Diagnosis and treatment of ectopic pregnancy

Historical features and physical findings

Ectopic pregnancy is usually diagnosed in the first trimester of pregnancy. The most common gestational age at diagnosis is 6 to 10 weeks, but fetal viability can be discovered until the time of delivery.15^,16 Ectopic pregnancy has about the same frequency across a wide range of maternal ages and ethnic origins. Documentation of risk factors (Table 19^,17^,18) is an essential part of history-taking, and asymptomatic clinic patients with risk factors may benefit from routine early imaging.19 However, more than half of identified ectopic pregnancies are in women without known risk factors.8^,9

The physical findings depend on whether tubal rupture has occurred. Women with intraperitoneal hemorrhage present with significant abdominal pain and tenderness, along with various degrees of hemodynamic instability. However, women without rupture may also present with pelvic pain or vaginal bleeding, or both.8^,9^,20^,21 Several investigators have measured the predictive value of specific risk factors and physical findings alone or in combination: no combination correctly and consistently ruled out ectopic pregnancy.8^,9^,21 Given the high prevalence of ectopic pregnancy among pregnant women presenting to emergency departments, further investigation is prudent for all patients presenting with first-trimester bleeding or pain.

Use of β human chorionic gonadotropin measurement

It is important to confirm pregnancy. In the emergency department, pregnancy is diagnosed by determining the urine or serum concentration of β human chorionic gonadotropin (β-hCG). This hormone is detectable in urine and blood as early as 1 week before an expected menstrual period. Serum testing detects levels as low as 5 IU/L, whereas urine testing detects levels as low as 20–50 IU/L.22 In most cases, screening is done with a urine test, since obtaining the results of a serum test is time-consuming and is not always possible in the evening and at night. However, if pregnancy is strongly suspected, even when the urine test has a negative result, serum testing will be definitive.

A single serum measurement of the β-hCG concentration, however, cannot identify the location of the gestational sac. Although women with an ectopic pregnancy tend to have lower β-hCG levels than those with an intrauterine pregnancy, there is considerable overlap (Table 2).23^,24

If a low serum β-hCG level (< 1000 IU/L) is associated with a higher relative risk of ectopic pregnancy, then can very low levels predict a benign clinical course? In general, no. Although a single very low serum level (< 100 IU/L) has been felt to be reassuring, in a review of 716 admitted patients with ectopic pregnancy, 29% of those with such a level were found to have tubal rupture at laparoscopy.25 The risk of tubal rupture was similar across a wide range of β-hCG values. Another study identified 38 instances of rupture among women with serum levels ranging from 10 to 189 720 IU/L.7 Thus, a single serum β-hCG measurement cannot exclude ectopic pregnancy or predict the risk of rupture unless it is less than 5 IU/L.

Serial β-hCG measurement is often used for women with first-trimester bleeding or pain, or both, but, as with a single measurement, serial measurement cannot confirm the location of the gestational sac. In a normal pregnancy, the first-trimester β-hCG concentration rapidly increases, doubling about every 2 days. An increase over 48 hours of at least 66% has been used as a cutoff point for viability.20^,26^,27 Ectopic pregnancy may present with rising, falling or plateau β-hCG levels; thus, serial measurement is most useful to confirm fetal viability rather than to identify ectopic pregnancy. In a patient with a subnormal increase in β-hCG concentration, nonviability is assumed, and more invasive investigations can be used to clarify the nature of the abnormality (i.e., miscarriage v. ectopic pregnancy). However, over-reliance on the doubling time may result in the interruption of a normal pregnancy through diagnostic dilatation and curettage (D&C) or administration of methotrexate. A recent study identified patients with only a 53% increase in serum β-hCG levels over 2 days who had a viable intrauterine pregnancy.28 Thus, demonstration of normal doubling of serum levels over 48 hours supports a diagnosis of fetal viability but does not rule out ectopic pregnancy, and a rising β-hCG concentration that fails to reach 50% suggests a failing or ectopic pregnancy, as does a plateau. Falling levels confirm nonviability but do not rule out ectopic pregnancy.

Use of progesterone measurement

Measurement of the serum concentration of progesterone has been investigated as a potentially useful adjunct to serum β-hCG measurement, since progesterone levels are stable and independent of gestational age in the first trimester.14 A meta-analysis, published in 1998, of studies assessing a single progesterone level demonstrated good capacity of low levels (≤ 5 ng/mL) to correctly diagnose pregnancy failure, but this cutoff was unable to discriminate between ectopic pregnancy and intrauterine pregnancy.29 Both high (> 22 ng/mL) and low (≤ 5 ng/mL) cutoff points have since been studied for their ability to correctly identify nonviable pregnancy and ectopic pregnancy (Table 3).30^,31 Rapid progesterone analysis can identify 2 important subgroups of patients in the emergency department with symptomatic first-trimester bleeding or pain, or both: stable patients with progesterone levels above 22 ng/mL, who have a high (but not certain) likelihood of viable intrauterine pregnancy; and patients with levels of 5 ng/mL or less, who almost certainly have a nonviable pregnancy. Invasive diagnostic testing (e.g., D&C) could be postponed in the former patients but offered to the latter, as could treatment with methotrexate, without fear of interrupting a potentially viable intrauterine pregnancy.

Ultrasound imaging

Transvaginal ultrasonography has transformed the assessment of women with problematic early pregnancy, allowing earlier, clearer visualization of both normally developing embryos and abnormalities. A normal gestational sac, an ovoid collection of fluid adjacent to the endometrial stripe, can be visualized by means of the transvaginal probe at a gestational age of about 5 weeks. It can often be seen when 2 or 3 mm in diameter and should be consistently seen at 5 mm. Since the hormonal environment in ectopic pregnancy can produce an intrauterine fluid collection that mimics a gestational sac (the “pseudogestational sac” shown in Fig. 1, arrow), a sac alone cannot confirm intrauterine pregnancy.32

As the embryo matures, more sonographic signs become visible. Once the sac is implanted within the endometrium, its position relative to the endometrial wall changes, producing the intradecidual-sac sign and then the double decidual-sac sign. The earliest embryonic landmark, the yolk sac, appears when the sac is 8 mm or more in diameter, usually during the fifth week of gestation (Fig. 2). Cardiac activity can be seen with endovaginal scanning when the embryo reaches 4 to 5 mm in diameter, at a gestational age of 6–6.5 weeks.32

The concept of the discriminatory threshold (the β-hCG level at which an intrauterine gestational sac can be reliably seen in a normal pregnancy) has existed since the early 1980s.33 A β-hCG level that has risen above the discriminatory threshold in the absence of sonographic signs of early pregnancy is considered presumptive evidence of an ectopic pregnancy. With the evolution in ultrasound technology, the discriminatory threshold has dropped from 6500 IU/L with a transabdominal approach to between 1000 and 2000 IU/L with transvaginal imaging.34 This threshold is user- and machine-dependent and thus will vary slightly from institution to institution. Caution should be used in assuming an ectopic pregnancy when a nondiagnostic ultrasound image accompanies a single β-hCG level above the discriminatory threshold: several articles have reported a small number of patients with indeterminate ultrasound images and β-hCG levels above the threshold who have been eventually found to have a viable intrauterine pregnancy.34^,35^,36 In addition, there could be unseen multiple intrauterine gestational sacs, since the β-hCG values relative to gestational age are higher in patients with multiple embryos.37

Ultrasonographic identification of an intrauterine pregnancy (gestational sac plus yolk sac or other embryonic sign) rules out ectopic pregnancy in most patients.32 The exception is in patients with ovulation induction and assisted conception, who are at risk of heterotopic pregnancy (dizygotic twins, 1 intrauterine and 1 extrauterine). Although this phenomenon is exceedingly rare in the general population (estimated frequency 1 per 3889 to 30 000 pregnancies),38 in the setting of assisted reproduction it may occur in 1 in 100 pregnancies.39

The spectrum of sonographic findings in ectopic pregnancy is broad. Identification of an extrauterine gestational sac containing a yolk sac (with or without an embryo) confirms the diagnosis. Suggestive findings include an empty uterus, cystic or solid adnexal or tubal masses (including the tubal-ring sign, representing a tubal gestational sac), hematosalpinx and echogenic or sonolucent cul-de-sac fluid (Fig. 3).

Many prospective studies have shown that “formal” transvaginal ultrasound imaging (i.e., that performed by ultrasound technicians and interpreted by radiologists) in the emergency department has high accuracy in confirming intrauterine and ectopic pregnancy. Most protocols can establish a diagnosis with the initial scan in more than 75% of emergency department patients.23^,35^,40 A diagnosis can often be established even in the subgroup of patients with β-hCG levels below the discriminatory threshold. In some studies, transvaginal scanning has identified up to one-third of the patients with below-threshold β-hCG levels who had ectopic pregnancy.35^,41 Given the likelihood of a definitive diagnosis, even with below-threshold β-hCG levels, ultrasonography is the best initial investigation in problematic early pregnancy.

Because expertise in transvaginal ultrasonography is not available in all hospitals and may not be quickly available in some larger centres, there have been several studies of ultrasonography performed by emergency physicians in the assessment of patients with first-trimester bleeding or pain. Ultrasonography in the emergency department (ED-based ultrasonography) has evolved over the last decade and is now part of the diagnostic work-up for many clinical problems in major Canadian centres, as well as in a large number of smaller community emergency departments (Dr. Ray Wiss, Emergency Department Echo course director: personal communication, 2005). The evaluation involves 2 “Yes/No” questions: Can an intrauterine pregnancy be identified? Is there free pelvic or intra-abdominal fluid? This approach is in contrast to the goal of a “formal” pelvic ultrasound study, which is to describe the anatomic appearance and visible abnormalities of the uterus, adnexa and cul-de-sac. Absence of an intrauterine pregnancy translates to a risk of ectopic pregnancy of about 36%,42 and free fluid in the cul-de-sac represents an even higher risk. Several studies have documented the ability of emergency physicians to quickly and accurately identify both intrauterine pregnancy and intra-abdominal free fluid by means of ED-based ultrasonography after brief standardized training.12^,42^,43 The addition of ED-based ultrasonography to structured protocols for assessing symptomatic patients in the 1st trimester of pregnancy has led to a dramatically decreased stay in the emergency department44 as well as a decrease in the incidence of complications associated with missed ectopic pregnancy and tubal rupture.12

Transvaginal ultrasonography should therefore be the initial investigation for pregnant patients presenting to the emergency department with first-trimester bleeding or pain. Not only is it highly accurate in identifying ectopic pregnancy, but also it offers patients what they are most expecting from their visit: information about the health and viability of their pregnancy. No combination of history-taking, physical examination and laboratory tests can make the same claim. The use of ED-based ultrasonography offers rapid bedside detection of a viable intrauterine pregnancy or a high risk of ectopic pregnancy. Emergency physicians without access to bedside ED-based ultrasonography should arrange formal ultrasonography for all patients with early-pregnancy complaints. This investigation can be performed during the initial visit or, if the patient is stable and has minimal symptoms, the next day in an outpatient visit. However, in the case of outpatient investigation, mechanisms for timely follow-up, re-examination and further investigation must be in place.

Fig. 4 outlines the recommended approach to imaging and follow-up.

Expectant management

Ectopic pregnancy can resolve spontaneously through regression or tubal abortion. However, about 90% of women with ectopic pregnancy and serum β-hCG levels greater than 2000 IU/L require operative intervention owing to increasing symptoms or tubal rupture.3^,4 Tubal rupture can also occur when serum β-hCG levels are low or declining, or both.45 Expectant management should be offered only when transvaginal ultrasonography fails to show the location of the gestational sac and the serum levels of β-hCG and progesterone are low and declining. Because of the possibility of tubal rupture, these patients must be carefully monitored until the serum β-hCG concentration falls below 15 IU/L; at this point almost all ectopic pregnancies resolve spontaneously, without rupture.

Surgical management

Surgical management of ectopic pregnancy should be reserved for patients who refuse or have contraindications to medical treatment, those in whom medical treatment has failed and those who are hemodynamically unstable.

Three randomized studies have demonstrated that, compared with laparotomy, laparoscopic treatment of ectopic pregnancy is associated with lower cost, shorter hospital stay, less operative time, less blood loss, less analgesic requirement and faster recovery.46^,47^,48 Patients randomly assigned to laparoscopy also had fewer adhesions than patients treated with laparotomy (19% v. 64%).49

Tube-sparing salpingostomy (in which the gestational sac is removed, without the tube, through a 1-cm-long incision on the tubal wall) is preferred to salpingectomy (removal of the tube), as the former is less invasive but has comparable rates of subsequent fertility and ectopic pregnancy.50^,51^,52 However, 8% of patients have persistent ectopic pregnancy after salpingostomy.50 Follow-up determinations are required until β-hCG is undetectable. Regardless of the type of surgery, contralateral tubal abnormalities predispose the patient to recurrent ectopic pregnancy. In a retrospective study of 276 women with ectopic pregnancy, the cumulative rates of spontaneous intrauterine pregnancy over 7 year were 89% after conservative surgery and 66% after radical surgery.50 There was no significant difference in the risk of repeat ectopic pregnancy (17% after conservative surgery and 16% after radical surgery).

In summary, salpingostomy is preferred, particularly for women who wish to have another pregnancy. Salpingectomy may be necessary for women with uncontrolled bleeding, recurrent ectopic pregnancy in the same tube, a severely damaged tube or a tubal gestational sac greater than 5 cm in diameter.53

Surgical versus medical treatment

Several randomized studies found that MTX treatment in selected patients with ectopic pregnancy was as effective as laparoscopic treatment (Table 4).56^,57^,58^,59 The 2 treatments were also equally effective in tubal preservation; however, the β-hCG concentration declined more quickly after laparoscopic surgery.56 After MTX treatment the health-related quality of life may diminish, possibly owing to both long-term persistence of the ectopic pregnancy and the long treatment course. There were more physical symptoms after 2 days and 2 weeks in those given MTX, although symptoms were increased in both treatment groups. However, because of the noninvasive nature of MTX treatment, most patients are willing to cope with this short-term burden.60 MTX treatment is less expensive than laparoscopic surgery,59^,60 although in one study this was true only if the initial serum β-hCG level was less than 1500 IU/L.60