CHAPTER 6

EARLY PREGNANCY DEATHS

Summary

This Chapter considers all deaths from ectopic pregnancy, spontaneous abortion, or termination of pregnancy, before 24 weeks' gestation. Corresponding Chapters in previous Reports included deaths before 20 weeks. Since all such deaths reported in 1994-96 occurred before 20 weeks, direct comparison with previous triennia is possible. Other deaths in early pregnancy are counted in the relevant Chapters of this Report.

There were 12 deaths from ectopic pregnancy (compared to eight in 1991-93), two following spontaneous miscarriage (compared to three in 1991-93) and one as a direct consequence of a legal termination of pregnancy (compared to five in 1991-93). These figures are also shown in Table 6.1.

In addition, there were 15 Direct deaths from pulmonary embolism in pregnancies before 24 weeks' gestation (which are discussed and counted in Chapter 2), and one Indirect death from myocardial infarction following termination of pregnancy (counted in Chapter 10).

It is gratifying to identify a decrease in deaths associated with termination of pregnancy. Nevertheless many of the early pregnancy deaths showed features of substandard care and, in some cases, deviation from guidelines published in previous Reports. Excluding two cases with insufficient details the proportion of deaths from ectopic pregnancies associated with substandard care was 80%, as in the previous Report. Both cases of spontaneous miscarriage, and the one death following a termination of pregnancy were also associated with substandard care.

Early pregnancy: key recommendations Ectopic pregnancy It is essential that GPs and other clinicians, including staff in Accident and Emergency Departments, consider the diagnosis of ectopic pregnancy in any woman of reproductive age who complains of abdominal pain. It is important to recognise that the clinical presentation is often not "classical". ß -hCG testing should be considered in any young woman with unexplained abdominal pain whether or not she has missed a period or had abnormal vaginal bleeding. Gastrointestinal symptoms may be prominent in ectopic pregnancy, notably diarrhoea and painful defaecation. An ultrasonically "empty" uterus in a woman who presents with vaginal bleeding in pregnancy may indicate an ectopic pregnancy. When ectopic pregnancy is suspected, or when a young woman with a predisposing history complains of severe abdominal pain, rapid referral and assessment are vital. The decision to treat ectopic pregnancy by laparoscopic rather than open surgery should be based on the experience and expertise of the operator and a judgement about the suitability of the procedure for the individual woman. Women in haemorrhagic shock following rupture of ectopic pregnancy need to be transferred promptly to the operating theatre. Transfer must not be delayed by attempts to try to re-establish a normal circulating plasma volume. Several of the women who died were socially disadvantaged, and included recent immigrants, women with little English, and those with itinerant occupations or previous substance abuse. Such factors may make assessment or communication difficult or lead to a delay in seeking help. An RCOG guideline on treatment of ectopic pregnancy, including laparoscopic treatment, is being prepared and should be available during 1999. Spontaneous abortion When there is evidence of serious clinical decline postoperatively, or in association with miscarriage, and especially when sepsis is suspected, it is vital that there is direct and rapid involvement of senior obstetric staff and other relevant specialists including intensive care doctors. In cases of septic abortion, evacuation of the uterus should be performed by experienced surgeons, optimally around one hour after intravenous antibiotics. Termination of pregnancy Ideally, all women should undergo ultrasound examination before termination of pregnancy to establish gestational age, viability, and site. Failure to obtain tissue at suction termination of what appears, ultrasonically, to be an intrauterine pregnancy suggests an abnormal site of implantation, including cornual pregnancy. Laparoscopy, and/or laparotomy, is essential if perforation of the uterus occurs during suction termination of pregnancy, because of the risk of bowel damage and life-threatening sequelae. There remain questions as to whether dilatation and evacuation is a safe and appropriate method of terminating second trimester pregnancies when effective medical alternatives exist.

As previously discussed in the section on Definitions, the denominator used for early pregnancy deaths is the number of "estimated pregnancies". This is a combination of the number of maternities, together with legal terminations, hospital admissions for spontaneous miscarriages (at less than 24 weeks' gestation) and ectopic pregnancies, with an adjustment to allow for the period of gestation and maternal age at conception. It is the preferred denominator for deaths in early pregnancy. The estimate for the United Kingdom 1994-96 was 2,914,600. However, this is still an underestimate of the actual number of pregnancies since the figure does not include other pregnancies which miscarry early, those where the woman is not admitted to hospital, or those where the woman herself may not even know she is pregnant. Further details are available in Appendix 1 to this Report.

Table 6.1 shows the number of women dying from ectopic pregnancy and rates per thousand estimated pregnancies for 1987-96. Table 6.2 shows the rates, per 100,000 estimated pregnancies, for deaths following spontaneous abortion and legal termination of pregnancy.

Table 6.1 - Deaths from ectopic pregnancies and rates per 1,000 estimated pregnancies; England and Wales 1988-90 and United Kingdom 1991-96.

Table 6.2 - Direct abortion deaths by type of abortion, rates per million maternities and estimated pregnancies; United Kingdom 1985-96.

Ectopic pregnancy

Of the 12 deaths from ectopic pregnancy, eight were considered to be associated with substandard care. This resulted from delayed diagnosis and inappropriate investigation and treatment. In two other cases insufficient information was made available for the Assessors to make a judgement as to whether care was acceptable or not. In the two remaining cases, collapse was so sudden that there was little scope for any effective medical intervention.

Ectopic pregnancy is often difficult to diagnose. It is essential that GPs, and other clinicians, always consider the possibility of ectopic pregnancy in a woman of reproductive age who complains of abdominal pain. The clinical presentation may not be "classical" and, in particular, that there may be no history of a missed period. Gastrointestinal symptoms were prominent in some of the cases described here, notably diarrhoea and painful defaecation. Several of the women also showed features of social exclusion, including immigrant status, little English, itinerant occupation, or previous substance abuse. As in previous Reports, deaths occurred despite clear clinical pointers to the diagnosis of ectopic pregnancy and, in some cases, despite repeated consultations with medical staff.

A patient underwent an ultrasound scan because of vaginal bleeding at eight weeks of pregnancy. This showed no evidence of intrauterine pregnancy and a diagnosis of probable complete miscarriage was made by the ultrasonographer. Subsequent attendances to general practitioners and A&E departments with acute abdominal pain, vaginal bleeding, fainting episodes, and pallor produced diagnoses of pelvic inflammatory disease and constipation which preceded her final collapse and death from ruptured ectopic pregnancy.

Previous Reports have emphasised the fact that an ultrasonically empty uterus in a woman who presents with vaginal bleeding in pregnancy may indicate an ectopic pregnancy. Further investigation should be based on suggestive symptoms, clinical signs and ß -hCG estimation, which is readily available. In this case insufficient attention was paid to symptoms and signs.

Ultrasound can also be misleading in the diagnosis of cornual pregnancy (see later) and of more advanced extrauterine pregnancies, as in the next case. This also applies to another woman who died of a pulmonary embolus and is counted and discussed in Chapter 2:

A fetus was thought to be small for gestational age on ultrasound examination prompted by a raised serum alpha-fetoprotein result. Detailed examination was undertaken by an experienced sonographer who confirmed this impression and also noted an echogenic mass in the lower abdomen. Fetal karyotyping was performed. Less than two weeks later, the patient re-presented to hospital with abdominal pain and tenderness and tachycardia, and was admitted. A junior doctor prescribed repeated doses of opiate analgesia for increasing pain without apparently making a diagnosis. The patient was found dead in bed the following morning. Autopsy showed an "abrupted" abdominal pregnancy.

With hindsight, the pelvic mass seen on ultrasonography was presumably the empty uterus. Although advanced extrauterine pregnancy can be a very difficult diagnosis to make by ultrasound, it is regrettable that it was not apparently considered here. Junior doctors should not prescribe regular doses of opiate without making a diagnosis or asking a more senior colleague to assess the patient.

A young woman complained repeatedly of intermittent abdominal pain and episodes of faintness. She was amenorrhoeic and had a positive pregnancy test at six weeks' gestation. She was noted to be pale by one of several GPs that she consulted. A month later she was visited twice in one day by her GP, because of abdominal pain and diarrhoea. The gastrointestinal symptoms masked the clear diagnosis of ectopic pregnancy. No referral was made to a specialist and no ultrasound examination was requested. The patient collapsed at home the next day and was found to be asystolic by the attending paramedics. Cardiorespiratory resuscitation was effected during transfer to the local A&E Department where resuscitation continued. Eventual ultrasound examination identified free fluid in the peritoneal cavity, and the gynaecological team was called. Despite rapid laparotomy and salpingectomy for a ruptured ectopic pregnancy, the patient died shortly afterwards in the ICU.

This case illustrates how presentation to different doctors, together with some atypical symptoms, can lead to failure to consider the diagnosis of ectopic pregnancy despite strong clinical indicators. It is also vital that staff in A&E departments consider the diagnosis in women of reproductive age who present with hypovolaemic shock. Ultrasound examination in such circumstances merely delays definitive surgical intervention; the presence of intraperitoneal blood can be demonstrated much more rapidly by paracentesis.

A young woman with a history of pregnancy terminations complained of severe abdominal pain one evening. Her GP gave advice over the telephone but later visited because the pain had worsened. There was no history of a missed period and because the patient complained of nausea, it was assumed that the problem was gastrointestinal in origin. Less than an hour later, the woman had a cardiac arrest in the presence of the GP, who had been re-called. Cardio-pulmonary resuscitation was initiated, and continued until arrival at hospital. Attempts were then made to resuscitate the patient by infusion through multiple intravenous cannulae, with minimal success, before laparotomy was undertaken. This showed a ruptured ectopic pregnancy and three litres of blood in the peritoneal cavity. The patient died in theatre.

In this case the delay in diagnosis was again caused by the presence of gastrointestinal symptoms. It was compounded by futile attempts at resuscitation before laparotomy. It is important to re-emphasise the traditional teaching that women in haemorrhagic shock following rupture of ectopic pregnancy need to be transferred promptly to the operating theatre without delays to try to re-establish a normal circulating plasma volume.

When ectopic pregnancy is suspected, or when a young woman with a predisposing history complains of severe abdominal pain, rapid medical attendance and assessment are vital:

Another young woman with a history of previous terminations of pregnancy was seen, at home, by a locum doctor during the night because of severe abdominal pain. A further call to the Practice the following day was made by a neighbour because the patient appeared unwell and continued to complain of pain. The GP arrived three hours later to find the woman dead. Autopsy showed her to have a ruptured ectopic pregnancy.

Delay in diagnosis may result from communication problems. Two women for whom English was not their first language died from ruptured ectopic pregnancies:

One had oligomenorrhoea following medical treatment of endometriosis and had attended her GP for another reason two weeks before emergency admission to an A&E Department, shocked from a ruptured ectopic pregnancy. Despite rapid resuscitation and surgery, death occurred subsequently from adult respiratory distress syndrome. Whether the outcome would have been different if the patient had been able to speak English, or had access to a clinician with knowledge of her language, remains speculative.

The other woman was admitted in asystole to an A&E Department; it is not known what, if any, prior symptoms she might have had.

There has been a trend in recent years towards less use of laparoscopy for diagnostic purposes in suspected ectopic pregnancy, with greater reliance on ultrasound and quantitative ß -hCG estimation. At the same time there has been more use of laparoscopy for surgical treatment of confirmed ectopic pregnancy. Ultrasound examination, however, can be especially misleading in the diagnosis of advanced extrauterine pregnancy, as mentioned above, and of cornual pregnancy:

A woman had a termination of pregnancy in a private clinic at eight weeks' gestation. A suction curettage was performed and apparently appropriate tissue noted in the cannula during the procedure. A week later, however, she was re-referred because she still felt pregnant and her pregnancy test remained positive. An ultrasound scan was reported as showing a nine-week intrauterine pregnancy. A repeat suction curettage was performed but no tissue was obtained. Consideration was given in theatre to performing a laparoscopy, but this was not done because a pregnancy test in theatre was negative. She was discharged home but, in light of a subsequently reported high serum ß -hCG result, an appointment was made for her to re-attend the clinic. In the meantime, she attended an A&E Department with acute abdominal pain and a diagnosis of probable urinary tract infection was made. Later the same day, she returned to the same A&E Department and had a cardiac arrest. Resuscitation was attempted but failed. At autopsy, there were 3.5 litres of blood in the peritoneal cavity and a ruptured cornual pregnancy in an otherwise normal uterus.

Although cornual pregnancy is difficult to diagnose by ultrasound, it should have been obvious after the second unsuccessful attempt at suction termination that the pregnancy was abnormally sited. An ultrasound scan or laparoscopy should have been performed in theatre to clarify the site of implantation. The patient should not have been discharged home without knowledge of the ß -hCG result. The fact that this took place in the private sector, where there may not have been "on site" laboratory services, may be relevant. When she attended the A&E Department with abdominal pain, the previous history should have alerted the staff to the serious nature of her complaint. The management of this woman was sub-standard in many respects.

"Minimal access" surgical treatment of ectopic pregnancy has become more popular but may be associated with hazards:

A parous woman with an intrauterine contraceptive device presented with acute lower abdominal pain and vaginal bleeding two weeks after what was thought to be a delayed period. There was marked pelvic tenderness and a diagnosis of pelvic inflammatory disease was made despite hypotension and absence of pyrexia. The diagnosis of ectopic pregnancy was made after ß -hCG testing and ultrasonography. Although two litres of blood were present in the peritoneal cavity, salpingostomy was performed via the laparoscope, by a non-consultant member of staff. The patient required laparotomy later the same day because of further intra-abdominal bleeding by which time a further two litres of blood had accumulated (from the ovarian vein). Salpingo-oophorectomy was performed. She died subsequently from ARDS despite intensive care, including ECMO.

In this case, the consequences of delay in diagnosis, despite the classical features of ectopic pregnancy, were compounded by ill-judged surgical treatment in a woman with major haemoperitoneum and a normal, contralateral fallopian tube. A similar case was reported during the previous triennium. An RCOG guideline on treatment of ectopic pregnancy, including laparoscopic treatment, is being prepared and should be available during 1999.

Another woman died after diagnostic laparoscopy for presumed ectopic pregnancy. She had positive ß -hCG results but no pregnancy was found at laparoscopy. Postoperatively, she developed septicaemic shock and was found to have a perforated sigmoid colon. She died despite intensive care. In this case insufficient information has been made available to the Assessors to permit comment on the standard of care.

In another case there was evidence of substandard care after appropriate initial treatment:

A nulliparous woman with a history of pelvic inflammatory disease presented, shocked, with a ruptured ectopic pregnancy at eight weeks' gestation. She underwent laparotomy and salpingectomy without any apparent surgical or anaesthetic difficulties. She became unwell after 36 hours with evidence of sepsis and respiratory embarrassment. Despite vigorous intensive care she died. Autopsy showed ARDS, sickle lung, endocarditis due to Candida, pulmonary embolism and myocardial infarction. There was nothing to suggest aspiration of stomach contents as the cause of respiratory disease.

Despite clinical evidence of serious decline postoperatively in this patient, there was no direct involvement of senior obstetric staff and requests to involve senior physicians were tardy.

In one further case, no information was available other than the death certificate which stated the cause of death as ectopic pregnancy.

It is again necessary to re-emphasise the conclusions in the previous Report. Ectopic pregnancy continues to be an important cause of maternal deaths and the need for early diagnosis is paramount. Awareness of the possibility of an ectopic pregnancy in any woman of reproductive age is essential. The emphasis is again placed on the importance of the history as well as suggestive signs on examination. Primary care doctors in particular must be constantly aware of the possibility of this diagnosis.

When a woman presents to her GP or to an A&E Department with unexplained abdominal pain, with or without vaginal bleeding, it is essential to exclude an ectopic pregnancy. The ready availability of sensitive ß -hCG kits means that the diagnosis of early pregnancy can be made in GPs' surgeries or A&E Departments. The test itself is very reliable: the limiting factor is thinking of using it. If the diagnosis of ectopic pregnancy is likely then vaginal examination is best deferred until the patient is in hospital.

Abortion

This section discusses the three deaths related to spontaneous and legal abortion. This is the fifth successive Report in which no identified deaths from illegal abortion are reported. Table 6.2 shows data on abortion deaths from 1985, when the Report first covered the United Kingdom as a whole.

Spontaneous miscarriage

Although deaths from septic abortion have been rare in the UK in recent decades, this remains a very important cause of death globally¹. When serious infection does occur, prompt action is necessary to minimise the risk to the mother. This includes appropriate antibiotic treatment, rapid removal of infected tissue, intensive care, and involvement of senior clinical and microbiological staff.

There is often uncertainty about the optimal timing of surgical evacuation of the uterus after initiating antibiotic treatment. The main initial aims of antibiotic treatment are to counter bacteraemia and to protect other tissues from becoming infected. Antibiotic penetration into necrotic, septic placental tissue is very limited and this is not the primary purpose of antibiotic treatment. It is therefore logical to perform evacuation when there are high levels of antibiotics in the bloodstream and increasing levels in normal tissues. With most antibiotics, blood levels are high around one hour after intravenous administration and then drop with increasing tissue uptake. The aim should be to perform surgical evacuation one hour after antibiotics are given. The timing of such procedures cannot, therefore, be left to the vagaries of the "emergency list" that exists in many general hospitals. Nor should the operation be left to the most junior member of staff.

In the following cases, there seemed to be delay in appreciating fully the serious nature of the problem:

A woman who had had several episodes of threatened miscarriage, re-presented to the hospital at 17 weeks' gestation with further bleeding. She miscarried nine hours later and a foul smelling discharge was noted, apparently for the first time. She was tachycardic, jaundiced, oliguric and anaemic. Disseminated intravascular coagulation was diagnosed, and appropriate therapy was initiated by a Consultant Haematologist. There was a delay of six hours before surgical evacuation of 'putrid' placental tissue by an obstetric registrar. Although intravenous antibiotic treatment had been started at the time of miscarriage, death occurred from E. coli septicaemia within hours.

A parous woman was admitted with an incomplete miscarriage at 10 weeks' gestation. Although she was markedly tachycardic and hypotensive, her haemoglobin concentration was normal. Her vital signs improved after surgical evacuation of the uterus but she was readmitted three days later with further signs of shock, although apyrexial. Despite intravenous antibiotic treatment, intensive care and hysterectomy to remove the source of infection, she died. There is no record of the responsible organism but pathological findings were compatible with septicaemia with primary infection in the uterus.

Termination of pregnancy

The previous Report emphasised the need for laparoscopy or laparotomy if perforation of the uterus occurs during suction termination of pregnancy, because of the risk of bowel damage and life-threatening sequelae. In the following case this guideline was not followed, with fatal consequences:

A young woman underwent a termination of pregnancy. The date of her last menstrual period was uncertain but, clinically, the uterus appeared to be 13-14 weeks' size. Ultrasound examination was not performed. Examination in theatre after Gemeprost pre-treatment showed the uterus to be 15 weeks' size. Dilatation and evacuation were performed followed by insertion of a suction catheter which was felt to pass through the fundus of the uterus. It was considered by the surgeon that no negative suction had been applied within the peritoneal cavity, and therefore no surgical inspection of intra-abdominal contents was performed. The patient was discharged home, apparently well, after overnight observation. She was admitted, moribund, to an A&E Department six days later and died despite attempts at resuscitation. Autopsy revealed two holes in the jejunum, and faecal peritonitis.

In addition to the need for laparoscopy after perforation, this case illustrates the value of ultrasound examination before termination if the gestational age is in doubt so that difficult and hazardous surgical terminations may be avoided. A policy of routine ultrasound would, in addition, allow the recognition of non-viable (or even non-existent) pregnancies thereby relieving some women of anxiety or a sense of guilt.

It was not clear if uterine perforation, in this case, occurred during dilatation and evacuation, or on insertion of the suction cannula. A similar death following dilatation and evacuation was reported in the 1988-90 triennium and raises questions as to whether this is an appropriate method of terminating second-trimester pregnancies when safe and effective medical alternatives exist. Research is required in this area.

Reference

1. Royston, E. & Armstrong, S. Preventing Maternal Deaths. Geneva: World Health Organization, 1989.

We welcome your comments on this site. Prepared 28 December 1998