Diagnosis of ectopic pregnancy and miscarriage

National Collaborating Centre for Women's and Children's Health (UK)

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women's and Children's Health (UK). Ectopic Pregnancy and Miscarriage: Diagnosis and Initial Management in Early Pregnancy of Ectopic Pregnancy and Miscarriage. London: RCOG Press; 2012 Dec. (NICE Clinical Guidelines, No. 154.)

See 2019 Partial Update

In April 2019 NICE updated its guideline on ectopic pregnancy and miscarriage. See the evidence reviews for the areas in which new recommendations were developed. The 2012 recommendations have been retained in the new guideline. This 2012 full guideline includes the evidence supporting the 2012 recommendations and has not been updated.

In April 2019 NICE updated its guideline on ectopic pregnancy and miscarriage. See the evidence reviews for the areas in which new recommendations were developed. The 2012 recommendations have been retained in the new guideline. This 2012 full guideline includes the evidence supporting the 2012 recommendations and has not been updated.

Cover of Ectopic Pregnancy and Miscarriage

Ectopic Pregnancy and Miscarriage: Diagnosis and Initial Management in Early Pregnancy of Ectopic Pregnancy and Miscarriage.

Show details

Contents

< Prev Next >

6Diagnosis of ectopic pregnancy and miscarriage

6.1. Signs and symptoms of ectopic pregnancy

Review question

What are the signs and symptoms associated with ectopic pregnancy?

Introduction

Ectopic pregnancy is a relatively common and potentially life-threatening complication of pregnancy. Despite this, morbidity and mortality attributable to failure to consider the diagnosis, and therefore missed or delayed diagnosis, continues to be problematic. This is often due to misconceptions and ignorance of symptoms and signs of an ectopic pregnancy. This review seeks to clarify the relative importance of these individual factors in diagnosing an ectopic pregnancy.

Description of included studies

Twenty-nine studies were included in this review (Aboud & Chaliha, 1998; Al-Suleiman & Khwaja, 1992; Banerjee et al., 1999; Barnhart et al., 2003; Barnhart et al., 2006; Bouyer et al., 2002; Buckley et al., 1998; Choi et al., 2011; Clancy & Illingworth, 1989; Condous et al., 2007; Diamond et al., 1994; Dimitry, 1989; Downey & Zun, 2011; Easley et al., 1987; Goksedef et al., 2011; Gonzalez & Waxman, 1981; Hutton & Narayan, 1986; Jabbar & Al-Wakeel, 1980; Jiao et al., 2008; Kazandi & Turan, 2011; Larrain et al., 2011; Makinen et al., 1984; Menon et al., 2007; Michelas et al., 1980; Powers, 1980; Raziel et al., 2004; Shaunik et al., 2011; Tsai et al., 1995; Wong & Suat, 2000).

The included studies consist of one case–control study (Barnhart et al., 2006) and four prospective observational studies (Banerjee et al., 1999; Buckley et al., 1998; Condous et al., 2007; Shaunik et al., 2011), while the remainder were retrospective case series.

The studies were conducted in the UK (Aboud & Chaliha, 1998; Banerjee et al., 1999; Clancy & Illingworth, 1989; Condous et al., 2007; Dimitry, 1989), the USA (Barnhart et al., 2003; Barnhart et al., 2006; Buckley et al., 1998; Diamond et al., 1994; Downey & Zun, 2011; Easley et al., 1987; Gonzalez & Waxman, 1981; Menon et al., 2007; Powers, 1980; Shaunik et al., 2011), France (Bouyer et al., 2002; Larrain et al., 2011;), Finland (Makinen et al., 1984), Greece (Michelas et al., 1980), Turkey (Goksedef et al., 2011; Kazandi & Turan, 2011), New Zealand (Hutton & Narayan, 1986), China (Jiao et al., 2008; Tsai et al., 1995), Singapore (Wong & Suat, 2000), South Korea (Choi et al., 2011), Israel (Raziel et al., 2004), and Saudi Arabia (Al-Suleiman & Khwaja, 1992; Jabbar & Al-Wakeel, 1980).

All studies evaluated the presenting signs and symptoms of women diagnosed with an ectopic pregnancy (EP), or reported the frequency of risk factors among the participants. Two studies included only cases of ovarian pregnancy (Choi et al., 2011; Raziel et al., 2004), one study only included cases of proximal ectopic pregnancy (Larrain et al., 2011) and one study included only cases of caesarean scar pregnancy (Jiao et al., 2008). Three studies evaluated the signs and symptoms of women initially classified as having pregnancies of unknown location (PUL) who were later diagnosed with an ectopic pregnancy (Banerjee et al., 1999; Condous et al., 2007; Shaunik et al., 2011). The GDG recognised that the key clinical question centred on recognition of signs and symptoms of ectopic pregnancy; however it felt it might also be helpful to report the risk factors for ectopic pregnancy if these were available from included studies as it wanted to highlight the limitations inherent in using risk factors to help make a diagnosis.

Evidence profile

When more than one study reported a risk factor, symptom or sign, the reported percentage is the median frequency derived from the reported frequency in each study. The range of frequencies given is simply the minimum and maximum frequency as reported in the studies that included that particular risk factor, symptom or sign. These have been presented in order to provide some guidance as to which risk factors, signs and symptoms are most frequently associated with ectopic pregnancy and to illustrate the wide range of possible presenting symptoms and signs.

Evidence quality has been downgraded if the studies were retrospective, had a small sample size (N of 50 or fewer), or if five or fewer studies reported the finding and there was a serious issue of indirectness with at least one of the study populations. Unusual study populations are detailed in the table.

Risk factors and symptoms are presented in order of decreasing frequency; signs are presented in the order in which they might be elicited.

Table 6.1

GRADE summary of findings for the risk factors, symptoms, and signs of ectopic pregnancy.

Evidence statements

Unless otherwise stated, all the evidence is of moderate quality.

Risk factors for ectopic pregnancy

Evidence from six studies showed that, on average, 37% of women with ectopic pregnancy had no risk factors for ectopic pregnancy.

Evidence from three studies showed that, on average, 48% of women with ectopic pregnancy smoked cigarettes (low quality).

Evidence from 15 studies showed that, on average, 23% of women with ectopic pregnancy had a prior pelvic or abdominal surgery.

The evidence showed that 10–20% of women with ectopic pregnancy had a history of a sexually transmitted infection (three studies), a previous elective abortion (seven studies), a history of infertility (10 studies), a previous miscarriage (seven studies), a history of pelvic inflammatory disease (18 studies), a previous ectopic pregnancy (16 studies) or a history of intrauterine contraceptive device (IUCD) use (13 studies).

The evidence showed that less than 10% of women with ectopic pregnancy had a history of oral contraceptive pill use (seven studies), prior tubal surgery (11 studies) or endometriosis (two studies, low quality).

Symptoms reported

The evidence showed that the majority of women with ectopic pregnancy presented with abdominal or pelvic pain (93%, 21 studies), amenorrhea (73%, 11 studies) or vaginal bleeding (64%, 25 studies).

The evidence showed that 20–30% of women with ectopic pregnancy presented with breast tenderness (three studies, low quality), gastro-intestinal symptoms (10 studies) or dizziness, fainting or syncope (12 studies).

The evidence showed that 10–20% of women with ectopic pregnancy presented with shoulder tip pain (seven studies).

The evidence showed that less than 10% of women with ectopic pregnancy presented with urinary symptoms (three studies, low quality), passage of tissue (two studies), rectal pressure or pain on defecation (three studies, low quality) or no symptoms (three studies, low quality).

Signs identified on examination

The evidence showed that the majority of women with ectopic pregnancy had pelvic tenderness (91%, one study), adnexal tenderness (82%, seven studies) or abdominal tenderness (78%, 11 studies).

The evidence showed that 40–75% of women with ectopic pregnancy had cervical motion tenderness (eight studies), pallor (one study) or rebound tenderness or peritoneal signs (nine studies).

The evidence showed that 20–40% of women with ectopic pregnancy had abdominal distension (two studies, low quality), an enlarged uterus (six studies), an adnexal mass (nine studies) or tachycardia or hypotension (five studies, low quality).

The evidence showed that less than 20% of women with ectopic pregnancy had a palpable pelvic mass (two studies), were collapsed or in shock (eight studies) or had orthostatic hypotension (three studies).

Evidence to recommendations

Relative value placed on the outcomes considered

In conducting this review, the GDG was keen to identify uncommon signs and symptoms associated with ectopic pregnancy.

Consideration of clinical benefits and harms

The group recognised that there was a wide range of symptoms associated with ectopic pregnancy, including some non-specific symptoms such as gastro-intestinal symptoms. Given this, it felt that there is value in healthcare professionals always considering pregnancy in women of childbearing age presenting with these symptoms, and thus they should consider conducting a pregnancy test. If a woman is found to be pregnant, the GDG agreed that ectopic pregnancy should be considered, as many of the symptoms of ectopic pregnancy are the same as those of pregnancy in general. The group felt strongly that all healthcare professionals who provide care to women of reproductive age should have access to pregnancy tests, so that women with a suspected ectopic pregnancy can be identified and referred promptly and appropriately.

The evidence showed that cervical motion tenderness, pelvic tenderness and pain or tenderness in the abdominal region were associated with ectopic pregnancy. As a result, the group recommended that women with tenderness or pain in these areas and a positive pregnancy test should be immediately referred for further assessment. ‘Immediately’ here means the further assessment should take place as soon as possible, and at least within a few hours of the initial assessment. Where possible the woman should go directly from the place of the initial assessment to the early pregnancy assessment service (EPAS) – or alternative out of hours gynaecology service if the EPAS is not open – so as not to incur further delay.

The GDG recognised that risk factors were not a helpful method for identifying women with an ectopic pregnancy, as about a third of women with an ectopic pregnancy had no identifiable risk factors. The group agreed that even in the absence of risk factors, it would still be necessary for a healthcare professional to rule out the possibility of ectopic pregnancy, and thus agreed that risk factors should not be used as a diagnostic aid.

Consideration of health benefits and resource use

The group recognised that lowering the index of suspicion for ectopic pregnancy was likely to lead to an increase in the number of women being offered a pregnancy test and thus an increase in the number of women referred to an EPAS. However, the group felt that this approach was very likely to be cost effective given the potential large loss of quality adjusted life years (QALYs) associated with missing a diagnosis of ectopic pregnancy.

Quality of evidence

The majority of evidence was of moderate quality and, as such, the GDG felt that it was sufficient to make recommendations.

Information giving and emotional support

From their clinical experience, the GDG members thought that uncertainty about what was happening would increase women's anxiety. Therefore, they felt that it was important that women who required referral were given information about why the referral was necessary, and what they might expect when they arrived at the EPAS.

Other considerations

The GDG was keen to emphasise that ectopic pregnancy can present with a variety of symptoms and agreed that it would be helpful for healthcare professionals to be able to refer to a list of potential signs and symptoms. Although reported in the evidence, the GDG members felt that it was not appropriate to include ‘adnexal mass’ or ‘palpable pelvic mass’ in the list of signs and symptoms. From their clinical experience, they noted that palpation can increase the risk of an ectopic pregnancy rupturing. They also felt that, while palpation to detect an internal mass might once have been used in the diagnosis of ectopic pregnancy, the development of new diagnostic modalities (such as transvaginal ultrasound and biochemical tests) has meant that it is no longer appropriate.

The GDG felt that a decision tool, incorporating risk factors, signs and symptoms, could be very valuable in decreasing the likelihood of women with ectopic pregnancies being misdiagnosed and therefore mismanaged. The group noted, however, that the evidence from this review was not sufficient to develop and validate such a tool, and therefore decided that a research recommendation was warranted. The group felt that such a tool would be extremely valuable in allowing healthcare professionals, particularly non-specialists, to evaluate a woman's likelihood of having an ectopic pregnancy and to determine the level of urgency of any resulting referral.

The GDG felt it important to note that not all women presenting to a healthcare professional with bleeding in early pregnancy need to be referred for a transvaginal ultrasound scan. For women who are reporting to be less than 6 weeks pregnant with bleeding but no pain, expectant management could be undertaken. The GDG felt that this was justified, because at gestations of earlier than 6 weeks, the pregnancy is likely to be too small to yield any information about viability. In addition, from their clinical experience, the GDG members agreed that many women experience spotting in early pregnancy which resolves without need for further intervention. They agreed that these women should be advised to take a pregnancy test 7–10 days later and that they should return in the interim if their symptoms continue or worsen.

The group agreed that in all other cases, that is where women are more than 6 weeks pregnant with blood loss or are in pain, or where there is uncertainty about the pregnancy's gestation, referral to a dedicated early pregnancy service should be made in order that an ultrasound scan can be carried out.

Recommendations

Number	Recommendation
9	Refer women who are haemodynamically unstable, or in whom there is significant concern about the degree of pain or bleeding, directly to A&E.
10	Be aware that atypical presentation for ectopic pregnancy is common.
11	Be aware that ectopic pregnancy can present with a variety of symptoms. Even if a symptom is less common, it may still be significant. Symptoms of ectopic pregnancy include: common symptoms: abdominal or pelvic pain amenorrhoea or missed period vaginal bleeding with or without clots other reported symptoms breast tenderness gastrointestinal symptoms dizziness, fainting or syncope shoulder tip pain urinary symptoms passage of tissue rectal pressure or pain on defecation.
12	Be aware that ectopic pregnancy can present with a variety of signs on examination by a healthcare professional. Signs of ectopic pregnancy include: more common signs: pelvic tenderness adnexal tenderness abdominal tenderness other reported signs: cervical motion tenderness rebound tenderness or peritoneal signs pallor abdominal distension enlarged uterus tachycardia (more than 100 beats per minute) or hypotension (less than 100/60 mmHg) shock or collapse orthostatic hypotension.
13	During clinical assessment of women of reproductive age, be aware that: they may be pregnant, and think about offering a pregnancy test even when symptoms are non-specific and the symptoms and signs of ectopic pregnancy can resemble the common symptoms and signs of other conditions – for example, gastrointestinal conditions or urinary tract infection.
14	All healthcare professionals involved in the care of women of reproductive age should have access to pregnancy tests.
15	Refer immediately to an early pregnancy assessment service (or out-of-hours gynaecology service if the early pregnancy assessment service is not available) for further assessment women with a positive pregnancy test and the following on examination: pain and abdominal tenderness or pelvic tenderness or cervical motion tenderness.
16	Exclude the possibility of ectopic pregnancy, even in the absence of risk factors (such as previous ectopic pregnancy), because about a third of women with an ectopic pregnancy will have no known risk factors.
17	Refer to an early pregnancy assessment service (or out-of-hours gynaecology service if the early pregnancy assessment service is not available) women with bleeding or other symptoms and signs of early pregnancy complications who have: pain or a pregnancy of 6 weeks gestation or more or a pregnancy of uncertain gestation. The urgency of this referral depends on the clinical situation.
18	Use expectant management for women with a pregnancy of less than 6 weeks gestation who are bleeding but not in pain. Advise these women: to repeat a urine pregnancy test after 7–10 days and to return if it is positive a negative pregnancy test means that the pregnancy has miscarried to return if their symptoms continue or worsen.
19	Refer women who return with worsening symptoms and signs that could suggest an ectopic pregnancy to an early pregnancy assessment service (or out-of-hours gynaecology service if the early pregnancy assessment service is not available) for further assessment. The decision about whether she should be seen immediately or within 24 hours will depend on the clinical situation.
20	If a woman is referred to an early pregnancy assessment service (or out-of-hours gynaecology service if the early pregnancy assessment service is not available), explain the reasons for the referral and what she can expect when she arrives there.

Number	Research recommendation
RR 3	Research should be undertaken to design and validate a decision tool for evaluating signs, symptoms and risk factors for correctly identifying ectopic pregnancy

6.2. Ultrasound for determining a viable intrauterine pregnancy

Review question

What is the diagnostic value of ultrasound for determining a viable intrauterine pregnancy?

Introduction

The application of ultrasound is well established and important in the assessment and evaluation of early pregnancy events and complications. Its use in early pregnancy assessment may be routine but practices vary considerably. Although high resolution transvaginal ultrasound has been widely adopted, the limitations of accuracy in defining ultra-small structures, such as a fetal heart at early gestations, are generally acknowledged. The aim of this review was to identify the point at which the viability of a pregnancy can be definitively confirmed using ultrasound. This threshold also represents the point at which miscarriage can be definitively diagnosed.

Description of included studies

Fifteen studies were included in this review (Abaid et al., 2007; Abdallah et al., 2011; Bree et al., 1989; Brown et al., 1990; Cacciatore et al., 1990; de Crespigny, 1988; Ferrazzi et al., 1993; Goldstein, 1992; Hassan et al., 2009; Levi et al., 1988; Levi et al., 1990; Pennell et al., 1991; Rempen, 1990; Rowling et al., 1999; Steinkampf et al., 1997).

The included studies consist of nine prospective observational studies (Abdallah et al., 2011; Bree et al., 1989; Cacciatore et al., 1990; de Crespigny, 1988; Goldstein, 1992; Hassan et al., 2009; Pennell et al., 1991; Rempen, 1990; Rowling et al., 1999), five retrospective observational studies (Abaid et al., 2007; Ferrazzi et al., 1993; Levi et al., 1988; Levi et al., 1990; Steinkampf et al., 1997) and one partially retrospective observational study (Brown et al., 1990).

The studies were conducted in the UK (Abdallah et al., 2011; Hassan et al., 2009), Germany (Rempen, 1990), Italy (Ferrazzi et al., 1993), Finland (Cacciatore et al., 1990), the USA (Abaid et al., 2007; Bree et al., 1989; Brown et al., 1990; Goldstein, 1992; Pennell et al., 1991; Rowling et al., 1999; Steinkampf et al., 1997), Canada (Levi et al., 1988; Levi et al., 1990) and Australia (de Crespigny, 1988).

All included studies evaluated the use of transvaginal ultrasound for visualising fetal cardiac activity in intrauterine pregnancies, and stratified their findings by gestational age, crown–rump length (CRL) or gestation sac size. Two studies additionally compared the performance of transabdominal ultrasound in visualising cardiac activity (Ferrazzi et al., 1993; Pennell et al., 1991).

Evidence profile

Table 6.2

GRADE summary of findings for evaluation of ultrasound for determining a viable intrauterine pregnancy.

Evidence statements

All evidence statements relate to the use of transvaginal ultrasound (TVU) unless otherwise stated.

Visualisation of cardiac activity by crown–rump length

One study provided high quality evidence that cardiac activity can be visualised in all viable fetuses with a CRL of at least 3 mm. One further study provided high quality evidence that cardiac activity can be visualised in all viable fetuses with a CRL of at least 5 mm using transvaginal ultrasound, and at least 9 mm using transabdominal ultrasound. Three studies provided moderate quality evidence that cardiac activity can be visualised in all viable fetuses with a CRL of at least 6.0 mm, at least 5 mm and at least 3.5 mm respectively. One study provided low quality evidence that viability can be correctly determined in all fetuses with a CRL of 5.3 mm. Two further studies provided low quality evidence that cardiac activity can be visualised in all viable fetuses with a CRL of at least 4 mm.

Visualisation of cardiac activity by gestation sac

One study provided high quality evidence that cardiac activity can be visualised in all viable fetuses with a chorionic cavity diameter of at least 18.3 mm. One study provided moderate quality evidence that viability can be correctly determined in all fetuses with a gestation sac of 21 mm, in the absence of an embryo, with or without a yolk sac. One study provided moderate quality evidence that cardiac activity can be visualised in all viable fetuses with a gestation sac diameter exceeding 9 mm. Two studies provided low quality evidence that cardiac activity can be visualised in all viable fetuses with a gestation sac diameter of at least 13 mm or exceeding 12 mm. One study provided very low quality evidence that cardiac activity can be visualised in all viable fetuses with a gestation sac diameter of at least 16 mm while another study provided very low quality evidence that cardiac activity can be visualised in 99% of viable fetuses with a gestation sac diameter of at least 19 mm. One study provided very low quality evidence that cardiac activity can be reliably detected when the gestation sac diameter exceeds 18 mm.

Visualisation of cardiac activity by gestational age

One study provided high quality evidence that cardiac activity can be visualised in all viable fetuses with a gestational age of at least 46 days. One study provided moderate quality evidence that cardiac activity can be visualised in all viable fetuses with a gestational age exceeding 40 days. One study provided very low quality evidence that cardiac activity can be visualised in 99% of viable fetuses with a gestational age of at least 45.5 days. One study provided very low quality evidence that cardiac activity can be visualised in all viable fetuses with a gestational age of at least 35 days using transvaginal ultrasound, and at least 37 days using transabdominal ultrasound. One further study provided very low quality evidence that cardiac activity can be reliably detected in all viable fetuses with a gestational age exceeding 43 days.

Evidence to recommendations

Relative value placed on the outcomes considered

The outcomes for this review were the thresholds of gestational age, CRL or gestation sac diameter at which 100% of fetuses that later proved to be viable had cardiac activity visible on ultrasound. The GDG noted that, even when women could be completely certain about the date of intercourse or last menstrual period, variation in the menstrual cycle and rate of fetal growth might result in inaccurate estimates of gestational age. Therefore, the GDG felt that the use of a gestational age threshold alone was not appropriate for the determination of viability. From their own clinical experience the GDG members felt that, where it was possible to measure the crown–rump length, this would provide the most accurate estimate of development. They were also aware of recent work by Pexsters et al. (2011) which found that measurement of mean gestational sac diameter was associated with higher inter-observer variability than measurement of crown–rump length. Therefore, measurement of mean gestational sac diameter was only recommended in cases where a fetal pole could not be identified.

Consideration of clinical benefits and harms

The evidence suggested that transvaginal ultrasound scans are more effective than transabdominal ultrasound scans for visualising structures and this was borne out by the GDG's clinical experience. The group agreed that the majority of scans are now conducted transvaginally and felt that it was appropriate to recommend that a transvaginal scan be recommended in the first instance. It accepted that some women may opt for a transabdominal scan, and that this could be offered as an alternative. However, GDG members felt that it was important that women be given information about the potential limitations of transabdominal scanning so that they could make a fully informed choice.

The GDG felt that it was appropriate to set thresholds for the determination of a viable intrauterine pregnancy; however, considering the consequences of misdiagnosing a viable intrauterine pregnancy as a miscarriage, the group felt that the thresholds should be conservative. The group noted that in the studies there was a 100% success rate in diagnosing viable intrauterine pregnancies using a CRL threshold of 6 mm or more. However, GDG members also noted that Pexsters et al. (2011) documented the potential for considerable intra- and inter-observer variation in measurements of CRL and mean gestational sac diameter. They also felt that there was additional potential for variation in measurements linked to the quality of scanning equipment and the skill level of the sonographer. In light of all of these considerations, the GDG determined that fetal non-viability should not be diagnosed based on the absence of a heartbeat in fetuses with a CRL of less than 7.0 mm or a mean gestational sac diameter of less than 25.0 mm, as measured on a single transvaginal ultrasound. They felt that, at such small sizes, it would not be possible to determine whether a miscarriage had truly occurred or whether the embryo was simply too small for there to be a visible heartbeat. Therefore, up to and including these thresholds, all women should have a repeat scan to confirm the findings of the initial scan. The GDG discussed what would be an appropriate interval between scans, balancing the fact that sufficient time would need to pass to be able to confirm the diagnosis with the fact that women might understandably want an answer as soon as possible. They agreed that when using a transvaginal ultrasound, it would be appropriate to wait for a minimum of 7 days before repeating the scan when measuring either the CRL or the mean gestational sac diameter.

Due to the significant consequences of misclassifying a viable pregnancy as a miscarriage, the GDG felt that it was reasonable to recommend that, in the absence of a visible heartbeat above these thresholds (that is, a CRL greater than or equal to 7 mm and/or a mean gestational sac diameter greater than or equal to 25 mm), all sonographers should seek a second opinion before definitively diagnosing a non-viable pregnancy. However, they also realised that some women might instead wish to wait and have a second confirmatory scan at a later date, and felt that this was a reasonable alternative.

Based on their clinical experience, and evidence from another review comparing transvaginal and transabdominal ultrasound for diagnosing ectopic pregnancy (see Section 6.3 below), the GDG members decided that transvaginal ultrasound would generally be the optimal mode of scanning. However, they recognised that, in some circumstances, a transvaginal ultrasound might not be appropriate or acceptable to women, and therefore a transabdominal scan could be offered as an alternative. Given the poorer quality of imaging when using a transabdominal scan, the group agreed that there should be an interval of at least 14 days between repeat transabdominal scans before providing a diagnosis, in order to ensure that any change in the size of the CRL or mean gestational sac diameter was identifiable. The group felt that while there were thresholds where a definitive diagnosis was possible following second opinion when using transvaginal scans, this was not possible with a transabdominal scan. Since there is a greater potential for key features not to be visible on a transabdominal scan, it was felt a second opinion would not be helpful.

Consideration of health benefits and resource uses

The GDG did not feel that recommending a repeat ultrasound scan would add significantly to the case load and resource use, because in practice this often happens anyway. However, it did feel that cost effectiveness should be a component of any research conducted in this area, and therefore incorporated it into a research recommendation. Given the number of women requesting scans in early pregnancy and the associated service and cost implications, the group felt that it was important that research be done in this area, to determine the timing and frequency of ultrasound examinations that would maximise improvements in diagnosis, outcomes and women's experience.

Quality of evidence

The quality of evidence ranged from high to very low quality and the GDG members felt that, in conjunction with their clinical experience, it was appropriate to make recommendations based on the findings.

Information giving and emotional support

The GDG recognised that throughout all episodes of care it is important that women be given evidence-based information about the risks, benefits and limitations of investigations being offered as well as determining what the woman expects from the investigation in order to ensure any misunderstandings can be clarified. They cross-referred here to NICE clinical guidance 138 Patient experience in adult NHS services.

From their own experience, and data about the risks of expectant management reported in another review question (see Section 6.3), the GDG members felt that there would be minimal risk in recommending that some women wait a week for a repeat scan, particularly when balanced against the consequences of accidentally terminating a viable pregnancy after misdiagnosis. However, they felt that it was important that women were informed about what to expect in the intervening week and what symptoms should prompt them to seek medical attention. Given this, they also recommended that women should be provided with a 24-hour contact telephone number. The GDG was of the opinion that, due to the fact that miscarriage is a potentially traumatic experience, it was important that women be able to contact someone who would be able to give them accurate information and appropriate support. Therefore, it recommended that this telephone number should allow women to speak directly with someone with experience of dealing with early pregnancy complications and should not simply be a non-specific service like an emergency department.

6.3. Accuracy of imaging techniques for diagnosis of an ectopic pregnancy

Review question

What is the accuracy of transvaginal ultrasound compared with transabdominal ultrasound for diagnosing ectopic pregnancy?

Introduction

The estimated prevalence of ectopic pregnancy is 1–2% worldwide, with nearly 12,000 ectopic pregnancies diagnosed each year in the UK. Associated costs are high due to repeated diagnostic tests, delayed diagnosis and its treatment (Jurkovic & Wilkinson, 2011). Critical evaluation of patient symptoms and signs remains important for the detection of ectopic pregnancy; however, ultrasound scanning remains the cornerstone of clinical diagnosis. Transabdominal ultrasound has, in recent years been largely replaced by transvaginal ultrasound imaging. It is important to note that the diagnostic accuracy of both ultrasound methods is experience-based and allied to constant vigilance for the potential presence of ectopic pregnancy. The GDG therefore considered the evidence for application of the two different methods in order to determine which should be used.

Description of included studies

Five studies were included in this review. Three studies were conducted in the USA (Kivikoski et al., 1990; Shapiro et al., 1988; Thorsen et al., 1990), one in Austria (Schurz et al., 1990) and one in Finland (Cacciatore, 1989). Four prospective studies (Cacciatore, 1989; Kivikoski et al., 1990; Shapiro et al., 1988; Thorsen et al., 1990) compared the diagnostic accuracy of transvaginal and transabdominal ultrasound in women with suspected ectopic pregnancy. One prospective study (Schurz et al., 1990) evaluated reliability and advantages of transabdominal and transvaginal ultrasound compared with clinical signs for detection of ectopic pregnancy.

The reference standard for diagnosis of ectopic pregnancy was reported as laparoscopy in one study (Schurz et al., 1990), surgery (the type of which was not specified) in three studies (Cacciatore, 1989; Shapiro et al., 1988; Thorsen et al., 1990) and either laparoscopy or surgery in one study (Kivikoski et al., 1990).

In one prospective study (Kivikoski et al., 1990) women were first seen at 4–12 weeks amenorrhea at the time of evaluation. Four prospective studies (Cacciatore, 1989; Schurz et al., 1990; Shapiro et al., 1988; Thorsen et al., 1990) did not report women's gestation at the time of ultrasound evaluation.

Evidence profile

The evidence is presented below in one profile. Diagnostic accuracy for ectopic pregnancy was measured using transvaginal and transabdominal ultrasound.

Table 6.3

GRADE summary of findings for accuracy of diagnosing ectopic pregnancy using transvaginal or transabdominal ultrasound.

Two very low quality studies also evaluated the diagnostic accuracy of transvaginal and transabdominal ultrasound for ectopic pregnancy in women with clinical suspicion of ectopic pregnancy. However, there was not adequate information reported in the studies to calculate diagnostic accuracy measures.

Table 6.4

Additional diagnostic accuracy findings for transvaginal and transabdominal ultrasound from two observational studies.

Evidence statements

In the evidence statements the following definitions have been used when summarising the levels of sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV):

High – 90% and above
Moderate – 75% to 89%
Low – below 75%

The following terms have been used when summarising likelihood ratios:

Positive likelihood ratio:
- Very useful – more than 10
- Moderately useful – 5 to 10
- Not useful – less than 5
Negative likelihood ratio:
- Very useful – 0 to 0.1
- Moderately useful – more than 0.1 up to 0.5
- Not useful – more than 0.5

Evidence was identified for transvaginal and transabdominal ultrasounds to determine diagnostic accuracy for ectopic pregnancy in women with clinical suspicion of ectopic pregnancy. The quality of the evidence was moderate, low and very low for the included studies.

Transvaginal ultrasound

Three studies evaluated the diagnostic accuracy of transvaginal ultrasound for ectopic pregnancy in women with suspected ectopic pregnancy. One very low quality study reported a high sensitivity, a low specificity, a high PPV, a low NPV, a not useful positive likelihood ratio and a moderately useful negative likelihood ratio. One moderate quality study reported low sensitivity, high specificity, a high PPV, a moderate NPV, a very useful positive likelihood ratio and a not useful negative likelihood ratio. One low quality study reported low sensitivity, high specificity, a high PPV, a low NPV, a very useful positive likelihood ratio and a moderately useful negative likelihood ratio.

Transabdominal ultrasound

The same three studies evaluated the diagnostic accuracy of transabdominal ultrasound for ectopic pregnancy in women with suspected ectopic pregnancy. One study of very low quality reported a low sensitivity with no information provided about other diagnostic accuracy measures. One moderate quality study reported low sensitivity, high specificity, a high PPV, a low NPV, a very useful positive likelihood ratio and a not useful negative likelihood ratio. One low quality study reported a low sensitivity, a high specificity, a high PPV, a low NPV, a very useful positive likelihood ratio and a not useful negative likelihood ratio.

Diagnosis of ectopic pregnancy

Two very low quality studies evaluated the diagnostic accuracy of transvaginal and transabdominal ultrasound for ectopic pregnancy in women with clinical suspicion of ectopic pregnancy. In one low quality study (Cacciatore, 1989) there was no statistically significant difference between the transvaginal and transabdominal ultrasounds in detection of an adnexal mass and gestational sac. More ectopic fetuses, ectopic sacs, un-ruptured ectopic pregnancies and yolk sacs or viable fetuses were detected by transvaginal ultrasound when compared with transabdominal ultrasound.

In the other very low quality study (Schurz et al., 1990) reliability and advantages of transabdominal and transvaginal ultrasound were compared with clinical signs for detection of ectopic pregnancy in two populations. Clinical findings were more likely to lead to a correct diagnosis of ectopic pregnancy than the findings obtained from transabdominal ultrasound. However; clinical findings were less likely to lead to a correct diagnosis of ectopic pregnancy than the findings obtained by transvaginal ultrasound.

Evidence to recommendations

Relative value placed on the outcomes considered

The GDG recognised the importance of both sensitivity and specificity when diagnosing ectopic pregnancies. Given the risks of failing to identify women with an ectopic pregnancy, it is important that the test be sensitive. However, the group also recognised that there is a danger with a false positive diagnosis, as this could potentially lead to women receiving treatment when they actually have a viable intrauterine pregnancy.

Consideration of clinical benefits and harms

The evidence showed mixed results for both ultrasound methods. Both methods of ultrasound displayed a high specificity (most studies reported findings of 100%). However, each study that looked at transvaginal and transabdominal ultrasound showed that the transvaginal method had a higher sensitivity. As a result, the GDG agreed that transvaginal ultrasound should be the preferred approach but that women's views should be taken into account and accommodated where possible. The GDG members also identified from their clinical experience that there might be occasions when transabdominal ultrasound would be the better option, such as when women have an enlarged uterus or other pelvic pathology.

Consideration of health benefits and resource uses

The group did not feel that there was a difference in cost between the two different methods as the time taken to perform them both would be the same. It was also recognised that some personnel performing ultrasound scans tend to perform an abdominal ultrasound prior to a transvaginal ultrasound. It was the view of the group that unless there were particular indications for performing two scans in this way this practice was not necessary, given the additional resource implications and the superior accuracy of transvaginal scans.

Quality of evidence

The evidence varied in quality from very low to moderate. The GDG was disappointed that there was no recent evidence to address this review question. The members agreed that although scanning technology had improved from the time the studies were conducted, it was likely that both technologies would have improved similarly, and thus that comparisons of the two techniques were still valid. Overall, they felt that taking into account the evidence reviewed and the group's clinical experience, along with the experience of two expert advisers, it was possible to make recommendations.

Information giving and emotional support

The GDG recognised that, while the method of performing an ultrasound scan can make a difference, it is also extremely important that the scanning is undertaken by someone with specific training and experience in identifying ectopic pregnancy. Not only can this can make a large difference to the validity of the diagnosis, it can also impact the experience of the woman undergoing the scan. From their clinical experience, the GDG members felt that having a scan performed by a practitioner without appropriate experience would be likely to make a woman feel anxious and uncertain about her prognosis. In contrast, having a scan performed by a healthcare professional with experience in diagnosing ectopic pregnancies would help to ensure that women felt informed and supported.

Recommendations

Number	Recommendation
21	Offer women who attend an early pregnancy assessment service (or out-of-hours gynaecology service if the early pregnancy assessment service is not available) a transvaginal ultrasound scan to identify the location of the pregnancy and whether there is a fetal pole and heartbeat.
22	Consider a transabdominal ultrasound scan for women with an enlarged uterus or other pelvic pathology, such as fibroids or an ovarian cyst.
23	If a transvaginal ultrasound scan is unacceptable to the woman, offer a transabdominal ultrasound scan and explain the limitations of this method of scanning.
24	Inform women that the diagnosis of miscarriage using 1 ultrasound scan cannot be guaranteed to be 100% accurate and there is a small chance that the diagnosis may be incorrect, particularly at very early gestational ages.
25	When performing an ultrasound scan to determine the viability of an intrauterine pregnancy, first look to identify a fetal heartbeat. If there is no visible heartbeat but there is a visible fetal pole, measure the crown–rump length. Only measure the mean gestational sac diameter if the fetal pole is not visible.
26	If the crown–rump length is less than 7.0 mm with a transvaginal ultrasound scan and there is no visible heartbeat, perform a second scan a minimum of 7 days after the first before making a diagnosis. Further scans may be needed before a diagnosis can be made.
27	If the crown–rump length is 7.0 mm or more with a transvaginal ultrasound scan and there is no visible heartbeat: seek a second opinion on the viability of the pregnancy and/or perform a second scan a minimum of 7 days after the first before making a diagnosis.
28	If there is no visible heartbeat when the crown–rump length is measured using a transabdominal ultrasound scan: record the size of the crown–rump length and perform a second scan a minimum of 14 days after the first before making a diagnosis.
29	If the mean gestational sac diameter is less than 25.0 mm with a transvaginal ultrasound scan and there is no visible fetal pole, perform a second scan a minimum of 7 days after the first before making a diagnosis. Further scans may be needed before a diagnosis can be made.
30	If the mean gestational sac diameter is 25.0 mm or more using a transvaginal ultrasound scan and there is no visible fetal pole: seek a second opinion on the viability of the pregnancy and/or perform a second scan a minimum of 7 days after the first before making a diagnosis.
31	If there is no visible fetal pole and the mean gestational sac diameter is measured using a transabdominal ultrasound scan: record the size of the mean gestational sac diameter and perform a second scan a minimum of 14 days after the first before making a diagnosis.
32	Do not use gestational age from the last menstrual period alone to determine whether a fetal heartbeat should be visible.
33	Inform women that the date of their last menstrual period may not give an accurate representation of gestational age because of variability in the menstrual cycle.
34	Inform women what to expect while waiting for a repeat scan and that waiting for a repeat scan has no detrimental effects on the outcome of the pregnancy.
35	Give women a 24-hour contact telephone number so that they can speak to someone with experience of caring for women with early pregnancy complications who understands their needs and can advise on appropriate care.^*
36	When diagnosing complete miscarriage on an ultrasound scan, in the absence of a previous scan confirming an intrauterine pregnancy, always be aware of the possibility of ectopic pregnancy. Advise these women to return for further review if their symptoms persist.
37	All ultrasound scans should be performed and reviewed by someone with training in, and experience of, diagnosing ectopic pregnancies.

*: See also recommendation 3 for details of further information that should be provided.

Number	Research recommendation
RR 4	How does the timing and frequency of ultrasound examination affect diagnosis and outcomes of early pregnancy complications, including women's experience and cost effectiveness?
	Why this is important
	The rationale behind the frequency of ultrasound to improve diagnosis and outcomes of early pregnancy complications addresses the problems associated with pregnancy of unknown location and intrauterine pregnancy of uncertain viability. The evidence base for the timing and frequency of scanning in early pregnancy is limited, and the number of scans is organised by individual units according to capacity and demand. Some healthcare professionals choose to wait 5 days between scans whereas others will wait 10 to 14 days. These decisions are driven by resource availability as well as clinical considerations, but in particular the effect of different strategies on cost and women's experience is not clear. The literature suggests that there is no clear consensus, but there is general agreement that by 14 days a diagnosis will be clear. To establish the most appropriate time for scans, the efficacy of scans taken after 14 days could be compared with scans taken after 7 days for diagnosis of ectopic pregnancy or viability.

6.4. Diagnostic accuracy of two or more human chorionic gonadotrophin (hCG) measurements for ectopic pregnancy

Review question

What is the diagnostic accuracy of two or more hCG measurements for determining an ectopic pregnancy in women with pain and bleeding and pregnancy of unknown location?

Introduction

The diagnosis of ectopic pregnancy remains problematic in a significant number of cases. Women may present with a positive urine or blood pregnancy test but no visible evidence of the location of the pregnancy on an ultrasound scan (pregnancy of unknown location [PUL]). In these circumstances there is often a tension between not missing an ectopic pregnancy and not subjecting women with an early intrauterine pregnancy to a battery of expensive and potentially unnecessary tests or interventions. One possible alternative is to use serial measurements of human chorionic gonadotrophin (hCG) as a diagnostic tool to try to identify women who might have an ectopic pregnancy and those who are likely to have an early viable intrauterine pregnancy. The following reviews evaluate the diagnostic accuracy of this approach, and consider whether there is added value in the use of a single progesterone measurement.

Description of included studies

Ten studies were included in this review (Barnhart et al., 2010; Condous et al., 2004; Condous et al., 2007; Dart et al., 1999; Daus et al., 1989; Hahlin et al., 1991; Mol et al., 1998; Morse et al., 2012; Stewart et al., 1995; Thorburn et al., 1992).

The included papers consist of five prospective cohort studies (Condous et al., 2004; Condous et al., 2007; Hahlin et al., 1991; Mol et al., 1998; Thorburn et al., 1992) and five retrospective cohort studies (Barnhart et al., 2010; Dart et al., 1999; Daus et al., 1989; Morse et al., 2012; Stewart et al., 1995). One study (Barnhart et al., 2010) was conducted in both the UK and USA; the other studies were conducted in the UK (Condous et al., 2004; Condous et al., 2007), the USA (Dart et al., 1999; Daus et al., 1989; Morse et al., 2012; Stewart et al., 1995), the Netherlands (Mol et al., 1998) and Sweden (Hahlin et al., 1991; Thorburn et al., 1992).

The study participants were women with pain and bleeding in the first trimester of pregnancy who had been classified as having a pregnancy of unknown location on ultrasound. All the studies used two or more serum hCG measurements for diagnosis, and either reported measures of diagnostic accuracy or presented data that allowed calculations to be performed by the technical team. Four studies evaluated the percentage change in hCG over 48 hours (Dart et al., 1999; Daus et al., 1989; Mol et al., 1998; Morse et al., 2012), one study evaluated the rate of change of log hCG using two different thresholds (Stewart et al., 1995) and two studies evaluated the diagnostic accuracy of an abnormal hCG score (Hahlin et al., 1991; Thorburn et al., 1992). A further two papers (Condous et al., 2004; Condous et al., 2007) evaluated the diagnostic accuracy of two different predictive models: M1 and M4. Model M1 incorporated the hCG ratio (ratio of concentration at 48 hours to concentration at 0 hours) and its performance was evaluated using three different sets of parameters. Firstly, the model was evaluated using a probability threshold to distinguish between ectopic pregnancies and non-ectopic pregnancies. Then, the authors incorporated different statistical costs for misclassifying outcomes, using costs of 1 for misclassifying failing pregnancies of unknown location and intrauterine pregnancies, and a cost of either 4 or 5 for misclassifying an ectopic pregnancy. The different costs represent the relative importance of making different types of prediction errors, in this case reflecting that misclassifying an ectopic pregnancy may have more serious consequences than misclassifying other conditions. Model M4 incorporated the average hCG concentration (from measurements at 0 and 48 hours), the ratio of the two hCG measurements and the quadratic effect of the hCG ratio. M4 was only evaluated using the optimal costs of 1, 1 and 4 for misclassifying a failing pregnancy of unknown location, an intrauterine pregnancy and an ectopic pregnancy, respectively.

The reference standards used for the confirmation of the diagnosis of ectopic pregnancy varied between the studies. Laparoscopy was used as the reference standard in four studies (Condous et al., 2004; Hahlin et al., 1991; Mol et al., 1998; Thorburn et al., 1992) and two further studies reported that the diagnosis was confirmed using tissue diagnosis after surgery, the type of which was not specified (Daus et al., 1989; Stewart et al., 1995). Three studies used different reference standards (laparoscopy, ultrasound visualisation or serial serum hCG measurements combined with no evidence of chorionic villi after dilation and evacuation) according to how the ectopic pregnancy was managed (Barnhart et al., 2010; Condous et al., 2007; Dart et al., 1999). In one study it is reported that the ectopic pregnancies included visualised and non-visualised pregnancies but no further details are given (Morse et al., 2012). Further details of how diagnoses were confirmed, including ultrasound criteria, can be found in the evidence tables (see Appendix H).

Evidence profile

Table 6.5

GRADE summary of findings for the diagnosis of ectopic pregnancy using two or more hCG measurements.

Evidence statements

The following definitions have been used when summarising the levels of sensitivity, specificity, PPV and NPV:

High – 90% and above
Moderate – 75% to 89.9%
Low – 75% and below

The following terms have been used when summarising the likelihood ratios:

Positive likelihood ratio:
- Very useful – more than 10
- Moderately useful – 5 to 10
- Not useful – less than 5
Negative likelihood ratio:
- Very useful – 0 to 0.1
- Moderately useful – more than 0.1 up to 0.5
- Not useful – more than 0.5

Percentage change in serum hCG concentration in 48 hours

One study evaluated the use of a decline or a rise to less than 50% in serum hCG over 48 hours for the diagnosis of ectopic pregnancy. The study reported a low sensitivity, low specificity, low PPV, low NPV, not useful positive likelihood ratio and not useful negative likelihood ratio. The evidence for this finding was of low quality.

One study evaluated the diagnostic accuracy of a decline or a rise to less than 63% in serum hCG over 48 hours for the diagnosis of ectopic pregnancy. The study reported a high sensitivity, low specificity, low PPV, high NPV, not useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of a decline or a rise to less than 66% in serum hCG over 48 hours for the diagnosis of ectopic pregnancy. The study reported a moderate sensitivity, low specificity, low PPV, moderate NPV, not useful positive likelihood ratio and not useful negative likelihood ratio. The evidence for this finding was of low quality.

One study evaluated the diagnostic accuracy of a change in serum hCG between a decline of 36–47% and a rise of 35%. The study reported a moderate sensitivity, low specificity, low PPV and high NPV. The study did not report likelihood ratios. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of a change in serum hCG between a decline of 36–47% and a rise of 53%. The study reported a high sensitivity, low specificity, low PPV and high NPV. The study did not report likelihood ratios. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of a change in serum hCG between a decline of 36–47% and a rise of 71%. The study reported a high sensitivity, low specificity, low PPV and high NPV. The study did not report likelihood ratios. The evidence for this finding was of very low quality.

Rate of change of log hCG concentration

One study evaluated the diagnostic accuracy of a rate of change of log hCG of less than 0.11 for the diagnosis of ectopic pregnancy. The study reported a moderate sensitivity, low specificity, low PPV, moderate NPV, not useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of a rate of change of log hCG of less than 0.14 for the diagnosis of ectopic pregnancy. The study reported a high sensitivity, low specificity, low PPV, high NPV, not useful positive likelihood ratio and useful negative likelihood ratio. The evidence for this finding was of very low quality.

hCG score

Two studies evaluated the diagnostic accuracy of an abnormal hCG score for the diagnosis of ectopic pregnancy. They both reported a moderate sensitivity, low specificity, low PPV, moderate NPV, not useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for these findings was of low quality.

Models

One study evaluated the diagnostic accuracy of model M1, using a probability threshold, for the diagnosis of ectopic pregnancy. The study reported a moderate sensitivity, moderate specificity, low PPV, high NPV, moderately useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of low quality.

Two studies evaluated the diagnostic accuracy of model M1, incorporating costs of 1, 1 and 4, for the diagnosis of ectopic pregnancy. One study reported a moderate sensitivity, moderate specificity, low PPV, high NPV, moderately useful positive likelihood ratio and moderately useful negative likelihood ratio. The other study reported a low sensitivity, moderate specificity, low PPV, high NPV, moderately useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for both sets of findings was of low quality.

One study evaluated the diagnostic accuracy of model M1, incorporating costs of 1, 1 and 5, for the diagnosis of ectopic pregnancy. The study reported a high sensitivity, moderate specificity, low PPV, high NPV, moderately useful positive likelihood ratio and very useful negative likelihood ratio. The evidence for this finding was of low quality.

Two studies evaluated the diagnostic accuracy of model M4 for the diagnosis of ectopic pregnancy. One study reported a moderate sensitivity, moderate specificity, low PPV, high NPV, moderately useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of low quality. One further study incorporated two populations. In the UK population, the study reported a moderate sensitivity, moderate specificity, low PPV, high NPV, moderately useful positive likelihood ratio and moderately useful negative likelihood ratio. In the adjusted USA population, the study reported a low sensitivity, moderate specificity, low PPV, moderate NPV, not useful PPV and not useful NPV. The evidence for this finding was of very low quality

Evidence to recommendations

Please see Section 6.7 where the evidence from all of the reviews which assess the use of hCG for diagnosis has been considered.

6.5. Diagnostic accuracy of two or more hCG measurements plus progesterone for ectopic pregnancy

Review question

What is the diagnostic accuracy of two or more hCG measurements plus progesterone for determining an ectopic pregnancy in women with pain and bleeding and pregnancy of unknown location?

Description of included studies

Three studies were included in this review (Condous et al., 2004; Gevaert et al., 2006; Hahlin et al., 1991). These were two prospective cohort studies (Condous et al., 2004; Hahlin et al., 1991) and one retrospective study (Gevaert et al., 2006). The studies were conducted in the UK (Condous et al., 2004; Gevaert et al., 2006) and Sweden (Hahlin et al., 1991).

The study participants were women with pain and bleeding in the first trimester of pregnancy who had been classified as having a pregnancy of unknown location on ultrasound. All of the studies used the combination of two or more serum hCG measurements and progesterone for diagnosis, and either reported measures of diagnostic accuracy or presented data that allowed calculations to be performed by the NCC-WCH technical team. One study evaluated the diagnostic accuracy of an abnormal hCG score in conjunction with a progesterone concentration of less than 30 nmol/l (Hahlin et al., 1991), and two studies evaluated the diagnostic accuracy of predictive models (Condous et al., 2004; Gevaert et al., 2006). The Bayesian model (parameter prior model [PPM]) incorporated the hCG ratio (ratio of concentration at 48 hours to concentration at 0 hours), progesterone concentration at 48 hours and the number of gestation days. The model M3 incorporated the hCG ratio, the average progesterone concentration (from measurements at 0 and 48 hours) and maternal age in years.

The reference standard used for the confirmation of the diagnosis of ectopic pregnancy was laparoscopy in two studies (Condous et al., 2004; Hahlin et al., 1991). In the third study, diagnosis was based on ultrasound visualisation and was then confirmed at laparoscopy in women who underwent surgery (Gevaert et al., 2006). Further details, including ultrasound criteria, can be found in the evidence tables (see Appendix H).

Table 6.6

GRADE summary of findings for the diagnosis of ectopic pregnancy using two or more hCG measurements plus progesterone.

Table 6.7

GRADE summary of findings for the diagnosis of ectopic pregnancy using two or more hCG measurements plus progesterone (Model M3).

Evidence statements

hCG score and progesterone concentration

One study evaluated the diagnostic accuracy of an abnormal hCG score in combination with a progesterone concentration of less than 30 nmol/l for the diagnosis of ectopic pregnancy. The study reported a low sensitivity, low specificity, low PPV, low NPV, not useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of low quality.

Models

One study evaluated the diagnostic accuracy of Bayesian model PPM for the diagnosis of ectopic pregnancy. The study reported a moderate sensitivity, moderate specificity, not useful positive likelihood ratio and moderate useful negative likelihood ratio; PPV and NPV were not reported. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of model M3 for the diagnosis of ectopic pregnancy. Only the area under the ROC curve was reported, therefore other diagnostic accuracy measures could not be assessed. The evidence for this finding was of low quality.

Evidence to recommendations

Please see Section 6.7 where the evidence from all of the reviews which assess the use of hCG for diagnosis has been considered.

6.6. Diagnostic accuracy of two or more hCG measurements for viable intrauterine pregnancy

Review question

What is the diagnostic accuracy of two or more hCG measurements for determining a viable intrauterine pregnancy in women with pain and bleeding and pregnancy of unknown location?

Description of included studies

Six studies were included in this review (Dart et al., 1999; Daus et al., 1989; Hahlin et al., 1991; Mol et al., 1998; Morse et al., 2012; Stewart et al., 1995).

The included papers consist of two prospective cohort studies (Hahlin et al., 1991; Mol et al., 1998) and four retrospective cohort studies (Dart et al., 1999; Daus et al., 1989; Morse et al., 2012; Stewart et al., 1995). The studies were conducted in the USA (Dart et al., 1999; Daus et al., 1989; Morse et al., 2012; Stewart et al., 1995), the Netherlands (Mol et al., 1998) and Sweden (Hahlin et al., 1991).

The reference standard was a repeat ultrasound scan in three studies (Hahlin et al., 1991; Mol et al., 1998; Morse et al., 2012) and either a repeat ultrasound scan or the birth of the baby in two studies (Dart et al., 1999; Stewart et al., 1995). Further details, including the time of the repeat ultrasound scan, can be found in the evidence tables (see appendix H). In one study details of how the viable intrauterine pregnancies were confirmed are not reported (Daus et al., 1989).

Evidence profile

Table 6.8

GRADE summary of findings for the diagnosis of viable intrauterine pregnancy using two or more hCG measurements.

Evidence statements

Percentage change in serum hCG concentration in 48 hours

One study evaluated the diagnostic accuracy of a rise of more than 35% in serum hCG over 48 hours for the diagnosis of viable intrauterine pregnancy. The study reported a high sensitivity, high specificity, moderate PPV and high NPV. The study did not report likelihood ratios. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of a rise of more than 50% in serum hCG over 48 hours for the diagnosis of viable intrauterine pregnancy. The study reported a high sensitivity, high specificity, low PPV, high NPV, very useful positive likelihood ratio and very useful negative likelihood ratio. The evidence for this finding was of low quality.

One study evaluated the diagnostic accuracy of a rise of more than 53% in serum hCG over 48 hours for the diagnosis of viable intrauterine pregnancy. The study reported a moderate sensitivity, high specificity, high PPV and high NPV. The study did not report likelihood ratios. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of a rise of more than 63% in serum hCG over 48 hours for the diagnosis of viable intrauterine pregnancy. The study reported a moderate sensitivity, high specificity, high PPV, low NPV, very useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of very low quality.

One study evaluated the diagnostic accuracy of a rise of more than 66% in serum hCG over 48 hours for the diagnosis of viable intrauterine pregnancy. The study reported a moderate sensitivity, high specificity, moderate PPV, high NPV, very useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of low quality.

One study evaluated the diagnostic accuracy of a rise of more than 71% in serum hCG over 48 hours for the diagnosis of viable intrauterine pregnancy. The study reported a low sensitivity, high specificity, high PPV and high NPV. The study did not report likelihood ratios. The evidence for this finding was of very low quality.

Rate of change of log hCG concentration

One study evaluated the diagnostic accuracy of a rate of change of log hCG of more than 0.11 for the diagnosis of viable intrauterine pregnancy. The study reported a moderate sensitivity, moderate specificity, low PPV, high NPV, moderately useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of very low quality

One study evaluated the diagnostic accuracy of a rate of change of log hCG of more than 0.14 for the diagnosis of viable intrauterine pregnancy. The study reported a low sensitivity, high specificity, high PPV, high NPV, very useful positive likelihood ratio and moderately useful negative likelihood ratio. The evidence for this finding was of very low quality.

hCG score

One study evaluated the diagnostic accuracy of a normal hCG score for the diagnosis of viable intrauterine pregnancy. The study reported a high sensitivity, high specificity, moderate PPV, high NPV, very useful positive likelihood ratio and very useful negative likelihood ratio. The evidence for this finding was of low quality.

Evidence to recommendations

Please see Section 6.7 where the evidence from all of the reviews which assess the use of hCG for diagnosis has been considered.

6.7. Diagnostic accuracy of two or more hCG measurements plus progesterone for viable intrauterine pregnancy

Review question

What is the diagnostic accuracy of two or more hCG measurements plus progesterone for determining a viable intrauterine pregnancy in women with pain and bleeding and pregnancy of unknown location?

Description of included studies

One study was included in this review (Hahlin et al., 1991). The included paper is a prospective cohort study, conducted in Sweden. The study used the combination of a normal hCG score and progesterone concentration of more than 30 nmol/l for diagnosis, and included women with pain and bleeding in the first trimester of pregnancy who had been classified as having a pregnancy of unknown location on ultrasound. The reference standard used to confirm a viable intrauterine pregnancy was a transvaginal ultrasound scan in the 8th–10th week of gestation showing normal fetal development, including heart activity.

Evidence profile

Table 6.9

GRADE summary of findings for the diagnosis of viable intrauterine pregnancy using two or more hCG measurements plus progesterone.

Evidence statements

One study evaluated the diagnostic accuracy of a normal hCG score in combination with a progesterone concentration of more than 30 nmol/l for diagnosing viable intrauterine pregnancy. The study reported a high sensitivity, high specificity, moderate positive predictive value, high negative predictive value, very useful positive likelihood ratio and very useful negative likelihood ratio. The evidence for this finding was of low quality.

Evidence to recommendations

Relative value placed on the outcomes considered

The GDG had hoped to identify in these reviews whether the use of two or more hCG measurements (with or without progesterone) was appropriate for diagnosing both an ectopic pregnancy and a viable intrauterine pregnancy. The group was therefore looking for a test which provided both high sensitivity and high specificity. However, while the evidence indicated a number of different ways of documenting the change in hCG levels (measuring percentage change, measuring log change or applying various mathematical models), none of these methods showed the use of hCG to be useful as a test for comprehensively and definitively diagnosing either an ectopic pregnancy or a viable intrauterine pregnancy.

When developing the protocol for this question, the GDG agreed that it was best to focus the search on serial hCG measurements. It was aware that a small number of units use a single hCG plus progesterone measurement but did not think that this was common practice. The GDG felt that it was unlikely that a single measurement would provide sufficient accuracy to be helpful to clinicians. As the primary focus of clinicians at this stage would be to identify those women who are likely to have an ectopic pregnancy, the group agreed that most would not be comfortable making this judgement based on one measurement alone.

Consideration of clinical benefits and harms

The GDG recognised that some of the papers considered mathematical models. While they understood the intrinsic potential value that such models could have, the models considered in the review had not been widely validated nor did they lend themselves to ease of use. Thus it was not felt appropriate to recommend their use.

The evidence showed that the use of the rate of log change hCG or a change in hCG levels of greater than 63% both provided a high negative predictive value but a low positive predictive value. In other words, these measures were effective as tests to rule out the presence of an ectopic pregnancy but not effective as tests to specifically identify the presence of an ectopic pregnancy. Thus the group felt that the value of the tests would be as a risk stratification tool to determine the urgency and type of care that each woman requires, with the most urgent care being focused on those women in whom an ectopic pregnancy was more likely.

The group recognised that both measuring the rate of log change hCG and measuring the rate of change of hCG gave similar diagnostic findings. It was felt that measuring an increase of greater than 63% would be easier to calculate and a more useful measure in clinical practice than calculating the rate of log change. In addition, the findings regarding the greater than 63% change were likely to be more valid as the sample size of the study was much larger. As a result, the group agreed to recommend the use of an hCG increase of greater than 63%. The evidence suggested that approximately 17% women with a PUL would fall into this category and would have a high chance of having a viable intrauterine pregnancy. Daus et al. (1989) reported that 60/357 women with a PUL had an hCG rise of greater than 63%. Of these 54/60 were ultimately diagnosed with a viable intrauterine pregnancy (IUP). Similar findings were reported by Dart et al. (1999) who found that 52/307 of women with a PUL had an hCG rise of greater than 66% and 40/52 of these were ultimately diagnosed with a viable IUP.

The GDG discussed whether clinicians should also take into account absolute hCG levels. While there wasn't specific evidence about this available from the studies, the group agreed that for women with an hCG level under 1500 international units per litre (IU/l), they would feel comfortable waiting 7 to 14 days for a second scan. However, for women with an hCG level over 1500 IU/l it would be prudent for clinicians to consider the possibility of an earlier scan given the increased risk if the pregnancy was ectopic. In the studies, the two hCG measurements were taken at different times, but generally with an interval of around 48 hours. The group agreed that this was a reasonable interval to use and for early pregnancy assessment services (EPASs) to aim to achieve.

The GDG interrogated the evidence to determine an appropriate lower threshold to identify those women likely to have a failing pregnancy and for whom a different management strategy could be used. Of the four papers which evaluated a decline of hCG, one (Daus et al., 1999) used ‘any decline’, one used a decline of 36–47% (Morse et al., 2012) and the other two (Mol et al., 1998; Dart et al., 1999) evaluated a decline of ‘greater than 50%’. These last two studies suggested that 32–36% (63/195 and 109/307 respectively) of women with a PUL would fall into the latter category (that is, a decline of hCG of greater than 50%) and were therefore likely to have a failing pregnancy. In these studies none of these women had a viable intrauterine pregnancy and the risk of an ectopic pregnancy was low (about 1% when both studies were combined). In Morse et al. (2012), which used a threshold for decline of 36–47%, a higher proportion of women (over 3%) were misclassified as having a miscarriage when they were eventually diagnosed with an ectopic pregnancy. From these data and their own clinical experience, the GDG members felt that a decline in hCG levels of greater than 50% was highly likely to indicate a failing pregnancy and that this was an appropriately conservative threshold to use.

The group agreed that for women with a decline in hCG levels of greater than 50% it would not be necessary to conduct a repeat ultrasound scan, but instead that they should be asked to do a urine pregnancy test in two weeks' time. If this was negative and the woman was asymptomatic no further action would be necessary. However, if it were positive then the woman should return to the dedicated early pregnancy service within 24 hours for a further clinical review and individualised management.

For the remaining group of women (those with an hCG level change between a decline of less than or equal to 50% and a rise of less than or equal to 63%) it was agreed that a review in the early pregnancy assessment service within 24 hours would be warranted.

The group discussed the relative urgency of referral which should be associated with each threshold. They agreed that women who have an hCG increase of more than 63% and no confirmed ultrasound diagnosis of an intrauterine pregnancy on a subsequent scan should be referred immediately, given the risk of a rupturing ectopic pregnancy. Women with an hCG decrease of more than 50% and a positive pregnancy test after 14 days and women with an hCG level between the two thresholds should be reviewed within 24 hours. The GDG felt it was important that both of these groups of women be seen promptly, given the chance of there being an ectopic pregnancy. However, the GDG agreed that in these instances, the chance of rupture was reduced and thus an immediate referral was not required.

The group wished to highlight that hCG levels are a measurement of trophoblastic proliferation only and should not be used for a confirmatory diagnosis. Final confirmation can only be provided by either an ultrasound scan or a negative pregnancy test (to identify a failed pregnancy).

The GDG considered the evidence that was available for the use of progesterone levels in conjunction with hCG. It noted that for both diagnosis of ectopic pregnancy and diagnosis of viable intrauterine pregnancy, there was little or no improvement in the negative predictive value of the tests compared with using hCG alone.

The group felt it important to stress the importance of symptoms over hCG. The group agreed that all women, regardless of their hCG level, should be given written information about what to do if they experience any new or worsening symptoms, including details about how to access emergency care.

Consideration of health benefits and resource uses

The GDG felt that the use of large numbers of serum hCG measurements was not an effective use of resources, both in terms of women's care and the cost of the tests. It wanted to avoid large numbers of tests being performed without a diagnosis being made, and agreed that the use of more than two serum hCG measurements should only be undertaken following review by a senior healthcare professional. Considering that there was no evidence of any added value of progesterone in making a diagnosis when combined with hCG measurements, and that performing the test has associated costs, the GDG agreed that progesterone should not be used with serial hCG measurements in the assessment of women with pain and bleeding and a PUL.

Quality of evidence

The quality of the evidence was either low or very low. The GDG was aware of the limitations of the studies but ultimately agreed that there was sufficient evidence to make recommendations.

Information giving and emotional support

The GDG felt that it was important that women be given a realistic likely prognosis for their pregnancy based on their hCG levels, and also that they be informed that further confirmation would be needed. For women whose pregnancies might be unlikely to continue, they thought that this would avoid giving women false hope and ensure that they felt informed about the likely outcome. The group also felt that it was very important that these women be given information about where and how to access support and counselling services.

Other considerations

The GDG discussed the reference standards used for the diagnosis of ectopic pregnancy. Most of the included studies used confirmation of the diagnosis at surgery (generally laparoscopy) as the reference standard. While the GDG members noted that there might be a small risk of misdiagnosis using laparoscopy, their opinion was that this was a very rare event and therefore its use as a reference standard was reasonable. They noted that in some of the studies other reference standards were used if women had their ectopic pregnancy managed expectantly or with methotrexate. They agreed that this might undermine the validity of those studies; however, they concluded that overall the body of evidence remained robust and the findings of the review valid.

Recommendations

Number	Recommendation
38	Be aware that women with a pregnancy of unknown location could have an ectopic pregnancy until the location is determined.
39	Do not use serum hCG measurements to determine the location of the pregnancy.
40	In a woman with a pregnancy of unknown location, place more importance on clinical symptoms than on serum hCG results, and review the woman's condition if any of her symptoms change, regardless of previous results and assessments.
41	Use serum hCG measurements only for assessing trophoblastic proliferation to help to determine subsequent management.
42	Take 2 serum hCG measurements as near as possible to 48 hours apart (but no earlier) to determine subsequent management of a pregnancy of unknown location. Take further measurements only after review by a senior healthcare professional.
43	Regardless of serum hCG levels, give women with a pregnancy of unknown location written information about what to do if they experience any new or worsening symptoms, including details about how to access emergency care 24 hours a day. Advise women to return if there are new symptoms or if existing symptoms worsen.
44	For a woman with an increase in serum hCG concentration greater than 63% after 48 hours: Inform her that she is likely to have a developing intrauterine pregnancy (although the possibility of an ectopic pregnancy cannot be excluded). Offer her a transvaginal ultrasound scan to determine the location of the pregnancy between 7 and 14 days later. Consider an earlier scan for women with a serum hCG level greater than or equal to 1500 IU/litre. If a viable intrauterine pregnancy is confirmed, offer her routine antenatal care^* If a viable intrauterine pregnancy is not confirmed, refer her for immediate clinical review by a senior gynaecologist.
45	For a woman with a decrease in serum hCG concentration greater than 50% after 48 hours: inform her that the pregnancy is unlikely to continue but that this is not confirmed and provide her with oral and written information about where she can access support and counselling services^†and ask her to take a urine pregnancy test 14 days after the second serum hCG test, and explain that: if the test is negative, no further action is necessary if the test is positive, she should return to the early pregnancy assessment service for clinical review within 24 hours.
46	For a woman with a change in serum hCG concentration between a 50% decline and 63% rise inclusive, refer her for clinical review in the early pregnancy assessment service within 24 hours.
47	For women with a pregnancy of unknown location, when using serial serum hCG measurements, do not use serum progesterone measurements as an adjunct to diagnose either viable intrauterine pregnancy or ectopic pregnancy.

*: See Antenatal care (NICE clinical guideline 62)
†: See recommendation 3 for details of further information that should be provided

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

Bookshelf ID: NBK132768

Contents

< Prev Next >

PubReader
Print View
Cite this Page
National Collaborating Centre for Women's and Children's Health (UK). Ectopic Pregnancy and Miscarriage: Diagnosis and Initial Management in Early Pregnancy of Ectopic Pregnancy and Miscarriage. London: RCOG Press; 2012 Dec. (NICE Clinical Guidelines, No. 154.) 6, Diagnosis of ectopic pregnancy and miscarriage.
PDF version of this title (3.1M)

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on