CHAPTER 5

AMNIOTIC FLUID EMBOLISM

Summary

There were 17 deaths due to amniotic fluid embolism in the United Kingdom in 1994-96, compared with 10 in 1991-93, 11 proven and one diagnosed on clinical grounds in 1988-90 and nine in 1985-87. All but one of the women were aged over 25, and 12 were over 30. Two women were nulliparous and three were of high parity (para 4 or more). Seven had received prostaglandin or oxytocin and two had had amniotomy. Fourteen were delivered by forceps or caesarean section, the indication in eleven of these cases being maternal collapse. In eight cases the interval between collapse and death was short, in two cases this information was not available and in the other seven cases the interval was greater than six hours.

Amniotic fluid embolism

Seventeen Direct deaths were attributed to amniotic fluid embolism (AFE). In10 cases the diagnosis was confirmed by the finding of squames or hair in the lungs on histological examination. There were four deaths in which a clinical diagnosis of AFE was made without confirmation of the diagnosis at autopsy. In the remaining three cases inadequate information was available to the Assessors. The acceptance of a clinical diagnosis of AFE began with the last Report: prior to 1991-93 the Reports accepted only deaths in which autopsy provided histological evidence of squames in the lungs.

Clinically diagnosed cases

An asthmatic woman who weighed over 100kg had a history of a previous caesarean section followed by a "collapsed lung" requiring intensive care. She developed polyhydramnios and required blood transfusion for anaemia. Elective caesarean section was carried out at 36 weeks. Just after delivery of the placenta, cardiac arrest occurred. At autopsy the lungs showed anaphylactoid reaction but no squames were seen. The heart was dilated and flabby, weighing 500g, compared to a normal weight of around 300g. This case is also discussed in Chapter 16 on Pathology.

A primigravid woman went into spontaneous labour at term. Labour was augmented by syntocinon and lasted less than eight hours. She felt unwell in the last two hours of labour. There was very little liquor in labour and thick meconium was noted in the second stage. Instrumental delivery was performed, syntometrine was given and the placenta was delivered. She became suddenly cyanosed while the episiotomy was being repaired. She was transferred to an intensive care unit (ICU) where pulmonary oedema was diagnosed but she died a few hours later. No autopsy was carried out.

Care was not substandard in this case. Although cyanosis and sudden collapse are suggestive of AFE, it is not typical for collapse to occur after the placenta has been delivered.

A parous woman underwent amniocentesis because of her age. She had undergone previous instrumental deliveries because of large babies but was persuaded to attempt a vaginal delivery. Several days after term, induction of labour was begun with 1mg of prostaglandin gel followed six hours later by 2mg. About three hours later a strong contraction was noted. An epidural was set up. Fetal distress was diagnosed, and a trial of forceps was planned but caesarean section was carried out. The baby was stillborn and was heavier than her previous babies. Severe haemorrhage occurred at operation. After some delay hysterectomy was carried out. The patient was transferred to an ICU but died three days later.

The clinical picture of a sustained strong uterine contraction followed by fetal distress, stillbirth and severe haemorrhage is strongly suggestive of amniotic fluid embolism. There was delay in carrying out hysterectomy and this was considered substandard care.

The fourth clinically diagnosed case is described below, in the section on substandard care.

Age

Table 5.1 shows how the risk of the condition increases with age. In the previous two triennia no case of amniotic fluid embolism occurred in a woman under the age of 25. In this triennium only one of the patients was under 25.

Table 5.1 - Number of maternal deaths and the death rates per million maternities from amniotic fluid embolism by age; United Kingdom 1988-96. (Numbers of maternities are shown in Table 1.13)

Parity

Amniotic fluid embolism has been thought to be associated with high parity, though in 1991-93 none of the deaths was associated with high parity. In 1994-96, of the 14 women for whom parity information was available, two were primigravid, three had one previous delivery, five had two previous deliveries, and four had three or more previous deliveries.

Antenatal complications

Antenatal complications were present in about half the cases. Two women had undergone amniocentesis. Another two had developed polyhydramnios, one of whom had an intrauterine death. One woman had placenta praevia, one had fibroids, and one had a cervical suture inserted.

Induction or augmentation of labour

Amniotic fluid embolism has been thought to be associated with induction of labour and the use of oxytocic drugs, but in this triennium four of the women collapsed before labour began and another four went into labour spontaneously.

Six women had labour induced: five with prostaglandins (gel in two cases and tablets in three cases) and one with amniotomy and oxytocin. One woman received oxytocin in labour and another underwent amniotomy in labour.

Mode of delivery

Two woman died undelivered and one miscarried after collapsing at 22 weeks' gestation. Nine women were delivered by caesarean section and four by forceps. Information was not available about one case. In most cases of operative delivery the indication was maternal collapse.

In three of the cases of forceps delivery, instrumental delivery was undertaken after the signs of AFE had occurred. In the fourth case, signs of AFE occurred after delivery, though the woman had felt unwell during the last two hours of her labour.

In two of the caesarean sections, symptoms of AFE appeared after the operation had been started. The other seven caesarean sections were undertaken after symptoms of AFE had occurred. One of these was a post-mortem caesarean section:

A parous woman underwent amniocentesis because of her age. In late pregnancy she developed hypertension and proteinuria. Labour was induced with prostaglandin 1mg, repeated twice. A few hours after the third pessary she collapsed, cyanosed and foaming at the mouth. A live baby was delivered by caesarean section but the woman could not be resuscitated. At autopsy histological examination of the lungs confirmed the presence of fetal squames.

Although this case shows an association between amniocentesis, induction of labour and amniotic fluid embolism, there was no substandard care in this case. This was one of the few occasions in these Reports when peri-mortem caesarean section resulted in the birth of a live baby.

Speed of collapse

In nine cases the collapse was sudden and resuscitation proved impossible. In seven of these cases the diagnosis was histologically confirmed.

In seven cases the interval between death and collapse was longer, varying between six hours and three days. Three of these cases were histologically confirmed.

Substandard care

Care was substandard in five cases. In one (summarised above) this was because there was delay in carrying out hysterectomy to treat severe haemorrhage. In the others there were deficiencies in the midwifery or anaesthetic care:

A "grande multipara" was admitted in early labour at term after a normal pregnancy (apart from anaemia in the third trimester). She remained in hesitant labour for 24 hours. Amniotomy was carried out when the cervix was 9cm dilated. When vaginal examination was repeated over 3 hours later she collapsed. The baby, which was large, was delivered by forceps and survived. The mother died shortly after delivery and histological examination confirmed the presence of squames in the lungs.

Care was substandard in this case. Hesitant labour is unusual in a "grande multipara" and once the cervix reaches 9cm dilatation it is very worrying if more than three hours elapse before the second stage of labour begins. Persistent strong contractions at the end of the first stage may have contributed to the outcome in this case. As fewer "grande multiparous" women are seen, staff may lose their ability to recognise danger signs in such labours.

A woman with a history of a cone biopsy, recurrent miscarriages, preterm deliveries and a caesarean section had a cervical suture inserted and had several admissions with abdominal pain. She went into preterm labour, was treated with ritodrine and developed supraventricular tachycardia. The cervical suture was removed and amniotomy was carried out. Fetal bradycardia led to caesarean section. She became cyanosed before the baby was delivered. Cardiac arrest occurred. She was resuscitated and transferred to the ICU but died soon afterwards. Histological examination confirmed the presence of squames in the lungs.

Although she had been seen by a consultant physician because of her supraventricular tachycardia, the anaesthetic was given by a senior house officer. Care was therefore substandard, though this probably did not affect the outcome. In view of the poor obstetric history, it is also disappointing that the caesarean section was not performed by a consultant obstetrician.

A parous woman developed polyhydramnios at 37 weeks' gestation. Glucose tolerance was normal. At 38 weeks she noticed diminished fetal movements and intrauterine fetal death was diagnosed four days later. Labour was induced with a cervagem pessary followed by another about two hours later "because of failure to progress". After a labour lasting just over two hours she collapsed with shortness of breath. The baby was quickly delivered by forceps, after which postpartum haemorrhage occurred. She was declared dead less than six hours after the insertion of the first pessary. AFE was confirmed histologically.

Cervagem pessaries contain 1mg Gemeprost. When they are used to induce abortion in cases of intrauterine fetal death in the second trimester, the manufacturer recommends insertion into the posterior fornix at 3-hourly intervals up to a maximum of five administrations. The manufacturer recommends that "they should not be used for the induction of labour or cervical softening at term as fetal effects have not been ascertained". The listed indications for Gemeprost do not include intrauterine death at term, and other forms of prostaglandin are generally used in such cases. A two-hour interval between pessaries, particularly in an older parous woman, was felt to represent substandard care.

A parous woman with uterine fibroids had labour induced at 38 weeks' gestation because of a history of "slight shoulder dystocia" and patient request. One 3mg prostin pessary was inserted, followed by another 3mg prostin pessary. One hour after the second pessary, contractions began and two hours later there was fetal bradycardia. The cervix was 2cm dilated and caesarean section was carried out. The baby was acidotic and the uterus was "slightly lax". Two and a half hours after operation the patient became dyspnoeic and shocked and had a convulsion. Coagulopathy developed and haemorrhage was treated by hysterectomy. She died five days later in ICU after a repeat laparotomy for bleeding.

Amniotic fluid embolus was not confirmed histologically but this did not rule out the diagnosis in view of the length of time between collapse and autopsy. Care was substandard because a lower dose of prostin should have been used and the indications for induction of labour were in any case weak. Particular caution should be exercised in inducing labour in an older parous woman with any uterine abnormality.

Comment

The number of deaths from amniotic fluid embolism has risen from 10 in 1991-93 to 17 in 1994-96. This is the highest total in any of the last four Reports. Only the present Report and the previous one have accepted cases diagnosed on clinical grounds: in 1991-93 there were two such cases and in the present triennium there were four (and another three cases in which information was inadequate). If these seven cases are excluded the number of histologically confirmed deaths has varied little over the last four Reports - from nine in 1985-87 to 10 in 1994-96.

The Reports have shown consistently that age is a risk factor for amniotic fluid embolism. One woman in the present triennium was aged under 25 (and obese), three were in their late twenties and the rest were 30 or over. The average age of childbearing in the UK continues to rise and this is a possible reason for an increase in the numbers of cases of amniotic fluid embolism.

No single obstetric intervention was identified as a risk factor for amniotic fluid embolism, but it should also be noted that in only one case did the condition complicate an entirely straightforward pregnancy in a woman of low parity. Antenatal complications included amniocentesis, polyhydramnios, placenta increta, fibroids, cervical suture and intrauterine fetal death.

The use of oxytocic drugs has been thought to be a risk factor for amniotic fluid embolism, but nine of the women in this triennium received neither prostaglandin nor oxytocin. Nevertheless, two of the nine underwent amniotomy and another two were women of high parity, in whom normal uterine contractions can be expected to be particularly strong.

The death rate from amniotic fluid embolism could be reduced by avoiding uterine over-stimulation and by prompt diagnosis of obstructed labour. This, however, applies only to a small number of cases. The previous Report suggested that a more promising strategy for reducing deaths was better treatment of women who survive long enough to be transferred to intensive care. In the present triennium there were seven such cases but only one of these had substandard care after delivery. Examination of the other cases has not suggested how the intensive care of this condition can be improved.

Amniotic fluid embolism remains a frustrating challenge and only general recommendations can be made. Rates of obstetric intervention in the form of amniocentesis and induction and augmentation of labour should be kept as low as possible, and research into the condition is still necessary.

The UK amniotic fluid embolism register

A confidential register of all cases of AFE is to be established for the United Kingdom. The aim is to identify any differences or common factors between survivors and fatalities, which may help to reduce maternal deaths from this condition.

The entry criteria are:

• Acute hypotension or cardiac arrest
• Acute hypoxia (dyspnoea, cyanosis or respiratory arrest)
• Coagulopathy (laboratory evidence of intravascular coagulation or severe haemorrhage)
• Onset of all of the above during labour, caesarean section or within 30 minutes of delivery
• No other clinical condition or potential explanation for the symptoms and signs.

All cases of suspected or proven AFE, whether they survived or not, should be reported to:

Mr Derek Tuffnell, Consultant Obstetrician,

Bradford Royal Infirmary,

Duckworth Lane

Bradford,

BD9 6RJ

West Yorkshire.

Tel 01274 364520

Fax 01274 366690

We welcome your comments on this site. Prepared 14 December 1998