexate, either alone or combined with conservative surgery (Birch Petersen, 2016; Cheung, 2015). Surgical procedures include visually guided suction curettage, hysteroscopic removal, or isthmic excision done abdominally or vaginally. These are completed solely or with adjunctive MTX (Jurkovic, 2016; Li, 2014a; Wang, 2014; Yang, 2009). Often uterine artery embolization (UAE) is used preoperatively to minimize hemorrhage risk (Zhang, 2012; Zhuang, 2009). Foley balloon catheter placement can be another option for procedure-associated bleeding (Timor-Tritsch, 2015a).

Following conservative treatment, subsequent pregnancies have good outcomes, but placenta accreta and recurrent CSP are risks (Gao, 2016; Wang, 2015). Uterine arteriovenous malformations are a potential long-term complication (Timor-Tritsch, 2015b).

Diagnosis

This rare ectopic pregnancy is defined by cervical glands noted histologically opposite the placental attachment site and by all or part of the placenta found below the entrance of the uterine vessels or below the peritoneal reflection on the anterior uterus. In a typical case, the endocervix is eroded by trophoblast, and the pregnancy develops in the fibrous cervical wall. Predisposing risks include ART and prior uterine curettage (Ginsburg, 1994; Jeng, 2007).

Painless vaginal bleeding is reported by 90 percent of women with a cervical pregnancy—a third of these have massive hemorrhage (Ushakov, 1997). As pregnancy progresses, a distended, thin-walled cervix with a partially dilated external os may be evident. Above the cervical mass, a slightly enlarged uterine fundus can be felt. Identification of cervical pregnancy is based on speculum examination, palpation, and TVS. Sonographic findings typical of cervical pregnancy are shown and described in Figure 19-10. MR imaging and 3-D sonography have also been used to confirm the diagnosis (Jung, 2001; Sherer, 2008).

Management

Cervical pregnancy may be treated medically or surgically. Conservative management strives to minimize hemorrhage, resolve the pregnancy, and preserve fertility. In many centers, including ours, methotrexate has become the first-line therapy in stable women, and administration follows protocols listed in Table 19-1 (Verma, 2011; Zakaria, 2011). The drug has also been injected directly into the gestational sac, alone or with systemic doses (Jeng, 2007; Murji, 2015). Others describe MTX infusion combined with uterine artery embolization—”chemoembolization” (Xiaolin, 2010).

With MTX regimens, resolution and uterine preservation are achieved for gestations <12 weeks in 91 percent of cases (Kung, 1997). In selecting appropriate candidates, Hung and colleagues (1996) noted higher risks of systemic MTX treatment failure in those with a gestational age >9 weeks, β-hCG levels >10,000 mIU/mL, crown-rump length >10 mm, and fetal cardiac activity. For this reason, many induce fetal death with intracardiac or intrathoracic injection of potassium chloride. With a single-dose intramuscular protocol, an MTX dose between 50 and 75 mg/m² × BSA is typical. To resolve fetal cardiac activity, a sonographically guided fetal intracardiac injection of 2 mL (2 mEq/mL) potassium chloride solution can be given (Verma, 2009). Song and associates (2009) described management of 50 cases and observed that sonographic resolution lagged far behind serum β-hCG regression.

As an adjunct to medical or surgical therapy, uterine artery embolization has been described either as a response to bleeding or as a preprocedural preventive tool (Hirakawa, 2009; Zakaria, 2011). Also, in the event of hemorrhage, a 26F Foley catheter with a 30-mL balloon can be placed intracervically and inflated to effect hemostasis by vessel tamponade and to monitor uterine drainage. The balloon remains inflated for 24 to 48 hours and is gradually decompressed over a few days (Ushakov, 1997).

Although conservative management is feasible for many women with cervical pregnancies, suction curettage or hysterectomy may be selected. Moreover, hysterectomy may be required with bleeding uncontrolled by conservative methods. Because of the close proximity of the ureters to the ballooned cervix, urinary tract injury rates are a concern during hysterectomy.

If cervical curettage is planned, intraoperative bleeding may be lessened by preoperative UAE, by intracervical vasopressin injection, or by a cerclage placed at the internal cervical os to compress feeding vessels (Chen, 2015; Fylstra, 2014; Wang, 2011). Also, cervical branches of the uterine artery can effectively be ligated with vaginal placement of hemostatic cervical sutures on the lateral aspects of the cervix at 3 and 9 o’clock (Bianchi, 2011). Following curettage, a Foley balloon can be placed to tamponade bleeding and is managed as described earlier. Suction curettage may be especially favored in rare cases of a heterotopic pregnancy composed of a cervical and a desired uterine pregnancy (Tsakos, 2015).

Diagnosis

These rare ectopic pregnancies are defined as an implantation in the peritoneal cavity exclusive of tubal, ovarian, or intraligamentous implantations. Although a zygote can traverse the tube and implant primarily in the peritoneal cavity, most abdominal pregnancies are thought to follow early tubal rupture or abortion with reimplantation. In cases of advanced extrauterine pregnancy, it is not unusual for the placenta to be still at least partially attached to the uterus or adnexa.

Diagnosis may be difficult. First, symptoms may be absent or vague. Laboratory tests are typically uninformative, although maternal serum alpha-fetoprotein levels can be elevated. Clinically, abnormal fetal positions may be palpated, or the cervix is displaced (Zeck, 2007). Sonographically, the diagnosis is often missed (Costa, 1991). Oligohydramnios is common but nonspecific. Other clues include a fetus seen separate from the uterus or eccentrically positioned within the pelvis; lack of myometrium between the fetus and the maternal anterior abdominal wall or bladder; extrauterine placental tissue; or bowel loops surrounding the gestational sac (Allibone, 1981; Chukus, 2015). If additional anatomical information is needed, MR imaging can help confirm the diagnosis and provide maximal information concerning placental implantation (Bertrand, 2009; Mittal, 2012).

Management

Treatment of an abdominal pregnancy depends on the gestational age at diagnosis. Conservative management carries a maternal risk for sudden and dangerous hemorrhage. Moreover, Stevens (1993) reported fetal malformations and deformations in 20 percent. Thus, we believe that termination generally is indicated when the diagnosis is made. Certainly, before 24 weeks, conservative treatment rarely is justified. Despite this, some have described waiting until fetal viability with close surveillance (Kim, 2013; Marcellin, 2014).

Once placental implantation has been assessed, several options to control intraoperative hemorrhage mimic those used for placenta accrete syndrome (Chap. 41, Adjunctive Surgical Procedures). The principal surgical objectives involve delivery of the fetus and careful assessment of placental implantation without provoking hemorrhage. Unnecessary exploration is avoided because the anatomy is commonly distorted and surrounding areas are extremely vascular. Importantly, placental removal may precipitate torrential hemorrhage because the normal hemostatic mechanism of myometrial contraction to constrict hypertrophied blood vessels is lacking. If it is obvious that the placenta can be safely removed or if there is already hemorrhage from its implantation site, then removal begins immediately. When possible, blood vessels supplying the placenta should be ligated first.

Some advocate leaving the placenta in place as the lesser of two evils. It decreases the chance of immediate life-threatening hemorrhage, but at the expense of long-term sequelae. If left in the abdominal cavity, the placenta commonly becomes infected, with subsequent formation of abscesses, adhesions, intestinal or ureteral obstruction, and wound dehiscence (Bergstrom, 1998; Martin, 1988). In many of these cases, surgical removal becomes inevitable. If the placenta is left, its involution may be monitored using sonography and serum β-hCG levels (France, 1980; Martin, 1990). Color Doppler sonography can be used to assess changes in blood flow. In some cases, and usually depending on its size, placental function rapidly declines, and the placenta is resorbed. But placental resorption may take years (Roberts, 2005; Valenzano, 2003).

If the placenta is left in place, postoperative methotrexate use is controversial. It has been recommended to hasten involution but has been reported to cause accelerated placental destruction with accumulation of necrotic tissue and infection with abscess formation (Rahman, 1982). It is difficult to envision a supporting role for the use of an antimetabolite for a senescent organ (Worley, 2008).

Ectopic implantation of the fertilized egg in the ovary is rare and is diagnosed if four clinical criteria are met. These were outlined by Spiegelberg (1878): (1) the ipsilateral tube is intact and distinct from the ovary; (2) the ectopic pregnancy occupies the ovary; (3) the ectopic pregnancy is connected by the uteroovarian ligament to the uterus; and (4) ovarian tissue can be demonstrated histologically amid the placental tissue. Risk factors are similar to those for tubal pregnancies, but ART or IUD failure seems to be disproportionately associated (Zhu, 2014). Presenting complaints and findings mirror those for tubal ectopic pregnancy. Although the ovary can accommodate the expanding pregnancy more easily than the fallopian tube, rupture at an early stage is the usual consequence (Melcer, 2016).

Transvaginal sonography use has resulted in a more frequent diagnosis of unruptured ovarian pregnancies. Sonographically, an internal anechoic area is surrounded by a wide echogenic ring, which in turn is surrounded by ovarian cortex (Comstock, 2005). In their review of 49 cases, Choi and associates (2011) noted that the diagnosis may not be made until surgery, as many cases are presumed tubal ectopic pregnancy. Moreover, at surgery, an early ovarian pregnancy may be considered to be a hemorrhagic corpus luteum.

Evidence-based management accrues mainly from case reports (Hassan, 2012; Scutiero, 2012). Classically, management for ovarian pregnancies has been surgical. Small lesions can be managed by ovarian wedge resection or cystectomy, whereas larger lesions require oophorectomy (Elwell, 2015; Melcer, 2015). With conservative surgery, β-hCG levels should be monitored to exclude remnant trophoblast.

Pregnancy implanted toward the mesosalpinx may rupture into a space formed between the broad ligament leaves and become an intraligamentous or broad ligament pregnancy. Rents in prior cesarean scars serve as another conduit (Rudra, 2013). These are rare, and information accrues from case reports. Clinical findings and management mirror those for abdominal pregnancy. Although laparotomy is required in most instances, a few case reports describe laparoscopic excision of early small pregnancies (Apantaku, 2006; Cormio, 2006).

Ectopic placental implantations in less expected sites have been described in case reports and include the omentum, liver, and retroperitoneum, among others (Brouard, 2015; Liang, 2014; Watrowski, 2015). Also, intramural uterine implantations at sites other than a cesarean scar have been noted in women with prior uterine surgeries, ART, or adenomyosis (Memtsa, 2013; Wu, 2013). Although laparotomy is preferred by many for these ectopic sites, laparoscopic excision by those with suitable skills is gaining acceptance.