Chapter 6 Complications of early pregnancy

Bleeding and pain in early pregnancy

Hyperemesis gravidarum

Pregnancy of unknown location Bleeding and pain in early pregnancy Epidemiology

Vaginal bleeding occurs in about 25% of pregnancies with a positive pregnancy test within the first trimester.

The causes of vaginal bleeding in early pregnancy include:

• miscarriage (threatened, spontaneous, missed, complete, incomplete, or septic)

• ectopic pregnancy

• gestational trophoblastic disease

• gynaecological causes of bleeding such as ectropion

• tumours of the reproductive tract (rare). Prognosis

In pregnancies complicated by bleeding in the first trimester, 50% will progress beyond 20 weeks, 30% will miscarry, 10% will be an ectopic pregnancy, and 0.2% will be a hydatidiform mole; about 5% will elect to terminate the pregnancy. Clinical approach

The first step in management is to assess the blood loss and the need for resuscitation. If severe blood loss is seen or suspected, resuscitative measures must begin immediately.

If the patient is clinically stable, diagnostic measures can be undertaken. Diagnosis

History

• Menstrual history with an accurate last menstrual period where known

• Amount of vaginal bleeding

• Associated backache or abdominal pain: central or lateral

• Shoulder-tip pain (may occur due to diaphragmatic irritation in intra-abdominal bleeding)

• Previous obstetric history.

Examination

• Vital signs, especially pulse rate and blood pressure

• Abdominal distension and tenderness (distension occurs in intra-abdominal bleeding, unilateral tenderness may denote an ectopic pregnancy)

• Speculum examination may help establish the cause of bleeding. A bleeding ectropion may be visible.

• Bimanual pelvic examination may help to identify an enlarged uterus consistent with an intrauterine pregnancy or reveal a tender adnexal mass in an ectopic pregnancy. Investigations

Urinary pregnancy test

• Serial β-human chorionic gonadotrophin (hCG). A doubling rate of less than or equal to 48 hours is suggestive of a normal intrauterine pregnancy. Falling levels denote a failing pregnancy. A suboptimal rise occurs in ectopic pregnancy. Unduly high levels occur with trophoblastic disease.

Ultrasound scan

• With a transvaginal scan, an intrauterine gestational sac may be seen as early as 4 weeks. Presence of a yolk sac from 5 weeks is suggestive of an intrauterine pregnancy. A viable intrauterine pregnancy may be visualized from 5 to 6 weeks’ gestation. The presence of free fluid with an adnexal mass and empty uterus are associated with an ectopic pregnancy. A snowstorm appearance in the uterus occurs with a hydatidiform mole.

• Serum progesterone levels are also used in some units. Counselling and management

• In the presence of a viable, intrauterine gestation, the patient is reassured and advised to return if there is excessive bleeding.

• In the presence of a non-viable intrauterine pregnancy, the patient is given the options of conservative, medical, or surgical management of miscarriage considering the clinical conditions.

• A diagnostic laparoscopy may need to be considered if an ectopic pregnancy is suspected with salpingectomy if necessary.

• Ongoing counselling is very important at every step, as it is often an emotionally difficult situation for the parents.

• Use of anti-D should be considered if the patient’s blood group is Rhesus negative.

• The parents must be reassured that most subsequent pregnancies will have a normal outcome.

It is important to remember that isolated abdominal pain may be due to an ectopic pregnancy in any woman of reproductive age.

Royal College of Obstetricians and Gynaecologists (RCOG). The management of early pregnancy loss. Guideline No. 25. London: RCOG 2006.

Royal College of Obstetricians and Gynaecologists. The management of tubal pregnancy. Guideline No. 21. London: RCOG 2004. Patient information and contacts

Royal College of Obstetricians and Gynaecologists. Patient information leaflet—Bleeding and pain in early pregnancy: information for you. London: RCOG 2008.

Hyperemesis gravidarum Definition

Nausea and vomiting in pregnancy typically in the first trimester resulting in dehydration and ketonuria severe enough to justify hospital admission and require intravenous fluid therapy, after exclusion of other causes of vomiting (Verberg 2005). Epidemiology

Nausea or vomiting in early pregnancy affects up to 80% of women, but hyperemesis gravidarum (HG) affects only 0.3–1.5%. Typically, symptoms start between 5 and 10 weeks of gestation and resolve by 20 weeks, although up to 10% of women will continue to vomit throughout the pregnancy. The rate of admission for HG falls from 8 weeks onwards (Fell 2006).

Associations

• Ethnicity: higher incidence in Asian women in UK, Asians and Africans in USA, and Pacific Islanders in New Zealand.

• Body mass index: BMI below the normal range (and in some studies above the normal range) is associated with HG.

• Younger maternal age.

• Nulliparity.

• Female fetal sex.

• Multiple pregnancy.

• Family history of hyperemesis.

• Gestational trophoblastic disease (GTD): a recent study suggests this may not now be the case as GTD tends to be diagnosed earlier due to the prevalence of early pregnancy transvaginal ultrasound (Kirk 2007).

• Coexistent medical disorders (hyperthyroidism, psychiatric illness, previous molar pregnancy, pre-existing DM, gastrointestinal disorders, and asthma), though these disorders themselves may be causal factors for vomiting.

• Reduced incidence with maternal smoking.

• Psychological stress or ambivalence about pregnancy (may be a cause or effect of HG). Aetiology

The aetiology of HG is unknown and is likely to be complex and multifactorial.

Anatomical

Excessive fluid secretion causing distension of upper gastrointestinal tract and vomiting.

Infective

Several studies confirm a higher incidence of Helicobacter pylori in HG patients than in normal pregnant controls (Golberg 2007) with the density of colonization correlated with the severity of symptoms. This may be due to a change in the gastric pH or an altered immune system during pregnancy. Alternatively, vomiting itself may cause damage to the oesophagus, facilitating colonization of the mucosa. There is no randomized controlled trial of H. pylori treatment to confirm treatment benefit.

Hormonal

• Human chorionic gonadotrophin (hCG) is often thought to be a causal factor and several studies show higher hCG levels in women with HG. This is consistent with the peak incidence of HG occurring when the corpus luteum is most active in producing hCG and with women with multiple pregnancies and molar pregnancies having a reported increased incidence of HG, but does not explain why some women have prolonged vomiting throughout pregnancy when hCG levels have declined. In addition, women with excess hCG levels (e.g. choriocarcinoma or exogenous hCG in in vitro fertilization (IVF) treatment) do not generally suffer HG. There may be different isoform patterns of hCG in different women, or individual variation in responsiveness to hCG.

• Lower, higher, and similar levels of progesterone have been found in women with HG than in those without HG. Progesterone does not help symptoms, and women with exogenous progesterone support in IVF or with multiple corpora lutea do not have a high incidence of HG; progesterone is unlikely to be a true causative factor.

• High levels of oestrogen have been found in some HG studies and may have a role in slowing the intestinal transit time and gastric emptying, altering gastric pH and causing increased H.

• Leptin, cortisol, adrenocorticotropic hormone (ACTH), growth hormone, and prolactin have been postulated to have causative roles but this is not strongly supported in the literature. In particular some studies suggest lower cortisol and ACTH levels in HG but this is not supported by a clear value in corticosteroid treatment.

• Vitamin B6 (pyridoxine) deficiency has been reported but no benefit has been demonstrated with replacement therapy, and the deficiency is likely to be an effect (rather than cause) of HG due to increased demands, poor absorption, and decreased intake.

Immunological

Altered immunological factors have been found in women with HG but it is unclear whether these associations are causal or a response to HG.

HG and thyroid disease

• Pregnancy induces physiological alterations in thyroid function (including changes in iodine metabolism, serum thyroid binding proteins and maternal goitre development) that should be taken into account in the interpretation of thyroid funtion test (TFT) results. Physiological stimulation of the thyroid gland is common in early pregnancy. This is usually assumed to relate to the structural similarity between β-hCG and thyroid stimulating hormone (TSH), causing cross-reactivity and excessive stimulation of the thyroid. This association is supported by the mirroring of hCG rise and TSH fall in early pregnancy and an increased incidence of thyroid overactivity in women with molar or multiple pregnancies. The TSH level has also been shown to correlate with HG symptom severity and electrolyte levels.

• If thyroid hormone concentrations fall outside the normal range (high T4, low TSH) this is then termed gestational transient thyrotoxicosis (GTT). In contrast to Grave’s disease, GTT is not associated with adverse fetal outcome, and without treatment T4 usually returns to normal levels by 15 weeks and TSH by 19 weeks (Tan 2002).

There is a 15% recurrence rate of HG in future pregnancies (compared with 0.7% if there is no previous HG) and it may be of increased severity in subsequent pregnancies. Clinical approach

Diagnosis

History

The history should include protracted vomiting. This can be assessed using the validated pregnancy-unique quantification of emesis and nausea (PUQE) score (Koren 2002), which quantifies symptoms over the preceding 12 hours. The history should also aim to exclude symptoms of alternative causes of vomiting such as urinary tract infection, gastrointestinal infection (e.g. diarrhoea), or preceding morbidity (e.g. diabetes). Epigastric pain and haematemesis should be specifically enquired about.

Examinations

Examination of pulse and blood pressure as well as assessment of dehydration from mucous membranes and skin turgour. Abdominal examination should include epigastric palpation, assessment for organomegaly, and lower abdominal examination.

Investigations

• Urinalysis (at least 1+ ketonuria for the diagnosis to be confirmed)

• MSU if positive urinalysis for nitrites or leucocytes.

• FBC to assess for raised haematocrit associated with dehydration as well as anaemia.

• Hypokalaemia is typical with hyponatraemia and raised urea and creatinine in severe cases.

• Elevated liver function tests (LFT) (commonly alanine transaminase or aspartate transaminase) occur in up to 67% of cases, probably from a combination of dehydration, malnutrition, and lactic acidosis. Usually resolves quickly with rehydration and establishment of oral feeding.

• Liver ultrasound and hepatitis screen should be reserved for women with significantly abnormal LFT or in whom liver function does not resolve rapidly with treatment of HG.

• Raised amylase has been reported with hyperemesis due to increased secretions of saliva rather than excessive pancreatic amylase production.

Ultrasound

Although classically twin and molar pregnancies are more common in HG, a recent study suggests that there is no increased prevalence in HG and the early pregnancy loss rate is significantly lower in the presence of HG. However, ultrasound is recommended for maternal reassurance.

Differential diagnosis

Urinary tract infection, gastrointestinal disease, hepatitis, pancreatitis, and Grave’s disease.

Management

General points

• Maternal weight, ketonuria, and serum electrolytes should be assessed daily during admission.

• Dextrose-containing solutions should be avoided as they increase the body’s requirements for thiamine and thus the chance of precipitating Wernicke’s encephalopathy.

• Potassium supplementation should be given initially with the fluid replacement and then tailored to the serum potassium concentration.

• There is good evidence of benefit from antihistamine antiemetics and little evidence of adverse outcomes. 5-Hydroxytryptamine 3 (5HT3) antagonists have a less clear safety record and thus should not be used unless other treatments fail, and after consultation with the woman.

• The benefit of i.v. corticosteroid treatment remains equivocal (Nelson-Piercy 2001). Oral prednisolone alone has not been shown to be effective.

• Serotonin inhibitors have been described for hyperemesis but are not routine.

• Ginger (Zingiber officinale) 500–1500 mg orally in divided doses has been shown to be effective in reducing nausea and vomiting in four randomized controlled trials.

• Pryidoxine (B6) also has some effect at reducing symptoms.

• There is equivocal evidence of benefit from acupressure at the P6 point (wrist).

• Thiamine replacement is indicated once vomiting has been occurring for at least 3 weeks (the minimum time shown for thiamine stores to be depleted), to prevent development of Wernicke’s encephalopathy.

• Treatment with antithyroid drugs or beta-adrenergic blockers is only indicated if clinical and biochemical features of hyperthyroidism are apparent.

Treatment approach

Treatment is supportive and aimed at managing the symptoms of nausea and vomiting, correcting dehydration and electrolyte imbalance, and at preventing complications such as Wernicke’s encephalopathy. Admission to hospital has been recommended for any woman who is ketotic and is unable to maintain adequate hydration, but outpatient management with i.v. fluid and antiemetics is increasingly used. A suitable treatment regime is described below.

Fluids

Normal saline: 1 L with 20 mmol potassium chloride over 2 hours, 1 L with 20 mmol potassium chloride over 4 hours, 1 L over 6 hours, 1 L over 8 hours, followed by 1 L every 8 hours as maintenance regime, with appropriate electrolyte (potassium) replacement according to serum potassium result

Antiemetic

• First line: cyclizine tds i.v. until oral tolerated. An alternative (suitable for outpatient management) is prochlorperazine 3 mg buccal twice daily, increased to 6 mg twice daily if not effective.

• Second line: metoclopromide 10 mg po/i.m./i.v. tds or promethazine 25 mg at night.

• Third line: hydrocortisone 100 mg i.v. bd until able to tolerate oral prednisolone 20 mg bd, reducing weekly to 15 mg bd then continued reducing regime over a further 4 weeks. The reduction should be stopped at a dose below which symptoms recur.

• Steroids should be discontinued if there is no treatment effect within 24 hours of commencement.

• Fourth line: specific 5HT3 antagonists such as granisetron 1 mg po or i.v. or ondansetron 4 mg orally or i.v.

Vitamin supplementation

Thiamine 50 mg orally three times daily. If oral is not tolerated then one pair of Pabrinex (includes thiamine hydrochloride 250 mg) ampoules (diluted in 100 mL of normal saline and infused over 30–60 minutes) as a single i.v. dose. Repeat weekly until oral thiamine is tolerated.

Folic acid 5 mg orally per day (once tolerating oral feeding) to replace diminished stores as a result of the vomiting.

Ranitidine

For women with epigastric pain ranitidine 50 mg i.v. tds followed by 150 mg orally twice daily once tolerated

Iron therapy

Iron therapy, which is likely to exacerbate symptoms, should be stopped and restarted once nausea has settled.

Diet advice

No specific dietary advice (e.g. nil by mouth for first 24 hours) is needed as there is no evidence of benefit of withholding food or fluids.

Antithrombotic prophylaxis

Thromboembolic deterrent (TED) stockings are necessary for all women admitted and low-molecular-weight heparin prophylaxis should be considered.

Management of severe prolonged hyperemesis

In severe cases nasogastric or nasoduodenal feeding has been used. Peripherally inserted central catheters (PICC) lines are associated with an increased rate of maternal infection and venous thromboembolism (VTE) with no extra benefit. Total parenteral nutrition may be indicated in extreme situations. Complications

Wernicke’s encephalopathy, due to vitamin B1 (thiamine) deficiency, is potentially fatal. The classical presentation with the triad of confusion, ocular abnormalities and ataxia occurs in 47%. It is mostly reversible but some residual impairment remains in 60%. The fetal loss rate (excluding termination of pregnancy) with Wernicke’s encephalopathy is 37%.

The earliest reported case developed 3 weeks after the onset of vomiting and it is therefore suggested that to prevent Wernike’s encephalopathy thiamine is given to women with symptoms for more than 3 weeks.

Serum thiamine levels are unhelpful in making the diagnosis as little vitamin B1 is carried in serum. Magnetic resonance imaging has a high sensitivity for diagnosis, although treatment should not be delayed pending investigation. Observation of rapid improvement helps confirm the diagnosis and any woman presenting with a neurological abnormality in association with HG should be treated with intravenous thiamine (100 mg thiamine i.v. daily (or Pabrinex) then 50 mg orally until a balanced diet resumes.

Other maternal complications

• Central pontine myelinolysis, vasospasm of cerebral arteries, rhabdomyolysis, coagulopathy, acute renal failure and peripheral neuropathy, rupture of the oesophagus and pneumomediastinum have been reported.

• Social, economic, and psychological complications are common with HG, with possible job loss, depression, anxiety, and fear of future pregnancies. Such effects may last into the postpartum period. In particular there is evidence that women with HG can feel uncared for or dismissed by healthcare providers.

• Termination of pregnancy may be carried out for psychological indications (such as the woman feeling she is unable to care for self or existing children) or if the maternal condition is life threatening.

Fetal complications

Retrospective studies suggest possible low birthweight, fetal growth restriction, preterm delivery, low APGAR scores, or congenital abnormality. However, these findings seem restricted to those with poor pregnancy weight gain (less than 7 kg). Further reading

Fell DB, Dodds L, Joseph KS, et al. Risk factors for hyperemesis gravidarum requiring hospital admission during pregnancy. Obstet Gynecol 2006;107:277–84.

Golberg D, Szilagyi A, Graves L Hyperemesis gravidarum and Helicobacter pylori infection: a systematic review Obstet Gynecol 2007;110:695–703.

Jewell D, Young, G. Interventions for nausea and vomiting in early pregnancy. Cochrane Database Syst Rev 2003;4:CD000145.

Kirk E, Papageorghiou AT, Condous G, et al. Hyperemesis gravidarum: is an ultrasound scan necessary? Hum Reprod 2006;21:2440–2.

Koren G, Boskovic R, Hard M, et al. Motherisk-PUQE (pregnancy-unique quantification of emesis and nausea) scoring system for nausea and vomiting of pregnancy. Am J Obstet Gynecol 2002;186:S228–31.

Nelson-Piercy C, Fayers P, de Swiet M. Randomised, double-blind, placebo-controlled trial of corticosteroids for the treatment of hyperemesis gravidarum. Br J Obstet Gynaecol 2001;108:9–15.

Tan JY, Loh KC, Yeo GS, Chee YC. Transient hyperthyroidism of hyperemesis gravidarum. Br J Obstet Gynaecol 2002;109:683–8.

Verberg MF, Gillott DJ, Al-Fardan N, Grudzinskas JG. Hyperemesis gravidarum, a literature review. Hum Reprod Update 2005;11:527–39. Patient information and contacts

www.nhsdirect.nhs.uk/articles/article.aspx?articleId=260&sectionId=11 (23 Sept 2008)

Pregnancy of unknown location Definition

Pregnancy of unknown location (PUL) is a descriptive term used to classify a pregnancy in a woman who has a positive pregnancy test, but no intra- or extrauterine pregnancy is visualized on a transvaginal ultrasound scan (TVS). There is no evidence of an intrauterine gestational sac or retained products of conception and no adnexal mass suggestive of an ectopic pregnancy. Epidemiology

It is reported that between 8% and 31% of women referred for an ultrasound in early pregnancy will be classified as having a PUL. It is thought that the higher the quality of ultrasound assessment provided, the lower the PUL rate will be, as hopefully more early intrauterine pregnancies and ectopic pregnancies will be visualized on the initial TVS assessment. It has been suggested that modern units should try to maintain a PUL rate of less than 15%. Aetiology

PUL has four possible clinical outcomes: intrauterine pregnancy, failing PUL, ectopic pregnancy, and persistent PUL. Although the main concern for many clinicians is that a diagnosis of ectopic pregnancy has been missed, fortunately the majority of PULs are not ectopic pregnancies.

Very rarely a positive pregnancy test is not due to a pregnancy. Other causes include placental trophoblastic tumours and posterior cranial fossa germ cell tumours.

Intrauterine pregnancy

There is no doubt that performing an ultrasound assessment too early in pregnancy may lead to the finding of a PUL. A large number of women will present with intrauterine pregnancies that were just too early to be seen on the initial TVS assessment.

Failing PUL

Studies have shown that the majority of women (50–70%) will have spontaneously resolving pregnancies. However, the true location of the pregnancy may never be known. Although the majority will almost certainly be failed intrauterine pregnancies, a proportion will have failed ectopic pregnancies that were never visualized.

Ectopic pregnancy

Studies have shown that the majority of ectopic pregnancies (>90%) should be visualized on TVS prior to treatment. However, not all of these are visualized on the initial TVS assessment. One study has shown that a quarter of ectopic pregnancies are initially classified as PULs. The reason they are not seen on the initial TVS is that they are probably too small and early in the disease process to be visualized. However, only 7–20% of all women presenting with a PUL will subsequently will diagnosed with an ectopic pregnancy.

Persistent PUL

A small group of women will be diagnosed as having a persistent PUL, which tends to behave biochemically as an ectopic pregnancy. The serum human chorionic gonadotrophin (hCG) levels fail to decline and no evidence of the pregnancy is ever identified by TVS or laparoscopy. Clinical approach: diagnosis

History

A woman can only be classified as having a PUL if she has a positive urinary pregnancy test or serum hCG level and has had a TVS performed which shows no evidence of either an intra- or an extrauterine pregnancy.

Important things to elucidate from the history include

• date of the last menstrual period

• date of the first positive pregnancy test

• history and nature of any pain

• history and nature of any bleeding: with or without clots, passage of any possible products of conception

• risk factors for ectopic pregnancy (previous ectopic pregnancy, history of pelvic inflammatory disease, history of infertility, previous tubal surgery, assisted conception, use of an intrauterine contraceptive device, or emergency contraception in this pregnancy)

• the results of any previous TVS examinations that may have already been performed.

Although the finding of an empty uterus on TVS with a history of heavy bleeding with clots is highly suggestive of a complete miscarriage, unless a previous ultrasound scan has shown an intrauterine pregnancy, such women should be classified as having a PUL. These women need to be followed up, as a proportion will have an underlying ectopic pregnancy. One study has shown that 6% of women with a history suggestive of a complete miscarriage had an ectopic pregnancy.

Examination

• Abdominal palpation may be unremarkable or confirm an acute abdomen.

• Vaginal examination may be unremarkable or confirm cervical excitation or adnexal tenderness.

Investigations

Serum hCG: Serum hCG is the most commonly used hormone in the management of PULs.

Single levels

A single serum hCG measurement is used by some as a discriminatory level to help with the detection of ectopic pregnancy within the PUL population. The concept was developed with respect to transabdominal ultrasound, when it was reported that the absence of an intrauterine sac at an hCG concentration of >6500 IU/L has a sensitivity of 100% for the detection of ectopic pregnancy. Lower discriminatory serum hCG levels of between 1000 and 2000 IU/L are often used with TVS. However, the use of a single value of serum hCG in a PUL population is of limited value. Many ectopic pregnancies in a PUL population have a low serum hCG level, well below used discriminatory zones, so the clinician may be falsely reassured. Conversely some failing PULs and intrauterine pregnancies may have an initial serum hCG level above the discriminatory zone leading to possible unnecessary intervention in this group.

Serial levels

Serial hCG levels taken 48 hours apart are more useful in the prediction of PUL outcome. Although there are suggested cut-offs for the minimal rise and minimal decrease in serum hCG levels over 48 hours for the prediction of intrauterine pregnancies and failing PULs respectively, unfortunately there is no single way to characterize the behaviour of serum hCG levels in ectopic pregnancies.

The concept of a minimal rise in serum hCG levels of 66% in 48 hours to predict an intrauterine pregnancy has been around for more than 20 years. Recently a more conservative minimal rise of 35% over 48 hours has been suggested to predict intrauterine pregnancy. However, it has been reported that 13–21% of ectopic pregnancies have an hCG rise that mimics an intrauterine pregnancy.

A decrease in serum hCG of >15% over 48 hours has been used to predict a failing PUL. Again, around 10% of ectopic pregnancies will have a fall in serum hCG similar to spontaneous miscarriages.

Serum progesterone: levels of 25 nmol/L are ‘likely to predict’ and levels >60 nmol/L are ‘strongly associated’ with viable pregnancies. However, although progesterone levels seem to be good at predicting pregnancy viability, they are relatively poor at predicting pregnancy location. Management

Expectant management

Expectant management has been shown to be safe for the majority of haemodynamically stable women with PULs. Intervention rates of 0.3–29% have been quoted, but at present there is no consensus as to what is an acceptable intervention rate in this group. Women should be followed up with hormone measurements and repeat TVS examinations until a diagnosis is confirmed or serum hCG levels decline to non-pregnant levels. Diagnostic laparoscopy and curettage may be needed in cases where the clinical situation remains unclear or if the woman is symptomatic or haemodynamically unstable.

Diagnostic laparoscopy

The combination of a positive pregnancy test, an empty uterus on TVS and a serum hCG level above a discriminatory level of hCG is an accepted indication for laparoscopy in many units. However, this policy may lead to a high number of unnecessary surgical interventions. Diagnostic laparoscopy should be reserved for symptomatic women or cases where a concomitant condition such as fibroids may reduce the ability to visualize a pregnancy on TVS.

Uterine curettage

Uterine curettage has been combined into various diagnostic algorithms along with serum hCG and progesterone levels for the prediction of PUL outcome. It should not be used routinely in the management of PULs. It may play a role in diagnosing the location of failing or persistent PULs, but should not be used until the possibility of a viable intrauterine pregnancy has been excluded. Further reading

Banerjee S, Aslam N, Woelfer B, et al. Expectant management of early pregnancies if unknown location: a prospective evaluation of methods to predict spontaneous resolution of pregnancy. Br J Obstet Gynaecol 2001;108:158–63.

Condous G, Okaro E, Khalid A, Bourne T. Do we need to follow up complete miscarriages with serum human chorionic gonadotrophin levels? Br J Obstet Gynaecol 2004;111:1–3.

Condous G, Timmerman D, Goldstein S, et al. Pregnancies of unknown location: consensus statement. Ultrasound Obstet Gynecol 2006;28:121–22.

Kirk E, Daemen A, Papageorghiou, et al. Why are some ectopic pregnancies characterized as pregnancies of unknown location at the initial transvaginal ultrasound examination? Acta Obstet Gynecol Scand 2008;87:1150–4.

Kirk E, Bourne T. Predicting outcomes in pregnancies of unknown location. Women’s Health 2008;4:491–9.

Sagili H, Mohamed K. Pregnancy of unknown location: an evidence-based approach to management. Obstet Gynaecol 2008;10:224–30.

Royal College of Obstetricians and Gynaecologists (RCOG). The management of tubal pregnancy. Green-top Guideline No. 21. London: RCOG 2004.

RCOG. Green-top Guideline No. 25. The management of early pregnancy loss. London: RCOG 2006. Patient resources

The Ectopic Pregnancy Trust www.ectopic.org.uk

The Miscarriage Association www.miscarriageassociation.org.uk