Unmedicated, medicated and copper-coated IUCDs prevent both intrauterine and extrauterine pregnancies. However, a woman who conceives with an IUCD in situ is seven times more likely to have a tubal pregnancy compared with conception without contraception (Vessey et al 1974). IUCDs are more effective in preventing intrauterine than extrauterine implantation. With copper IUCDs, 4% of all accidental pregnancies are tubal, whereas with progesterone-coated IUCDs, 17% of all contraceptive failures are tubal pregnancies. The different mechanisms of action of the two devices could partially explain the difference in failure rates. Although both devices prevent implantation, copper IUCDs also interfere with fertilization by inducing cytotoxic and phagocytotic effects on the sperm and oocytes. Progesterone-containing IUCDs are probably less effective in preventing fertilization. Although the incidence of pregnancy diminishes with long-term use of the IUCD, among women who become pregnant, the likelihood of ectopic pregnancy increases. Women who have used the IUCD for more than 24 months are 2.6 times more likely to have an ectopic pregnancy compared with short-term users (<24 months). The ‘lasting effect’ of the IUCD may be related to the loss of the cilia from the tubal epithelium, especially if the IUCD has been in situ for 3 years or more (Wollen et al 1984, Ory HW 1981).

Termination of pregnancy

Data from two French case–control studies suggest that induced abortion may be a risk factor for ectopic pregnancy for women with no history of ectopic pregnancy. There is an association between the number of previous induced abortions and ectopic pregnancy [odds ratio (OR) 1.4 for one previous induced abortion and 1.9 for two or more] (Tharaux-Deneux et al 1998). Whether this is related to the spread of asymptomatic C. trachomatis infection or to the procedure itself is uncertain.

Assisted conception

Induction of ovulation with either clomiphene citrate or human menopausal gonadotrophin is a predisposing factor to tubal implantation (McBain et al 1980; Marchbanks et al 1985). A number of studies indicate that 1–4% of pregnancies achieved following induction of ovulation are ectopic pregnancies. The majority of these patients had a normal pelvis and patent tubes. The incidence of tubal pregnancy following oocyte retrieval and embryo transfer is approximately 4.5%. It must be noted that some women who undergo an in-vitro fertilization (IVF) cycle have risk factors for an ectopic pregnancy (i.e. previous ectopic pregnancy, tubal pathology or surgery).

Salpingitis isthmica nodosa

Salpingitis isthmica nodosa (SIN) is diagnosed by the histological evidence of tubal isthmic diverticula, and may be suggested by characteristic changes on hysterosalpingogram. Its incidence in healthy women ranges from 0.6% to 11%, but it is significantly more common in the setting of ectopic pregnancy. Persaud (1970) reported that 49% of fallopian tubes excised for tubal pregnancy had diverticula and evidence of SIN. The reason for the high incidence of ectopic gestation in women with SIN remains largely unknown. Defective myoelectrical activity has been demonstrated over the diverticula. Entrapment of the embryo into the diverticula is a possible mechanical explanation.

Smoking

A French study found that the risk of ectopic pregnancy is significantly higher in women who smoke. The risk increases according to the number of cigarettes per day (Bouyer et al 1998). The relative risk for ectopic pregnancy is 1.3 for women who smoke one to nine cigarettes per day, 2 for women who smoke 10–12 cigarettes per day, and 2.5 for women who smoke more than 20 cigarettes per day. Inhibition of oocyte cumulus complex pick-up by the fimbrial end of the fallopian tube and a reduction of ciliary beat frequency are associated with nicotine intake (Knoll and Talbot 1998).

Diethylstilboestrol

Results of a collaborative study indicate that the risk of ectopic pregnancy in diethylstilboestrol (DES)-exposed women was 13% compared with 4% for women who had a normal uterus (Barnes et al 1980). A meta-analysis on risk factors for ectopic pregnancy confirms that exposure to DES in utero significantly increases the risk of ectopic pregnancy (Ankum et al 1996).

Pathology

Sites of ectopic pregnancy

A 21-year survey of 654 ectopic pregnancies (Breen 1970) revealed that the most common sites of ectopic pregnancy are as shown in Table 25.2. Similar distribution for tubal and abdominal pregnancy sites have been reported by Bouyer et al (2002) [ampullary (70.0%), isthmic (12.0%), fimbrial (11.1%), interstitial (2.4%), abdominal (1.3%)]. In this population, no cervical pregnancies were observed and slightly increased incidence was seen for ovarian pregnancies (3.2%) (Bouyer et al 2002).

Table 25.2 Sites of ectopic gestation

Fallopian tube
Ampullary segment	80%
Isthmic segment	12%
Fimbrial end	5%
Interstitial and cornual	2%
Abdominal	1.4%
Ovarian	0.2%
Cervical	0.2%

Natural progression of a tubal pregnancy

Unruptured tubal pregnancy

Occasionally, in the very early stages of a tubal pregnancy, there are no obvious macroscopic features, and an ectopic pregnancy could be overlooked even after a laparoscopy. However, as pregnancy progresses, local enlargement of the tube occurs at the point of implantation. At a later stage, a large segment of the tube is distended and the tubal wall appears discoloured, dark red or purple.

Tubal rupture

One of the fundamental aspects of ectopic pregnancy is the inability of the tissues into which the blastocyst implants to offer resistance or respond to the invading trophoblast. Uncontrolled invasion of the trophoblast results in destruction of vessels, local haemorrhage and thinning of the tubal wall. Rupture of the tubal wall results in the escape of large amounts of blood, with or without the products of conception, into the peritoneal cavity. The embryo rarely survives. In rare cases, pregnancy continues if an adequate portion of the placenta is retained or if secondary implantation occurs in other organs of the pelvic or peritoneal cavity. Rupture occurs more often at the antimesenteric part of the tubal wall. Rupture of the inferior and mesenteric part of the tubal wall results in haemorrhage between the two layers of the broad ligament. Intraligamentous haemorrhage could cause rupture of anterior or posterior layers of the broad ligament.

Spontaneous involution

Usually, the conceptus dies at an early stage without any notable symptoms.

Complete tubal miscarriage

At an early stage, the conceptus is extruded via the fimbriated end of the fallopian tube into the peritoneal cavity, and subsequently absorbed by the surrounding tissues.

Incomplete tubal miscarriage

The conceptus is partially extruded via the fimbrial end of the tube, and intervention is usually required.

Tubal blood mole or carneous mole

In some cases, recurrent choriodecidual haemorrhage around the dead conceptus contributes to the formation of a tubal blood mole or carneous mole. The presence and development of cellular and biochemical elements of connective tissue around the mole give rise to a semi-solid structure which could remain unresolved for years.

Time of rupture at various sites in the tube

Isthmic implantation

The isthmic segment of the fallopian tube is narrow and less distensible than the ampullary or interstitial segments. Isthmic rupture often occurs at 6–8 weeks of gestation and is usually dramatic.

Ampullary implantation

Rupture of an ampullary pregnancy usually occurs at 8–12 weeks of gestation. The ampulla is the wider segment of the fallopian tube and the site of 80% of all ectopic pregnancies.

Interstitial implantation

The interstitial segment of the fallopian tube is surrounded by myometrium which can hypertrophy to accommodate the enlarging conceptus. Rupture therefore occurs at a relatively late stage of 12–14 weeks of gestation. Rupture of an interstitial pregnancy causes damage of the highly vascularized cornual end of the uterus, resulting in severe intra-abdominal haemorrhage.

Histological changes

In the early stages of pregnancy, the myometrium responds in an identical pattern under the influence of hormones, irrespective of whether the gestation is ectopic or eutopic. The uterus becomes softened and slightly enlarged as a consequence of hypertrophy and hyperplasia of the myometrial cells. Where the gestation is ectopic, the endometrial glands demonstrate an atypical histological pattern referred to as the ‘Arias-Stella phenomenon’, which is characterized histologically by: hyperplasia of glandular cells, closely packed glands with evidence of hypersecretion, large irregular hyperchromatic nuclei, cytoplasmic vacuolation and loss of cellular polarity. It is important to note that the Arias-Stella reaction is non-specific and can be found in the endometrium of patients with an intrauterine pregnancy. However, the presence of the Arias-Stella reaction and the absence of chorionic villi from endometrial curettings are both highly suspicious signs of an extrauterine pregnancy. The presence of chorionic villi is the most reliable histological feature for the definite diagnosis of pregnancy.

Simultaneous with the glandular changes, the stroma is converted into decidual tissue containing large polyhedral cells with hyperchromatic nuclei. A failing pregnancy and lower levels of associated hormones results in gradual disintegration of the decidua, giving rise to the intermittent, occasionally heavy, vaginal bleeding of uterine origin that occurs in ectopic pregnancy. In some cases, the decidua may be detached abruptly and passed as a flat, triangular, reddish-brown piece of tissue called a ‘decidual cast’.

Diagnosis

Symptoms and signs

Ectopic pregnancy remains a diagnostic challenge. It should be considered as an important differential in any woman of reproductive age who presents with the triad of amenorrhoea, abdominal pain and irregular vaginal bleeding. This philosophy is particularly useful if the patient has any of the risk factor(s) identified in Table 25.1. The frequency with which various symptoms and signs were reported from a series of 300 consecutive cases are shown in Table 25.3 (Droegemueller 1982).

Table 25.3 Symptoms and signs in 300 consecutive cases of ectopic pregnancy at admission

Symptoms and signs	Cases (%)
Abdominal pain	99
Generalized	44
Unilateral	33
Radiating to the shoulder	22
Abnormal uterine bleeding	74
Amenorrhoea ≤2 weeks	68
Syncopal symptoms	37
Adenexal tenderness	96
Unilateral adenexal mass	54
Uterus
Normal size	71
6–8-week size	26
9–12-week size	3
Uterine cast passed away vaginally	7
Admission temperature >37°C	2

Abdominal pain

This is the most common symptom; however, it should be emphasized that there is no typical pain that is pathognomonic of ectopic pregnancy. Women can present with generalized abdominal or localized pain in the pelvis (unilateral or bilateral) and/or pain radiating to the shoulder.

The generalized abdominal pain is usually due to rupture of ectopic pregnancy and intraperitoneal haemorrhage. The pain is often severe. Shoulder pain is also an indirect indication of intraperitoneal haemorrhage. Accumulation of blood in the subdiaphragmatic region stimulates the phrenic nerve and creates shoulder tip pain. Localized pain may be due to distension of the fallopian tube. The pain may be sudden or progressive, and continuous or intermittent.

Amenorrhoea and abnormal uterine bleeding

Most patients present with amenorrhoea of at least 2 weeks duration. One-third of the women will either not recall the date of their last menstrual period or have irregular periods. Abnormal uterine bleeding occurs in 75% of women with an ectopic pregnancy. The bleeding is often light, recurrent and results from detachment of the uterine decidua. According to Stabile (1996a), ‘If a patient who is a few weeks pregnant complains of a little pain and heavy vaginal bleeding, the pregnancy is probably intrauterine, whereas if there is more pain and little bleeding, it is more likely to be an ectopic pregnancy’.

Other symptoms

Although abdominal pain, amenorrhoea and abnormal vaginal bleeding are the most common and typical symptoms, patients may present with additional features such as syncopal attacks. These are related to sudden-onset haemorrhage or to hypovolaemia or anaemia. Other atypical symptoms include diarrhoea or vomiting. In the 2003–2005 Confidential Enquiry into Maternal and Child Health report (Lewis, 2007), some women who presented with these symptoms were undiagnosed and subsequently died.

Physical examination

Physical examination should include an assessment of vital signs and examination of the abdomen and pelvis. Depending on the rate and amount of blood loss, the general condition of the patient may vary from slight pallor to haemodynamic shock. Palpation of the abdomen may reveal generalized or localized mild tenderness. Occasionally, guarding and rebound tenderness are also elicited.

An unusual feature is the Cullen’s sign. This is bluish discoloration of the skin around the umbilicus caused by a considerable quantity of free blood in the peritoneal cavity. However, this sign is rare and its absence does not exclude massive intraperitoneal haemorrhage.

The findings on pelvic examination vary from completely negative examination to the presence of a large, fixed, soft and tender mass. An adnexal mass may be palpable in up to 55% of cases. In many cases, the mass is ill defined and it may consist not only of tubal pregnancy but also of adherent omentum, small and large bowel. The mass may also be an enlarged corpus luteum. The uterus may be slightly enlarged but its size does not normally correspond to the gestational age. Cervical motion tenderness may or may not be present. A tender boggy mass in the pouch of Douglas, when present, represents either a collection of blood or a dilated tube adherent to the posterior uterine wall.

Types of presentation

The presentation of symptomatic patients with a tubal ectopic pregnancy may be acute or subacute.

Acute presentation

This is usually a consequence of rupture of the ectopic gestation and the ensuing intraperitoneal haemorrhage and haemodynamic shock. These symptoms are due to the intra-abdominal haemorrhage and collection of blood into the subdiaphragmatic region and pouch of Douglas. The patient is often pale, hypotensive and tachycardic. She may complain of shoulder tip pain or urge to defaecate. Abdominal examination reveals generalized and rebound tenderness. Vaginal examination will reveal tenderness in the adnexal region and cervical motion tenderness. However, vaginal examination in patients who present with an acute abdomen due to a ruptured ectopic pregnancy is generally considered unnecessary and potentially dangerous for the following reasons: (a) generalized haemoperitoneum and pain often mean that specific information cannot be elicited, (b) the patient is very uncomfortable and the assessment is often difficult and inadequate, and (c) it could result in total rupture of the ectopic pregnancy and delay management of the patient. Although it is reported that 30% of all women with an ectopic pregnancy present after rupture (Barnhart et al 1994), acute presentation is becoming less common. This is due primarily to increased patient awareness and early referral to the hospital for evaluation of ‘suspected ectopic’ pregnancies. Finally, the availability of more sensitive and rapid biochemical tests for beta-human chorionic gonadotrophin (β-hCG) quantification and the wider availability of transvaginal ultrasonography and laparoscopy have significantly reduced the interval between presentation and treatment.

Subacute presentation

When the process of tubal rupture or abortion is very gradual, the presentation of ectopic pregnancy is subacute. According to Stabile (1996a,b), this is the group of women who are symptomatic but clinically stable. A history of a missed period and recurrent episodes of light vaginal bleeding may exist. The circulatory system adjusts the blood pressure, and the patient is haemodynamically stable. Progressively increasing lower abdominal pain and, occasionally, shoulder pain are typical symptoms. On bimanual examination, there may be localized tenderness in one of the fornices, and cervical motion tenderness is often present. Subacute presentation occurs in 80–90% of ectopic pregnancies. In cases with such a presentation, the establishment of an accurate diagnosis becomes more difficult, hence the need for further investigations.

Further Investigations

Biochemical tests

Quantitation of β-hCG subunit

hCG is a glycoprotein with a mass of 36,700 Da that is secreted by the syncytiotrophoblasts. It is heterodimeric, composed of two non-covalently linked α and β subunits. The α subunit is identical to that of follicle-stimulating hormone and luteinizing hormone, whereas the β subunit is specific to hCG. In response to β-hCG, the corpus luteum produces increasing concentrations of progesterone and other steroid hormones required to maintain early pregnancy, until the placenta takes over by the eighth week of gestation. β-hCG first enters the maternal circulation on the day of blastocyst implantation, i.e. 6–7 days after conception. Serum concentrations of β-hCG are approximately 1000 IU/l (third international standard) at approximately 4 weeks of pregnancy, increase exponentially during the first 6 weeks of gestation, and reach 20,000–250,000 IU/l at 10 weeks of gestation. Concentrations then decrease to 10,000–20,000 IU/l by 20 weeks of gestation and plateau thereafter.

Using highly sensitive assays (detection limit 0.1–0.3 IU/l), β-hCG can be detected in the maternal circulation around the time of implantation. This will appear in urine approximately 2 days after it appears in the blood. Most commercially available monoclonal-antibody-based urine pregnancy tests can detect β-hCG concentrations above 25 IU/l, which corresponds to day 24–25 of a regular 28-day cycle. The sensitivity of modern assays is 99–100%.

Single β-hCG measurement

A single serum β-hCG concentration has been used as a discriminatory level to detect an ectopic pregnancy as described below. A single measurement, however, has limited clinical value as there is considerable overlap of values for normal and abnormal pregnancies. When using radioimmunometric assays with a detection limit of 5 IU/l, a single measurement of β-hCG may be useful because, if negative, it can exclude the diagnosis of an ectopic pregnancy.

Serial β-hCG measurements

Serum β-hCG concentrations double every 1.4–1.6 days from the time of first detection up to the 35th day of pregnancy, and then double every 2.0–2.7 days from the 35th to the 42nd day of pregnancy (Pittaway et al 1985). Since the normal doubling time of β-hCG is 2.2 days and its half-life is 32–37 h, serial quantitative assessments of β-hCG may help to distinguish normal from abnormal pregnancies. Kadar et al (1981) first reported a method for screening for ectopic pregnancy based on β-hCG doubling time. An increase in serum β-hCG of less than 66% over 48 h was suggestive of an ectopic pregnancy (using an 85% CI for β-hCG levels). More recently, studies have used an increase in serum β-hCG levels of 35–53% (using 99% CI) to diagnose viable intrauterine pregnancies and to reduce the potential risk of terminating an intrauterine pregnancy (Seeber et al 2006). However, another method used to help with the diagnosis of an ectopic pregnancy is the ‘plateau’ in serum β-hCG levels. Plateau is defined as a β-hCG doubling time of 7 days or more (Kadar and Romero 1988). If the half-life of serum β-hCG is less than 1.4 days, spontaneous miscarriage is likely, whereas a half-life of more than 7 days is more likely to be indicative of an ectopic pregnancy. Therefore, falling levels of β-hCG can distinguish between an ectopic pregnancy and a spontaneous miscarriage. As a proportion of ectopic pregnancies are tubal miscarriages (with biochemical changes similar to those of intrauterine miscarriages), and approximately 15–20% of all ectopic pregnancies can have doubling serum β-hCG levels similar to those of normal intrauterine pregnancies (Silva et al 2006), suboptimal serial β-hCG changes are not specific or sensitive enough to diagnose ectopic pregnancies.

Serum progesterone

Progesterone, a C-21 hormone, is essential in early pregnancy for decidual function, implantation, reduction in immune response and myometrial quiescence. It is produced entirely by the corpus luteum until the sixth week of gestation. The trophoblastic contribution begins from the seventh week and becomes the predominant source by the 12th week. Progesterone concentrations continue to rise until 7 weeks of gestation and then plateau until 10 weeks, after which there is a gradual increase until term. In the conception cycle, progesterone concentrations are 2–4 nmol/l on the day of the luteinizing hormone surge and rise to 20–70 nmol/l a week later. At term, concentrations vary from 200 to 600 nmol/l.

Progesterone concentrations have been widely used for the diagnosis and management of pregnancies of unknown location (PUL, i.e where ultrasound is inconclusive). In failing pregnancies, whether ectopic or miscarriage, progesterone concentrations are expected to be low compared with values in healthy ongoing pregnancies (Hahlin et al 1990). Most studies report cut-off concentrations of less than 16 nmol/l for failing pregnancies and more than 80 nmol/l for healthy ongoing pregnancies (Mathews et al 1986, Sau and Hamilton-Fairley 2003, Bishry and Ganta 2008). Progesterone levels over 25 nmol/l are ‘likely to indicate’ and levels over 60 nmol/l are ‘strongly associated with’ pregnancies subsequently shown to be normal. A progesterone concentration below 25 nmol/l in an anembryonic pregnancy has been shown to be diagnostic of non-viability (Elson et al 2003). Concentrations less than 20 nmol/l have a sensitivity of 93% and a specificity of 94% for the prediction of spontaneous resolution of PULs (Banerjee et al 2001). A meta-analysis has demonstrated that a single serum progesterone measurement is good at predicting a viable intrauterine or failed pregnancy, but is not useful for locating the site of pregnancy (Mol et al 1998). When interpreting progesterone measurements, variations in concentrations should be taken into account because of the assay methods, and a departmental protocol should state the normal range for that unit.

Other protein and steroid markers

In an effort to detect an ectopic pregnancy at an early stage, various steroid and protein markers have been studied including Schwangerschafts protein-1, human placental lactogen, pregnancy-associated plasma protein A, inhibin and insulin growth factor binding protein-1. However, none of these markers has been found to make a significant impact on clinical practice. Recently, serum CA125, creatine kinase, activin A and vascular endothelial growth factor have also been studied for the diagnosis of ectopic pregnancy; however, their sensitivity and specificity are not sufficient to allow their application in clinical practice.

Ultrasonography

Ultrasound has become an essential tool in the assessment and diagnosis of suspected ectopic pregnancies. The presence of an intrauterine pregnancy does not conclusively exclude an ectopic pregnancy; however, the occurrence of a heterotopic pregnancy is rare. The transvaginal approach is far superior to the abdominal approach, as the proximity of the vaginal probe to the pelvic structures and the use of high-frequency transducers (5–7 MHz) significantly improves resolution.

The earliest a normal intrauterine gestational sac (approximately 2 mm) can be seen is 4 weeks with a transvaginal ultrasound and 5 weeks with an abdominal ultrasound. In 20% of patients with an ectopic pregnancy, collection of fluid within the uterine cavity results in the formation of a pseudo gestational sac which mimics a true intrauterine sac. Ultrasonographic signs of a true gestational sac include an eccentrically placed gestational sac within the uterine cavity, the double ring or double decidual sac sign (DDSS) and an intact midline endometrial echo, with the pregnancy seen implanted beneath the endometrial surface. The double ring comprises chorionic villi, intervillous lakes, the extravillous trophoblast and the maternal decidua. This DDSS can also be seen in one-third of ectopic pregnancies. On the other hand, a pseudo sac is surrounded by a single layer of tissue, tends to follow the contour of the cavity, and in a longitudinal section of the uterus, the midline endometrial echo cannot be seen because of the fluid.

The most reliable sign to differentiate between a pseudo sac and a true sac is the appearance of a yolk sac within the true gestational sac. The yolk sac confirms an intrauterine pregnancy even before a live embryo is detected. A yolk sac is identified by ultrasound when the mean gestational sac diameter is greater than 8 mm.

A review of the literature by Brown and Doubilet (1994) revealed the frequency of various ultrasound features in ectopic pregnancies as detailed below:

• an empty uterus (28%);

• an empty uterus and an adnexal mass (35%);

• an intrauterine sac or pseudo gestational sac (25%);

• an empty uterus and an ectopic gestational sac (12%) with or without a yolk sac and cardiac activity; and

• varying amount of fluid in the pouch of Douglas (25%).

In ectopic pregnancies, a hyperechogenic tubal ring (’doughnut’ or ‘bagel’ sign) is the most common finding on ultrasound scan. Others include a mixed adnexal mass representing either tubal miscarriage or tubal rupture, an ectopic sac with a yolk sac, or an embryo with or without a fetal heartbeat. The corpus luteum may be present on the ipsilateral side in 85% of cases. It has been reported that 74% of ectopic pregnancies can be visualized on initial transvaginal ultrasound scan, and more than 90% of ectopic pregnancies can be visualized prior to treatment (Kirk et al 2007).

Quantitative assessment of β-hCG levels is essential for the accurate interpretation of ultrasonographic findings. A single serum β-hCG measurement has been used as a discriminatory level to help with the diagnosis of ectopic pregnancy. Absence of an intrauterine gestational sac with serum β-hCG levels of 6500 IU/l or more has an 86% positive predictive value and a 100% negative predictive value for the presence of an ectopic pregnancy (Romero et al 1985). With transvaginal ultrasound scan, different discriminatory levels of serum β-hCG (e.g. 1000, 1500 and 2000 IU/l) have been used with no significant difference in the detection rates of ectopic pregnancies (Barnhart et al 1999, Condous et al 2005). Failure to visualize an intrauterine gestational sac by transvaginal ultrasound if β-hCG concentrations are 1000–2000 IU/l or more indicates either an abnormal intrauterine pregnancy, a recent miscarriage or an ectopic pregnancy. It must be emphasized that although the discriminatory zone for an intrauterine pregnancy is well established, there is no such discriminatory zone for ectopic pregnancies. In 15–20% of women with clinical suspicion of an early pregnancy failure, ultrasound findings are not diagnostic.

Doppler ultrasonography

Doppler ultrasound is a non-invasive biophysical method of investigating patterns of blood flow. In ectopic pregnancies, the normal high-resistance blood flow in the uterine or ovarian artery branches changes to low resistance.

Jurkovic et al (1992), in a prospective study, used colour Doppler to detect and compare changes of blood flow in the uterine and spiral arteries and the corpus luteum in ectopic and intrauterine pregnancies. In intrauterine pregnancies, the impedance to flow (resistance index) in uterine arteries decreased with gestational age, but this remained constant in ectopic pregnancies. Peak blood velocity in the uterine arteries increased with gestational age in intrauterine pregnancies, and the values were significantly higher than those seen in ectopic pregnancies. However, uterine blood flow velocity was lower in the ectopic group, indicating an overall reduction in uterine blood supply. Local vascular changes associated with a true gestational sac differentiate an intrauterine pregnancy from the pseudo sac of an ectopic pregnancy. Doppler flow imaging may thus suggest that an adnexal mass is an ectopic pregnancy from the presence of high vascularity at the periphery of the adnexal mass.

Colour and pulsed Doppler increase the sensitivity of transvaginal ultrasound and allow earlier detection of ectopic pregnancies, thereby giving the opportunity for the application of medical treatment or conservative surgery.

Three-dimensional ultrasonography

Three-dimensional (3D) ultrasound is emerging as a possible additional diagnostic tool for ectopic pregnancy. Rempen (1998) conducted a prospective follow-up study in order to evaluate the potential of 3D ultrasound to differentiate intrauterine gestations from extrauterine gestations. Fifty-four pregnancies with a gestational age of less than 10 weeks and with an intrauterine gestational sac more than 5 mm in diameter were included in the study. The configuration of the endometrium in the frontal plane of the uterus was correlated with pregnancy outcomes. It was found that this configuration was asymmetrical in 84% of intrauterine pregnancies, whereas the endometrium showed asymmetry in 90% of extrauterine pregnancies (P ≤ 0.0001). The conclusion was that evaluation of the endometrial shape in the frontal plane may be a useful additional means of distinguishing intrauterine from extrauterine pregnancies.

Harika et al (1995) conducted a study to assess the role of 3D imaging in the early diagnosis of ectopic pregnancy on 12 asymptomatic patients whose gestational age was less than 6 weeks. Nine of these patients had an ectopic pregnancy at laparoscopy, and of these, 3D transvaginal ultrasound showed a small ectopic gestational sac in four cases. These preliminary results suggest that 3D ultrasound may be an effective procedure for the diagnosis of ectopic pregnancy in asymptomatic patients prior to the sixth gestational week; however, its accuracy would have to be refined.

Diagnostic laparoscopy

With the availability of sensitive quantitative serum β-hCG assays and high-resolution transvaginal ultrasound, diagnostic laparoscopy is seldom used nowadays as a primary diagnostic tool. In some units, it is used to diagnose ectopic pregnancy in the presence of an empty uterus with serum β-hCG levels above the discriminatory level. A major disadvantage of laparoscopy is that a small ectopic pregnancy may not be visualized, and in 3–4% of cases the diagnosis will be missed. Therefore, careful selection of patients is crucial.

Culdocentesis

With the availability of modern equipment and biochemical tests, culdocentesis is rarely used for the diagnosis of ectopic pregnancy. However, in parts of the world where these facilities are lacking, aspirating blood from the pouch of Douglas which does not clot on standing is most likely suggestive of an ectopic pregnancy.

Differential Diagnosis

The differential diagnoses of an ectopic pregnancy should include ruptured corpus luteum cyst, threatened or incomplete miscarriage, PID and degenerating fibroids. Quantitative assessment of β-hCG, transvaginal ultrasonography and laparoscopy are powerful tools which should be used to resolve uncertain cases.

Pregnancy of Unknown Location

In recent years, PUL has been used as a diagnosis for patients with a positive pregnancy test with no signs of intrauterine or extrauterine pregnancy or retained products of conception on an initial transvaginal ultrasound scan. Studies report that PUL is identified in 8–31% of women referred for ultrasound assessment in early pregnancy (Hahlin et al 1995, Condous et al 2004).

Subsequent follow-up in woman with PUL will lead to the diagnosis of either: (a) an intrauterine pregnancy that was too early to be visualized on an initial ultrasound scan; (b) an ectopic pregnancy that was too early to be visualized or was missed on an initial ultrasound scan (7–20%); (c) a spontaneously resolving pregnancy known as a ‘failing PUL’ (50–70%), which would include both intrauterine and extrauterine pregnancies; or (d) persistent PUL, where biochemical findings mimic an ectopic pregnancy with suboptimal serum β-hCG levels that fail to decline, and failure to localize pregnancy on ultrasound scan or laparoscopy.

PULs present a management dilemma for health professionals in the early pregnancy unit, and inevitably increase workload and time consumption. For clinically stable women, expectant management has been demonstrated to be safe. PULs are followed up with serial hormone measurements, repeat transvaginal ultrasound scans and possibly laparoscopy with or without uterine curettage. It is important to note that in women with a history and transvaginal ultrasound findings suggestive of complete miscarriage, serial β-hCG measurements should be performed as approximately 6% of these can be ectopic pregnancies (Condous et al 2004b).

Recently, the use of β-hCG ratios (β-hCG 48 h/β-hCG 0 h) rather than absolute β-hCG values has been assessed in the diagnosis and management of PUL. A β-hCG ratio below 0.87 (or a β-hCG decrease >13%) has a 92.7% sensitivity (95% CI 85.6–96.5) and a 96.7% specificity (95% CI 90.0–99.1) for the prediction of a failing pregnancy (Condous et al 2006, Bignardi et al 2008). Various mathematical models using log regression and the Bayesian approach have been developed to predict the outcome of PULs; however (Condous et al 2004), these need to be tested prospectively in multicentre trials. Figure 25.1 shows an algorithm for the management of women with PUL.

Figure 25.1 Algorithm for the management of women with pregnancy of unknown location (PUL). β-hCG, beta human chorionic gonadotrophin.

Management of Ectopic Pregnancy

Figure 25.2 shows a summary of the algorithm for managing ectopic pregnancies. The following section should be read in tandem with this figure.

Figure 25.2 Algorithm for management of ectopic pregnancies. PUL, pregnancy of unknown location; β-hCG, beta human chorionic gonadotrophin; D4, day 4; D7, day 7.

Surgical treatment

Laparoscopy or laparotomy

All the surgical procedures for the treatment of ectopic pregnancy can be accomplished via laparoscopy or laparotomy. The main factors which will determine the preferred approach are:

• haemodynamic condition of the patient;

• size and location of the ectopic pregnancy;

• laparoscopic experience of the surgeon; and

• availability of appropriate equipment for laparoscopic surgery.

Laparotomy should be considered where the patient is haemodynamically unstable, there is a suspicion of extensive abdominal and pelvic adhesions making laparoscopy difficult, or the pregnancy is cornual or ovarian. A ruptured ectopic pregnancy does not necessarily require laparotomy, the main aim being use of the most expedient method. In the hands of an experienced laparoscopist, all the indications for laparotomy are relative rather than absolute.

A systematic review has shown that laparoscopic salpingostomy is significantly less successful than the open surgical approach in the elimination of tubal ectopic pregnancy (two randomized controlled trials, n = 165, OR 0.28, 95% CI 0.09–0.86) due to a significantly higher persistent trophoblast rate in laparoscopic surgery (OR 3.5, 95% CI 1.1–11). However, the laparoscopic approach is significantly less costly than open surgery. Long-term follow-up (n = 127) shows no evidence of a difference in intrauterine pregnancy rate, but there is a non-significant tendency to a lower repeat ectopic pregnancy rate (Hajenius et al 2007). The laparoscopic approach is associated with less intraoperative blood loss, lower analgesic requirements and a shorter duration of hospitalization and convalescence. An additional benefit of the laparoscopic approach is significantly fewer adhesions at the surgical site compared with those treated by laparotomy (Lundorff et al 1991).

Salpingectomy

Removal of the fallopian tube is still the most common method for the treatment of ectopic pregnancy. It can be performed via laparotomy or laparoscopy. When laparotomy is performed, the easiest way to remove the fallopian tube is by progressive division and ligation of the mesosalpinx. Laparoscopic salpingectomy can be performed with a pretied loop ligature, bipolar diathermy or disposable stapling devices. However, the cost associated with the use of the stapling devices has limited their popularity.

Several methods have been proposed for removal of the fallopian tube from the peritoneal cavity following laparoscopic salpingectomy:

• placement of the fallopian tube into a specially designed plastic bag introduced via a 10-mm port. The bag is then sealed at its neck and removed intact through the port;

• morcellation of the fallopian tube and removal by a 10-mm grasping forceps; or

• posterior colpotomy and removal of the tube intact via the pouch of Douglas.

When to consider salpingectomy

Salpingectomy is indicated for women with:

• ruptured tubal pregnancy;

• recurrent ectopic pregnancy in a tube already treated conservatively;

• previous sterilization and reversal of sterilization;

• previous tubal surgery for infertility; or

• pre-existing tubal damage as a consequence of a frozen pelvis.

No randomized controlled trials have compared salpingectomy and salpingotomy. Results from four cohort studies suggest that there may be a higher subsequent intrauterine pregnancy rate associated with salpingotomy, but the magnitude of this benefit may be small (Royal College of Obstetricians and Gynaecologists 2004). The possibility of further ectopic pregnancies in the conserved tube should also be discussed if salpingotomy is being considered. In the presence of a healthy contralateral tube, there is no clear evidence that salpingotomy should be used in preference to salpingectomy.

Finally, the debate over salpingectomy or salpingo-oophorectomy no longer exists. Removal of a healthy ovary is now considered both unnecessary and unjustified.

Linear salpingotomy

An incision is made on the antimesenteric border of the fallopian tube with either a needle-point monopolar cutting diathermy, scalpel, scissors or laser, and the tubal lumen is exposed. The products of conception are removed with grasping forceps. Sometimes the laparoscopic suction/irrigation probe may be used for aqua dissection and removal of the products of conception by suction. Irrigation of the tubal lumen after removal of the products of conception is essential. Thereafter, haemostasis is achieved with electrocoagulation. The incision can either be closed with a fine non-absorbable suture material such as 6/0 Prolene, or left open (salpingotomy) for healing to occur by secondary intention. Either approach results in satisfactory healing of the tubal wall. A study by Tulandi and Guralnick (1991) showed that there was no significant difference in the number of subsequent intrauterine pregnancies, number of ectopic pregnancies or incidence of adhesion formation with either method.

Linear salpingotomy can be performed via an open or laparoscopic approach. However, regardless of the route used to access the ectopic pregnancy, the surgeon should remember that fallopian tubes are delicate structures. Basic microsurgical principles such as gentle handling of tissues, avoidance of peritoneal damage, thorough irrigation and suction of the peritoneal cavity to remove all the blood clots, and minimizing postoperative adhesion formation should always be applied.

Laparoscopic linear salpingotomy should be considered as the primary procedure of choice for the treatment of tubal ectopic pregnancy in the presence of contralateral tubal disease and where the patient wishes to retain her potential for future fertility.

The use of conservative surgery exposes the women to a small risk of tubal bleeding in the immediate postoperative period, persistent trophoblast and the risk of subsequent ectopic pregnancy in the conserved tube. For identification and treatment of persistent trophoblast, these women should be followed by serial β-hCG measurements post surgery. Treatment with methotrexate can be initiated if β-hCG concentrations fail to fall as expected. Sauer et al (1997) initiated treatment if serum β-hCG was more than 10% of the preoperative level by 10 days after surgery, and Pouly et al (1991) initiated treatment if serum β-hCG was more than 65% of the initial level at 48 h post surgery. Currently, there is insufficient evidence to recommend ideal cut-off concentrations of serum β-hCG for commencing treatment, but protocols for identification and treatment of persistent trophoblast should be present in every unit.

Segmental resection

When the ectopic pregnancy is located in the isthmic segment of the fallopian tube, segmental resection and end-to-end anastomosis constitute one of the conservative surgical therapeutic options. The anastomosis is usually isthmoampullary and can be performed either at the time of the resection (primary) or at a later stage (secondary). However, for isthmic tubal pregnancy, linear salpingotomy is technically less difficult and has a shorter operative time.

Fimbrial evacuation

Fimbrial evacuation is indicated when the pregnancy lies in the fimbrial segment of the fallopian tube. It involves gentle and progressive compression of the tube, starting just proximal to the side of the pregnancy and moving it systematically to the fimbrial end of the tube. When the pregnancy is located at the ampullary part, fimbrial evacuation or ‘milking of the tube’ is not recommended. The external pressure required to propel the ectopic pregnancy and achieve a tubal miscarriage is likely to cause severe damage to the endosalpinx and excessive bleeding from the implantation site, and the procedure is associated with a greater risk of persistent trophoblastic disease (Brosens et al 1984). When milking was compared with linear salpingotomy for ampullary ectopic pregnancy, milking was associated with a two-fold increase in the recurrent ectopic pregnancy rate (Smith et al 1987).

Reproductive outcome

The evaluation of reproductive outcome following surgical treatment for ectopic pregnancy is based on three main parameters:

• tubal patency as assessed by hysterosalpingography or second-look laparoscopy;

• subsequent intrauterine pregnancy rate; and

• recurrent ectopic pregnancy rate.

In patients who underwent radical surgery (salpingectomy), the subsequent intrauterine pregnancy rate was 49.3% (range 36.6–71.4%) and the recurrent ectopic pregnancy rate was 10% (range 5.8–12%). In patients who underwent conservative surgery, the subsequent intrauterine pregnancy rate was 53% (range 38–83%) and the recurrent ectopic pregnancy rate was 14% (range 6.4–21.2%). These results are based on a retrospective analysis of nine studies in a total of 2635 patients, and indicate that conservative surgery is associated with higher subsequent intrauterine pregnancy and higher recurrent ectopic rates compared with radical surgery (Yao and Tulandi 1997). However, the variations in the design of each study make strict comparison of the results very difficult. Similar figures of intrauterine pregnancy (83.7%) and recurrent ectopic pregnancy (8%) have been reported when expectant management is compared with salpingectomy (Helmy et al 2007).

Medical treatment with methotrexate

Early detection of an ectopic pregnancy, the desire to retain fertility, and minimizing costs and surgical morbidity are the main reasons for incorporating medical treatment as an option for managing ectopic pregnancies. The most frequently used drug is methotrexate and this will therefore be discussed in detail, although other agents have been used. Gynaecologists are familiar with methotrexate because the drug has been used very effectively for the treament of gestational trophoblastic disease since 1955 (Li et al 1956). Hreschchyshyn et al (1965) were the first to report on the use of methotrexate in the management of ectopic pregnancy. In 1982, Tanaka et al reported on the treatment of an unruptured interstitial pregnancy with a course of intramuscular methotrexate.

Mode of action and pharmacokinetics

Methotrexate, formerly known as amethopterin, is a potent folic acid antagonist. It inhibits the action of the enzyme dehydrofolate reductase and the coenzyme methyl tetrahydrofolate. These enzymes are essential for the conversion of dehydrofolic acid into tetrahydrofolic acid, and the conversion of deoxyuridylic acid into thymidylic acid which is critical for the synthesis of DNA during the S phase of the cell cycle. Prevention of incorporation of thymidylic acid into DNA results in lack of DNA synthesis and subsequent arrest of cellular division and multiplication. Methotrexate is ideal for inhibition of rapidly growing cells such as trophoblasts.

Methotrexate is almost completely absorbed if administered intramuscularly, with peak serum concentrations reached within 30–60 min. Only 10% of methotrexate undergoes hepatic and intracellular metabolism to polyglutamated products, which also act as active antimetabolites. Up to 90% of the drug is secreted unchanged in the urine within 48 h, and therefore impaired renal function can result in prolonged exposure of tissues to high levels of methotrexate or its metabolic products.

Side-effects of methotrexate

The two most significant factors that determine the intensity of methotrexate toxicity are the serum concentration of drug and the duration of exposure to methotrexate. The drug affects all rapidly proliferating cells, and therefore the main side-effects originate from suppression of bone marrow and epithelial cells of the gastrointestinal tract. The spectrum of possible side-effects include: abdominal pain in 75% of women following treatment, conjunctivitis, stomatitis, gastritis, diarrhoea, alopecia (reversible), myelosuppression, liver toxicity, photosensitivity (skin reaction), pneumonitis, renal toxicity and anaphylaxis.

The incidence of side-effects among patients treated with high-dose or multidose methotrexate is reported to be 20–30%. Single-dose regimens, however, currently used for the non-surgical treatment of ectopic pregnancy are associated with minor or no side-effects. Resolution of the adverse effects usually occurs within 72 h after termination of therapy.

Guidelines for the prescription of methotrexate

In order to achieve early identification and prevention of serious side-effects, the Committee on the Safety of Medicines has issued the following guidelines for the prescription of methotrexate.

Myelosuppression

Pretreatment full blood count needs to be documented as a baseline blood count. If an acute decrease in the white cell or platelet count occurs, treatment should be discontinued immediately.

Liver toxicity

Pre-existing abnormal liver function or concurrent abnormal liver function tests should alert medical staff to potential liver toxicity. Again, pretreatment baseline liver function tests are essential.

Renal function

Assessment of renal function is essential prior to treatment because the primary route of excretion of methotrexate is the kidney. Caution should be exercised when there is concurrent medication use with non-steroidal anti-inflammatory drugs as these may adversely affect renal function and further increase serum methotrexate concentrations.

Folinic acid rescue

Methotrexate morbidity can be minimized by the concurrent administration of the ‘antidote’ folinic acid in the form of calcium folinate which is then converted to 5-methyltetrahydrofolate.

Who will benefit from methotrexate?

The key for successful and safe medical treatment of ectopic pregnancy is proper selection of patients. Methotrexate should be reserved for patients who have an unruptured tubal pregnancy, are clinically asymptomatic and haemodynamically stable. The indications and contraindications for methotrexate therapy are listed below.

Indications

• Initial pretreatment plasma β-hCG concentration below 3000 IU/l.

• Tubal diameter less than 2 cm as defined by direct visualization or ultrasound.

• Absent fetal heartbeat on ultrasound scan.

Contraindications

• Symptomatic patients.

• Haemodynamically unstable patients.

• Presence of haemoperitoneum.

• Initial pretreatment β-hCG concentration greater than 3000 IU/l.

• Tubal diameter more than 2 cm.

• Presence of fetal heartbeat on ultrasound scan.

• Evidence of liver disease, renal insufficiency or myelosuppression.

• Inability of the patient to return for follow-up care.

Although medical management can be successful at serum β-hCG concentrations greater than 3000 IU/l, quality-of-life data suggest that methotrexate should only be used with serum β-hCG concentrations below 3000 IU/l. Shalev et al (1995) reported that following intratubal administration of methotrexate, the failure rate was 24% when the tubal diameter was less than 2 cm and 48% when the diameter was more than 2 cm. The authors recommend that methotrexate should only be used when the size of the ectopic pregnancy is less than 2 cm. Stovall and Ling (1993) found that the administration of methotrexate to patients who had a fetal heartbeat on ultrasound scan was associated with a 14.3% failure rate. In the group of patients without fetal cardiac activity, the failure rate was 4.7%.

Prior to commencing medical treatment, women should be given clear information about the possible need for further treatment and adverse effects following treatment, and they should be able to return easily for assessment at any time during follow-up. Patients should be advised regarding adequate compliance; no use of alcohol, aspirin, non-steroidal inflammatory drugs or folic acid supplements; avoidance of sexual intercourse and sunlight exposure; fluid intake of at least 1.5 l/day; saline mouth washes daily; and delay of pregnancy for at least 3 months after treatment because of the teratogenecity of methotrexate.

Routes of administration and dose

Methotrexate can be administered systemically (intravenously or intramuscularly) or locally using a laparoscopic or transvaginal ultrasound-guided approach. The transvaginal administration of methotrexate under sonographic guidance requires visualization of an ectopic gestational sac, and specific skills and expertise of the clinician. This mode of administration is less invasive and more effective than the laparoscopically ‘blind’ intratubal injection. Two randomized controlled trials (Fernandez et al 1994, Cohen et al 1996) suggested that local injection of methotrexate was equivalent in effect to systemic methotrexate. The major advantages of direct injection of methotrexate into the tube include smaller dose of the drug, higher tissue concentration and fewer systemic side-effects. In comparison, systemic methotrexate is practical, easier to administer and less dependent on clinical skills. The published literature includes a wide spectrum of protocols for the systemic administration of methotrexate, including:

• standard single dose (60–200 mg IM);

• fixed multiple dose regimen: methotrexate 1 mg/kg body weight IM on days 0, 2, 4, 6 alternated with folinic acid 0.1 mg/kg orally on days 1, 3, 5, 7 followed by 6 days without

• medication. A second course is given on day 14 if the serum hCG concentration on that day is 40% above the initial value on day 0; and

• variable single dose regimen (1 mg/kg or 50 mg/m² of body surface).

The success rate of all three regimens is comparable. Three trials have shown that the administration of methotrexate in a single IM dose of 50 mg/m² body surface area is associated with a complete reabsorption rate of 92%, a subsequent intrauterine pregnancy rate of 58% and a recurrent ectopic pregnancy rate of 9% (Hajenius et al 2007). Reabsorption rate is defined as the time interval between initiation of therapy and undetectable serum hCG concentration. The mean reabsorption time is 30 days, ranging from 5 to 100 days. Tubal patency is reported to be 50–100% (mean 71%) after systemic administration of methotrexate.

Laparoscopic surgery vs methotrexate

O’Shea et al (1994) published the results of a prospective randomized trial comparing intra-amniotic methotrexate with carbon dioxide laser laparoscopic salpingotomy. The salpingotomy group and the local methotrexate group had comparable success rates: 87.5% and 89.7%, respectively. A systematic review showed that systemic methotrexate in a fixed multiple dose IM regimen has a non-significant tendency to higher treatment success than laparoscopic salpingostomy (OR 1.8, 95% CI 0.73–4.6). No significant differences are found in long-term follow-up for intrauterine pregnancy and repeat ectopic pregnancy (Hajenius et al 2007). Methotrexate appears to be effective in a selective group of patients with ectopic pregnancy. Direct and indirect costs of medical therapy are less than half of those associated with laparoscopy. Intramuscular administration eliminates anaesthetic and surgical risks arising from laparoscopic injection of the drug. However, the long-term effect of locally injected pharmacological substances upon the endosalpinx remains largely unknown.

Serum β-hCG levels should be checked 4 and 7 days after the first dose of methotrexate, and a further dose is given if β-hCG levels have failed to fall by more than 15% between days 4 and 7. The level of serum β-hCG is likely to increase between days 1 and 4 after treatment. It has been suggested that although methotrexate is arresting mitosis in the cytotrophoblast, the syncytiotrophoblast may still be able to produce β-hCG. Also, destruction of the cells releases more β-hCG into the systemic circulation. Approximately 15% of women require more than a single dose of methotrexate, and 7% experience tubal rupture during follow-up (Yao and Tulandi 1997, Sowter et al 2001).

It is important to recognize that abdominal pain occurs in 30–50% of patients 6–7 days after the administration of methotrexate. The gynaecologist has to decide if the pain is due to imminent tubal rupture, existing tubal rupture or simply represents postmethotrexate pain. Postmethotrexate pain is localized in the lower abdomen. Its precise aetiology remains largely unknown, but it is likely to be due to a combination of factors, including destruction of the trophoblastic tissue and transtubal haemorrhage leading to peritoneal irritation. Close monitoring of patients is essential in order to avoid unnecessary surgical intervention in the form of laparoscopy or laparotomy. According to Stovall (1995), regular assessment of haemoglobin is the most useful parameter. The pain usually subsides within 24–48 h and occasionally precedes or follows an episode of vaginal bleeding.

Brown et al (1991) assessed the clinical value of serial transvaginal ultrasonography in the monitoring of ectopic pregnancies treated with methotrexate, and concluded that routine ultrasonography is not necessary after methotrexate treatment. There was no correlation between the pattern of resolution of β-hCG levels and sonographic findings.

Mifepristone and methotrexate

Mifepristone (RU486) has been used in combination with methotrexate for the medical treatment of ectopic pregnancy (Gazvani and Emery 1999). A non-randomized study from France (Perdu et al 1998) only reported one failure in the 30 patients treated with IM methotrexate and oral mifepristone, and 11 failures in 42 patients treated with methotrexate alone. The authors concluded that the combination of mifepristone and methotrexate decreased the risk of failure in medical treatment of ectopic pregnancy. However, further prospective randomized studies are needed to evaluate the role of mifepristone in the medical management of ectopic pregnancy.

Controlled expectant management

When only expectant management was available, it was associated with high morbidity and mortality because the patients who failed expectant management usually presented with significant clinical symptoms, indicating rupture of the ectopic pregnancy. Since then, due to the advancement in ultrasound and rapid access to quantitative β-hCG assays, the nature of expectant management has changed in that it is now controlled and monitored.

Ten prospective studies, with a total of 347 patients managed expectantly, demonstrated that 69% of the ectopic gestations resolved spontaneously. All patients were haemodynamically stable and had decreasing serum β-hCG levels. However, other variables such as size and location of the ectopic pregnancy or presence of fetal cardiac activity were not always specified (Yao and Tulandi 1997). Interventions are required in approximately 23–29% of cases with expectant management (Hahlin et al 1995).

The essential prerequisite for successful expectant management of ectopic gestation is appropriate selection of patients. The Royal College of Obstetricians and Gynaecologists (2004) issued guidelines for the management of tubal pregnancy. The conclusion was that expectant management is more likely to be successful in the following circumstances:

• clinically stable asymptomatic woman;

• baseline serum β-hCG concentrations below 1000–2000 IU/l;

• haemoperitoneum of less than 50–100 ml;

• a tubal mass less than 2 cm; and

• absence of recognizable fetal parts on ultrasound.

Women managed expectantly should be followed twice weekly with serial β-hCG measurements (ideally β-hCG should be <50% of the initial level within 7 days) and weekly transvaginal ultrasound examination to assess reduction in size of adnexal mass by the end of the first week. Thereafter, weekly β-hCG concentrations and transvaginal ultrasound examination should be undertaken until β-hCG concentrations are less than 20 IU/l. It must be emphasized, however, that tubal rupture has been reported in cases of very low and decreasing serum β-hCG concentration, and expectant management often involves long periods of hospitalization and follow-up. Women should be counselled about the importance of compliance and should be within easy access of medical help.

Anti-D immunoglobulin

All non-sensitized, Rhesus-negative women with suspected or confirmed ectopic pregnancy should receive 250 IU (50 µg) of anti-D immunoglobulin.

Non-Tubal Ectopic Pregnancy

Cervical pregnancy

A cervical pregnancy is one that implants entirely within the cervical canal. The reported incidence ranges from one in 2500 to one in 18,000 deliveries (0.2% of ectopic pregnancies). Predisposing factors for the development of cervical pregnancy include previous instrumentation of the endocervical canal, previous caesarean delivery, Asherman’s syndrome and exposure to DES. Sporadic cases of cervical pregnancy have been reported after controlled ovarian stimulation with human menopausal gonadotrophins and intrauterine insemination, as well as IVF (Weyerman et al 1989).

Cervical pregnancy produces profuse vaginal bleeding without associated cramping pain. In some cases of advanced pregnancies, it can be associated with urinary symptoms. The clinical and ultrasound criteria for the diagnosis of cervical pregnancy are summarized below (Hofmann et al 1987).

Clinical criteria

The external os is usually dilated and the cervix feels distended or globular with a small uterus on bimanual examination. Curettage of the endometrial cavity reveals no evidence of trophoblastic tissue. The products of conception are confined within the cervix and the internal cervical os is closed.

Ultrasound criteria

No evidence of an intrauterine gestational sac, ballooned cervical canal/barrel-shaped cervix, gestational sac in the endocervix below the uterine arteries, closed internal os and Doppler blood flow around the sac.

The differential diagnoses include an intrauterine pregnancy with a low implantation site (isthmicocervical pregnancy), an ongoing spontaneous intrauterine miscarriage, a cervical carcinoma, a cervical or prolapsed submucosal myoma, placenta praevia and trophoblastic tumour.

Early forms of cervical pregnancy are managed conservatively. Conservative treatment includes the use of systemic methotrexate, intra-amniotic feticide, reduction of blood supply and tamponade. Transvaginal ultrasound-guided local injection of the ectopic pregnancy with methotrexate or potassium chloride has been shown to be successful (Benson and Doubilet 1996). This technique has been used for the treatment of heterotopic coexistent cervical and intrauterine pregnancy. Similarly, systemic methotrexate alone is 91% efficacious in the treatment of cervical ectopic pregnancy, and a combination of intra-amniotic and systemic methotrexate therapy has been used to increase the chance of successful treatment (Kung and Chang 1999). Currently, there is no recommendation for the optimal dose and route of administration of methotrexate for the treatment of cervical ectopic pregnancy, although gestational age less than 9 weeks, serum β-hCG below 10,000 IU/l, absent fetal cardiac activity and crown–rump length less than 10 mm are associated with a greater chance of success with conservative treatment (Hung et al 1998).

Removal of the products of conception from the cervical canal by suction curettage is likely to stop the haemorrhage. Conservative measures used to arrest bleeding include packing of the uterus and cervix, and insertion of an intracervical 30-ml Foley catheter to arrest bleeding, insertion of sutures to ligate the lateral cervical vessels or placement of a cervical cerclage (Kirk et al 2006). In some cases, because of the depth of trophoblastic invasion, major blood vessels are involved and more radical measures such as bilateral uterine artery or internal iliac artery ligation are necessary to control the bleeding. When all the conservative measures have failed to arrest the bleeding, hysterectomy is required. Preoperatively, the patient must be informed about the possibility of hysterectomy and sign the appropriate consent.