CHAPTER 16

PATHOLOGY

This Chapter differs from the others in that it is designed to stand alone and be used by pathologists without the need to read the entire Report.

The quality of the autopsy

The quality of the maternal autopsy has been adversely commented upon in previous Reports and it is disappointing that the standard of so many autopsies remains inadequate when judged against criteria defined by the Royal College of Pathologists' Bulletin (1993)¹. Some autopsies were technically deficient in one or more ways - no clinical history, inadequate descriptions of organs, no organ weights, or no histology. For example:

A woman had a tricuspid valve replacement some years ago. During her uneventful pregnancy she had been reviewed by cardiologists who recommended prophylactic antibiotics in labour. Prophylactic anticoagulation was not considered necessary. She was admitted at term with prolonged rupture of membranes and whilst in labour developed pyrexia and tachycardia, for which she received intravenous antibiotics. Poor fetal progress resulted in a caesarean section, following which the patient made a good recovery and was discharged home. She was found dead at home about three weeks postpartum.

The cause of death at autopsy was given as cardiac failure due to tricuspid incompetence. The lungs were described as grossly oedematous but neither the lungs nor the heart were weighed. The right ventricle was described as hypertrophied, but neither ventricle was either measured or weighed. The tricuspid valve showed extensive nodular calcification of the undersurface adjacent to the valve ring but there was no attempt to culture the valve. There was no attempt to explain how tricuspid incompetence gave rise to severe pulmonary oedema, particularly when the strain of pregnancy had not produced any clinical cardiovascular deterioration. Apart from acute congestion of the liver all other organs were described as normal but there were no weights and the descriptions of the organs were minimal. No histology was performed.

An older woman was eight weeks pregnant and suddenly collapsed at home. Emergency laparotomy showed extensive right-sided retroperitoneal bleeding but she was moribund and could not be resuscitated.

At autopsy, ruptured iliac veins were given as the cause of the extensive retroperitoneal bleeding, but there were no details with regard to its site, pathological appearance or relation to the uterus and pelvic organs. There was no description of the uterus or the conceptus.

Although thrombus in the inferior vena cava was identified, its extent and location were not described. Despite the very unusual nature of this case, no histology was taken.

Whilst some autopsies are technically deficient, in many the cause of an unexpected death is clearly identified from the autopsy, but there is failure to address other clinical problems relating to the case:

An obese multiparous woman suddenly died at 12 weeks of pregnancy. She had a history of myocarditis and at 10 weeks gestation had had chorionic villus sampling.

The autopsy clearly identified death as due to pulmonary embolus with an iliac vein thrombosis. However, there was no mention of some organs in the report - pancreas, bladder and ovaries. There was no description of the placental bed site and the descriptions of the pulmonary artery emboli and of the venous tree were very simple. As no histology was performed, confirmatory evidence of the previous myocarditis was not obtained and alternative cardiac pathologies were not excluded. Also there was no search for any evidence of pulmonary hypertension or previous pulmonary emboli as an alternative explanation of her previous 'myocarditis'.

An obese multiparous woman had a history of severe asthma and was a heavy smoker. Her current pregnancy was complicated by a placental cord tumour with polyhydramnios, and elective caesarean section was planned. Bronchodilators were given, together with preoperative hydrocortisone. The patient collapsed immediately after delivery of the placenta, severe disseminated intravascular coagulation supervened and she quickly died. A clinical diagnosis of amniotic fluid embolus was made.

No clinical history was given in the post-mortem report, which recorded an abnormal heart weighing
500 g. The ventricles were not separately weighed, and there was no attempt to identify the cause of this abnormality. No cardiac pathologist's opinion was sought. Clots and remnants of amniotic membranes adherent to the uterine wall were identified, but there was no documentation of any uterine wall tears or damage. The liver, at twice the normal weight, was described as fatty. There was no reference to the pathological changes of asthma within the lungs. Histologically, only the lungs were examined and there was no reference to the changes of asthma nor was there any indication of any special search for amniotic fluid embolus. There was no histological examination of the heart, the placental bed site, or any other organ.

A primipara in her twenties had a normal pregnancy up to 37 weeks, when she became breathless. She was admitted the following day and a cardiologist diagnosed cardiomyopathy. She was anticoagulated and labour was induced. After delivery she was transferred to the Intensive Care Unit and made rapid recovery. Massive pulmonary embolus clinically occurred on the 16^th postpartum day.

In the autopsy report there was no review of the clinical history and only the heart was weighed. Description of the heart identified right ventricular hypertrophy, without detail, and up to 40% narrowing of the left anterior descending coronary artery. Pulmonary emboli within the pulmonary trunk and major arteries were confirmed as the immediate cause of death but there was no attempt to identify any previous episode of thromboembolisation. Despite the discrepancies in the pathological findings and with the clinical diagnosis of cardiomyopathy, there was no histology performed on the heart or any of the other organs, nor was specialist cardiac pathological opinion sought.

The inadequacy of the autopsy does not necessarily relate to failure to address specific pathological features of Direct maternal deaths:

A primigravida had suffered from epilepsy for many years. Her therapy was changed from phenytoin to carbamazepine during her pregnancy. Her drug levels were subsequently monitored and were within the therapeutic range three weeks before death. When 36 weeks pregnant, she had a 'stomach upset' and she was found dead in bed the following day. There was a previous medical history of depression with five suicide attempts.

None of the previous medical history or features of drug therapy were identified in the post-mortem report and, although histology was performed on the heart, lungs, liver and kidneys, there was no histology on the placental bed or placenta. More particularly, even though death was attributed to epilepsy, no drug levels were obtained either to establish therapeutic levels at the time of death or to exclude overdose as an alternative cause of death.

Good autopsies

Conversely, a good autopsy contributes significantly to the assessment:

A woman in her thirties had a spontaneous vaginal delivery at 41 weeks' gestation. Within minutes of delivery there was sudden collapse with rapid demise. Clinically, amniotic fluid embolus was diagnosed.

At autopsy, massive haemoperitoneum with a ruptured aneurysm of the splenic artery was identified. There was detailed comprehensive description of all the organs but, more especially, there was detailed histology. Amniotic fluid embolism was carefully excluded and the lack of any systemic arterial pathology was demonstrated.

An obese, multiparous woman had had a deep vein thrombosis (DVT) in her first pregnancy. Since then she had been on warfarin, recently changed to aspirin. She was not taking the oral contraceptive. She was a late booking at 26 weeks in her second pregnancy and was a poor attender at antenatal clinic. It was considered that heparin self-administration would be unreliable and that she should be anticoagulated with warfarin postpartum. At 39 weeks she presented with a swollen hot, tender, right thigh. Labour was induced. A small postpartum haemorrhage occurred but there were no retained products. She was out of bed after 24 hours, ambulant at 48 hours and discharged at 5 days post-partum. Heparin was given postpartum, but warfarin was not re-started. Four months postpartum she was referred to physicians with recurring breathlessness and weight loss since her pregnancy. A diagnosis of hyperthyroidism was made and she was treated with carbimazole. When seen again six weeks later she still had distressing shortness of breath and was admitted for investigation. She collapsed and died the following day.

The autopsy included a detailed review of the clinical history and documented the height and weight of the body. At the autopsy, pulmonary emboli of varying ages were identified, there was a detailed search of the venous tree and right ventricular hypertrophy was documented as the major abnormality in the enlarged heart. There was detailed histology of all organs and recanalised pulmonary emboli, plexiform lesions and pulmonary hypertension were identified in the lungs. The thyroid was histologically normal.

Autopsies by Chapter

Chapter 2: Thromboembolism:

Of the 48 deaths attributed to thromboembolism in this Report several did not have a post-mortem and in some further cases no autopsy report was available. Of the remainder, only eight could be described as model (five) or good (three). These reports not only established the cause of death and searched for the source, but also specifically searched for pathological changes that would contribute to understanding other significant episodes in the clinical history.

An obese woman, whose mother had had multiple pulmonary emboli, developed proteinuria at 41 weeks of her pregnancy. Labour was induced and a large baby was delivered by ventouse and forceps. She was immobile after delivery. Two weeks postpartum she had an episode of pleuritic chest pain treated conservatively but without anticoagulation. Some weeks later she became suddenly short of breath and the following day she collapsed and died.

Autopsy demonstrated a large saddle embolus in the pulmonary trunk with widespread embolism of small pulmonary arteries throughout both lungs (which also showed infarction). Thrombus was identified in the right external iliac vein and there was a detailed description of the patent veins of the abdomen, pelvis and legs. Extensive histology confirmed that there was fresh embolization of the lungs, but also showed organisation of some thrombi to pulmonary arterial walls, indicating a previous episode of embolization, entirely consistent with the clinical episode two weeks prior to death.

Sadly, not all cases reach this standard. Many have been classified as adequate because the autopsy findings satisfactorily explained the clinical history, even if no histology was performed and the College Guidelines were not met.

Ten reports would be considered inadequate/poor even by these relaxed criteria, the following being typical examples:

A parous woman in her thirties was admitted to hospital in early pregnancy with hyperemesis gravidarum. She was rehydrated, treated with pyridoxine and given TED stockings to wear, but a week later she collapsed and died with chest pain, severe shortness of breath and cyanosis. She had a history of severe headaches with left-sided weakness which had been diagnosed on CT scan as infarction adjacent to the anterior horn of the left ventricle.

The autopsy report, including patient details and cause of death, was limited to one side of A4 paper. There was no history, and there were no organ weights. The brain was described as "normal external and cut surfaces" . Emboli were identified in the lungs, but the venous source was not identified. No search of pelvic veins was made. Histology of the brain, liver, kidneys and lungs was not undertaken. The brain, liver and kidneys were described as normal. There was no attempt to identify the area of radiologically diagnosed cerebral infarction, there was no description of the hypothalamus or mamilliary bodies to exclude acute Wernicke's encephalopathy, and the search for the source of the pulmonary emboli was inadequate.

A known multiparous asthmatic had an uneventful pregnancy until term, when failure to progress in labour resulted in a caesarean section under spinal anaesthesia. Postoperatively she had puerperal sepsis with positive high vaginal swabs and blood cultures. This was vigorously treated with intravenous antibiotics. She was discharged after a week on penicillin, metronidazole and ciprofloxacin. She had complained of pain in her calf postoperatively and, although there was no evidence of thrombosis on Doppler ultrasound, TED stockings were prescribed. She had a transient undiagnosed illness 11 days after discharge, but otherwise seemed reasonably well until three weeks postpartum when she complained of feeling very unwell, and then suddenly collapsed and died.

Autopsy macroscopically was very thorough and detailed, identifying the massive pulmonary emboli, a source in the leg veins and patent pelvic veins. However, the clinical history of asthma and puerperal sepsis was not addressed, and no histology was taken. There was no attempt therefore to assess whether previous episodes of embolism had occurred.

Chapter 3: Hypertensive disorders of pregnancy

There were 20 deaths directly attributed to pregnancy-induced hypertension. As has been identified in the Chapter, the mean time between the last antenatal attendance and the development of severe pre-eclampsia or eclampsia was 6.5 days. Many cases also occurred between 26 and 33 weeks of gestation. It is therefore entirely conceivable that fulminating pre-eclampsia, giving rise to unexpected death, may occur de novo in the intervals between antenatal assessments.

A young primigravida developed proteinuria at 31 weeks' gestation but had a normal blood pressure and, therefore, was not referred for assessment. One week later she died at home. She had been complaining of feeling unwell with a swollen face when last seen by friends.

The autopsy was thorough and detailed and clearly demonstrated findings consistent with fulminating pre-eclampsia.

A multiparous woman had been treated for one month for a microcytic hypochromic anaemia. There was no history of epilepsy and during her pregnancy, blood pressure and urinalysis had been normal. At 31 weeks' gestation she was found drowned in the bath.

A detailed and thorough autopsy showed minimal macroscopic pathology but histology clearly identified changes of pre-eclampsia in the liver, kidneys, placenta and placental bed. A drug screen was negative. Death was attributed to drowning secondary to an eclamptic fit.

A primigravida was admitted unconscious with hypertension and proteinuria. She had been normotensive 11 days prior to admission. CT scan of the brain showed a large, right parietal haematoma.

A very detailed autopsy with histology confirmed the intracerebral haemorrhage with changes of pre-eclampsia within the brain, liver, kidneys and placenta.

In contrast to the last case there is an increasing tendency for death from cerebral haemorrhage not to come to autopsy. Frequently, therefore, the cause of the cerebral haemorrhage - arteriovenous malformation, berry aneurysm, systemic hypertension, or pregnancy-induced hypertension is not ascertained.

A multiparous woman had suffered from severe essential hypertension since her teens. She also had asthma for which she used salbutamol. She was under the care of renal physicians. During her pregnancy she was on Isradapine and frusemide. At 28 weeks' gestation she was admitted for an assessment of her blood pressure, which was 160/110 mmHg. One week later she was admitted with signs of a stroke which clinically was diagnosed as subarachnoid haemorrhage. There was no post-mortem.

The lack of post-mortem is surprising as severe essential hypertension from such a young age is unusual in itself. Also, the true nature of the cerebral haemorrhage was not ascertained and pre-eclampsia/eclampsia was not satisfactorily excluded.

A primigravida with a history of asthma and a family history of hypertension had borderline raised blood pressure during the course of her pregnancy. She was admitted at 36 weeks' gestation with a blood pressure of 140/85 mmHg and digital and ankle oedema. There was no proteinuria. Spontaneous vaginal delivery occurred the next day, following which her blood pressure settled to 110-135 systolic and 80-90 diastolic. She was discharged on day five post-partum, but was re-admitted four days later with headache and vomiting, and a blood pressure of 154/100 mmHg. She was orientated in time, place and person, but two hours later she was drowsy, writhing and moaning. One hour later she had a severe, right hemiplegia. CT scan showed intracerebral haemorrhage and she was declared brain dead.

There was no post-mortem, and the cause of the cerebral haemorrhage remained undetermined.

Even when active treatment has been initiated, death can be due to other causes:

An older woman was admitted at 35 weeks' gestation with hypertension (150/95 mmHg) and proteinuria. Labour was induced with prostaglandins. Shortly afterwards she started to feel unwell and medical staff were urgently summoned. When they attended, she was cyanosed, foaming at the mouth and unresponsive to commands. Cardiac arrest procedures were performed, but she could not be resuscitated. At autopsy the changes of eclampsia were minimal, but numerous fetal squames were found in the pulmonary capillaries.

Chapter 5: Amniotic fluid embolism (AFE)

This condition has a classical clinical presentation - sudden collapse during labour with cyanosis and rapid onset of disseminated intravascular coagulation (DIC) - which is so strongly suggestive that in the last triennial Report cases were included in this category that were not confirmed at autopsy. Similarly in this Report deaths have been attributed to AFE that have not been substantiated at autopsy.

Reasons for this vary. Some autopsy reports are so bad that no reliance can be placed on their findings. In others a search for AFE was apparently made and none found but it is unclear what techniques were used or how diligent was the search. In other instances survival in ICU raises doubts about how long fetal squames persist in the maternal pulmonary circulation. Finally, because AFE can masquerade as intrapartum/postpartum haemorrhage it is rare to encounter an autopsy report where this diagnosis has been carefully excluded from the causes of massive uterine haemorrhage.

An obese severe asthmatic who weighed over 100 kg had an elective caesarean section at 36 weeks but had a cardiac arrest just after the placenta was delivered. Clinically, amniotic fluid embolus was suspected.

The very brief autopsy report failed to address any of the issues. The heart was enlarged at 500 g but no separate ventricular weights were undertaken and the macroscopic description was terse. No specialist opinion was sought. Clots and remnants of amniotic fluid membranes were identified but there was no documentation of any damage or tears in the uterus. There was no histological examination of any organs apart from the lungs. Although it was stated that there were no fetal squames in the maternal circulation, no special techniques were employed and there was no evaluation of the changes of asthma in her lungs, nor was any search for DIC made. These factors create a lack of confidence that AFE has been satisfactorily excluded.

An older woman with known uterine fibroids was admitted at 24 weeks with severe right iliac fossa pain, thought to be due to infarction of her fibroids. After two days the pain had settled and she was discharged, only to be quickly readmitted the same day with increasing pain. The following day she collapsed, became cyanosed and developed coagulopathy. She had a spontaneous abortion and died 24 hours later in Intensive Care .

Autopsy showed evidence of DIC macroscopically and the lungs were oedematous and congested. The infarcted uterine fibroid was confirmed and some tears were noted in the cervix. Histology confirmed DIC and 'probable' fetal squames were found in occasional pulmonary capillaries. Given the gestation and the unusual clinical presentation it would have been appropriate to confirm the fetal squames in the maternal circulation immunocytochemically.

A parous woman was admitted at term + 10 days for induction of labour. Following rupture of membranes fetal distress was noted and emergency caesarean section was performed. The fetal head was jammed in the pelvis. During the section blood came out of the urinary catheter and there was vaginal bleeding but initially it was not certain whether this was due to trauma to the urogenital tract or to DIC. The bleeding could not be controlled so eventually total hysterectomy was performed. After this the patient was transferred to Intensive Care where DIC was confirmed but despite supportive measures, she died the following day.

A detailed autopsy was performed which included detailed comment on the absence of surgical trauma, review of the resected uterus and extensive histology on a wide range of organs. Histology confirmed the widespread DIC and confirmed the clinical diagnosis of AFE.

Modern immunocytochemical techniques are more sensitive than the standard histochemical methods. It is therefore recommended that unless fetal squames are easily found on standard histology, any search for AFE should incorporate probing the maternal lung sections with cytokeratin markers. Whilst CAM 5.2 is the most commonly used cytokeratin probe, LP34 gives stronger marking of fetal squames and a cleaner pulmonary background.

Such findings need to be interpreted against a background of knowledge about the frequency with which fetal squames can be detected in maternal lungs without the clinical syndrome normally associated with AFE. We also need to have information about the survival time of fetal squames in the lung, now that putative clinical cases are surviving.

Chapter 6: Deaths in early pregnancy

These are most commonly due to ectopic pregnancy but also include deaths from abortion before 24 weeks. Deaths from ruptured ectopic pregnancy are usually due to hypovolaemic shock from massive haemorrhage, or DIC supervening after successful resuscitation. The cause of death is therefore usually obvious at autopsy if clinically undiagnosed, or clinically apparent if resuscitation has been attempted. However, neither of these reasons excuses inadequate documentation of findings. Adequate documentation would become even more important if conservative treatment for ectopic pregnancies of less than 4 cm diameter becomes the established norm.

A multiparous woman presented to A&E in a collapsed state, and asystole rapidly supervened. She had not been to her GP about her early pregnancy.

At autopsy there was fresh massive haemoperitoneum and a ruptured left fallopian tube. The rupture measured over 1cm in diameter and spongy placental tissue was present in the pelvis near the tube.

There are occasions when complications, particularly after abortion, demand a more detailed autopsy:

A young mother had several episodes of threatened miscarriage and re-presented to the hospital at 17 weeks with further bleeding. She miscarried 9 hours later and a foul smelling discharge was noted. At this stage she was tachycardic, jaundiced, oliguric and anaemic. Treatment for DIC was initiated and evacuation of putrid placental tissue was undertaken but death occurred some hours later.

Autopsy included detailed histology and microbiology. There was no histological evidence of DIC. Swabs taken from the uterine cavity, and tissue taken from several organs grew Escherichia coli.

Unfortunately this is not always the standard provided:

A multiparous woman had retained products evacuated after an incomplete miscarriage at 10 weeks. Four days later she was re-admitted with a three-day history of shortness of breath and leg swelling. She was hypotensive with a tachycardia but her temperature was normal. Intravenous antibiotics were started. She was noted to be jaundiced with a falling platelet count. Subtotal hysterectomy for suspected gangrene was undertaken and the patient was transferred to ICU but she died the next day.

At autopsy the lungs were heavy and airless and the liver and spleen were grossly enlarged. Despite the clinical history there was no attempt to culture any tissues. Histology confirmed widespread changes of septicaemia in the heart and kidneys and there was liver "necrosis" which was not further defined.

It is necessary to repeat the recommendation from the last Report that it is inadequate not to have taken tissues and swabs for culture and not to have undertaken extensive histology in such cases.

Chapter 7: Genital tract sepsis

There were 16 deaths from genital tract sepsis excluding abortion. Of these, three were post-surgical and 11 occurred after spontaneous delivery.

Streptococcal septicaemia caused 10 deaths. Because of the speed with which it progresses, a high index of clinical suspicion, rapid laboratory responses and vigorous therapy are needed to combat it. This in turn demands careful and thorough autopsy when death does supervene.

A primigravid patient presented at 37 weeks by ultrasound dates with a "concealed" pregnancy. She was normotensive and fetal movements and heart beat were present. She was admitted next day with abdominal pain and evidence of intrauterine death, which was confirmed by ultrasound. DIC rapidly supervened and labour was induced by ARM and oxytocin. The liquor was blood stained. The patient developed a mottled appearance with an increased respiratory rate and poor urine output. A clinical diagnosis of AFE was made. Because of poor progress in labour an emergency caesarean section was successfully performed with minimal postoperative bleeding and the patient was returned to ICU. Despite further supportive measures her condition inexorably deteriorated and she died that evening.

Autopsy showed features of DIC with extensive petechial haemorrhages throughout all organs. Massive bilateral adrenal haemorrhages were found and the spleen was enlarged. Extensive histology confirmed the DIC with haemorrhagic infarction of adrenals and liver but no amniotic fluid embolus was found in the lungs. However, there were microabscesses in the myocardium with Gram-positive cocci present, and Group B streptococci were grown from a post-mortem intra-uterine swab.

Other infections can be as deadly:

A woman in her thirties had an intrauterine death after an antepartum haemorrhage at 26 weeks of pregnancy. Overnight she collapsed and was transferred to Intensive Care. She died 24 hours later.

At autopsy there was evidence of a bleeding diathesis but the major abnormality was an enlarged uterus filled with blood and with numerous bubbles of gas throughout the endometrium. Histology confirmed the presence of DIC and there were numerous colonies of Gram-positive cocci throughout the necrotic myometrium, confirming the growth of Clostridium perfringens on culture.

Necrotising fasciitis can complicate both surgical and non-surgical genital tract sepsis:

A woman was admitted to hospital 12 days postpartum with several days' history of pain in the left leg and back. She had had a spontaneous vaginal delivery without complications and had not required an episiotomy. A few days prior to admission there had been an offensive blood-stained vaginal discharge. On examination she was shocked and a diagnosis of pelvic sepsis with septicaemia was made. Further investigation suggested necrotising fasciitis extending down the left thigh associated with ß haemolytic streptococcal infection. Despite aggressive antibiotic therapy combined with hyperbaric oxygen her condition deteriorated with extensive necrotising fasciitis, renal failure and DIC. Five days after admission subtotal hysterectomy was performed. Despite further aggressive treatment her condition relentlessly deteriorated and she died three weeks after delivery.

The autopsy macroscopically confirmed the extensive necrotising fasciitis together with evidence of septicaemia but there was no confirmatory histology or microbiology. Review of the uterus, however, did confirm extensive haemorrhagic infarction with a minimal inflammatory response, thus implicating the genital tract as the source of the infection.

Chapter 8: Genital tract trauma and other Direct deaths.

In cases of genital trauma it is important to provide a detailed description of the injuries, to identify the site of the placenta and to exclude amniotic fluid embolus.

A multiparous woman had a trisomy-13 pregnancy but declined termination. An intrauterine death occurred and labour was induced with prostaglandins at 38 weeks of pregnancy. During labour there was evidence of a ruptured uterus, with DIC rapidly supervening. Emergency hysterectomy was required and large volumes of blood were transfused. She was transferred to an ICU and subsequently required a second laparotomy for ureteric damage. She gradually deteriorated and showed signs of diabetes insipidus before death. Cause of death was given clinically as septicaemia.

Nothing of this complicated history was recorded in the pathology report. The description of the urogenital system was very brief and there was no attempt to either localise or describe the extent of the ureteric damage. Despite the history of diabetes insipidus the pituitary was not described. The liver was twice the normal weight and was described as fatty, but subsequent histology was described as normal. This discrepancy was not explained. No histology of the pituitary was taken. In the description of the cardiovascular system moderate atheroma was identified, but not detailed. Despite the clinical diagnosis of septicaemia, no cultures or attempts to localise the source of infection were made.

A young woman was admitted at 41 weeks for induction of labour. Fetal distress developed and a forceps delivery was attempted. This was unsuccessful and an emergency caesarean section was performed. Following the operation there was blood in the urine and from the vagina. The abdominal incision was reopened and tears were found in the bladder, upper vagina and lower uterine segment. Cardiorespiratory arrest occurred, with successful resuscitation. The tears were repaired but within an hour there were signs of DIC. She was transferred to Intensive Care but over the next few days had an intermittent pyrexia, became increasingly difficult to ventilate and haemodynamically unstable, and died.

Autopsy showed grossly abnormal lungs consistent with adult respiratory distress syndrome, and there were vegetations on the pulmonary valve. However, there was no detailed description of the trauma to the genital tract, no histological confirmation of the macroscopic findings, no exclusion of AFE and no culture of the pulmonary valve vegetations.

Chapter 10: Cardiac Indirect deaths

Congenital

There were ten deaths associated with congenital cardiac anomalies. Two were complicated by endocarditis clinically but autopsy was not performed in one. The other autopsy was conducted to a high standard and confirmed the clinical findings. Two deaths were due to pulmonary hypertension complicating Eisenmenger's Syndrome but the autopsy report was available for only one of these. This was very good and detailed, clearly identifying the extensive plexogenic arteriopathy with fibrinoid necrosis in the lungs. The report on the second was not available for assessment though a letter from the consultant obstetrician states that there were no unexpected findings at autopsy. A third death was from primary pulmonary hypertension of long standing, but again no report was available for assessment.

Autopsy was of value in two cases in which sudden unexpected death occurred. In the first case there was a grossly enlarged heart which was histologically confirmed as hypertrophic cardiomyopathy. In the second there were anomalous coronary arteries each of which rapidly subdivided into a leash of very small vessels from a normal ostium.

Acquired

Myocardial Infarction

There were six deaths related to myocardial infarction/ischaemic heart disease. A consistent pattern of heavy smoking in women in their thirties emerges although details of family history and blood lipids are rarely known. All deaths were sudden and unexpected apart from one patient who had central chest pain radiating into the left arm whilst in hospital and died six days later.

A woman in her twenties had a salpingectomy for an ectopic pregnancy. Ten days later at home she suddenly woke up in the early hours of the morning with an apparent panic attack. She died at home shortly afterwards.

Autopsy showed a 0.5 cm long atheromatous plaque producing a 70-80% stenosis of the lumen in the proximal left anterior descending coronary artery. Fresh thrombus occluded the residual lumen and this was histologically confirmed.

Although autopsy clearly demonstrates the cause of death in these cases with thrombotic occlusion or severe atherosclerotic stenosis of an artery, in two instances histology to date the myocardial infarction would have been helpful. One of these has already been described, the other is described below:

An older woman had a termination of pregnancy at 12 weeks. There were no complications. Five days later she was taken by ambulance to A&E with chest pain and headache but died.

At autopsy there was severe coronary artery atheroma with an occlusive small thrombus in the left anterior descending artery and evidence of myocardial infarction. Histology would have helped date the timing of the infarct.

Other cardiac

There were eight deaths attributed to cardiomyopathy or to myocarditis; half came to autopsy. Of the four available autopsy reports two are inadequate:

A primipara in her thirties, with a long history of asthma, developed breathlessness and cardiac failure a few days after delivery. She was treated medically for about six weeks, but was then transferred to a cardiac transplant centre for assessment. She arrived shocked with impalpable peripheral pulses and died some three days later.

The autopsy (conducted by a visiting pathologist) made little attempt to correlate findings with the clinical history, and no histology of the organs apart from the heart was taken. Even that histology was not reported on, as the consultant cardiologist commented that "we did not get any opportunity to have further details of the postpartum cardiomyopathy pathology". Perhaps this was because a copy of the autopsy report was only received at the hospital 41Ž2 months after the post-mortem was performed.

A primipara in her twenties was well until 37 weeks when she became breathless. She was admitted to hospital the next day, and after investigation by a cardiologist, was diagnosed as having cardiomyopathy. She was anticoagulated, labour was induced and ventouse delivery was effected followed by transfer to ICU. She was making good progress when she suddenly collapsed and died two weeks later.

Autopsy showed death to be due to a massive saddle pulmonary embolus, but did not record the presence, or absence, of earlier emboli. The heart was grossly abnormal at 554 g and right ventricular hypertrophy was mentioned but neither ventricle was separately weighed. Atheroma was said to have produced a 40% narrowing of the left anterior descending coronary artery. Despite all this gross pathology, no histology was performed and no cardiac pathology opinion was sought. Whilst it seems likely that this case is a cardiomyopathy decompensating during pregnancy and complicated by a terminal pulmonary embolus, the quality of the autopsy ensures that this will remain speculative.

Aneurysms

Autopsy clearly established the cause of death and its source in the seven cases of ruptured aneurysm. Two were dissecting aneurysms of the aorta, two of the coronary arteries and one each of the splenic, renal and superior mesenteric arteries. However, although histology was apparently often performed on the aneurysm it was rare for this report to be included with the documents, so that identification of any systemic arterial pathology is difficult to assess. Two cases illustrate how it should be done:

A parous woman in her thirties presented at 39 weeks with severe epigastric pain. Labour was induced and the pain spontaneously resolved only to recur two days postpartum. The patient suddenly collapsed and died shortly afterwards. There was a strong family history of dissecting aortic aneurysm.

The autopsy was detailed and a careful search was made for the stigmata of Marfan's syndrome, which were absent. Haemopericardium from a dissecting aortic aneurysm was found. Cystic medial necrosis of the aorta was identified histologically and the findings were confirmed by a second opinion from a cardiac pathologist.

A woman who had had a renal transplant but who was still hypertensive, had premature rupture of membranes at 32 weeks and had an emergency caesarean section. She was well until the sixth postpartum day when she developed severe abdominal and back pain needing morphine. Ten hours after onset of the pain she suddenly collapsed and died.

Autopsy was detailed, including identification of a high arched palate, and detailed description of the placental bed site and the renal transplants, as well as the haemopericardium from the ruptured dissecting aortic aneurysm. Histology, which was extensive, confirmed the myxoid degeneration of the aortic wall, and fibroblast culture was attempted.

Chapter 11: Other Indirect deaths

These cover a miscellany of diseases, and in many of these the relationship and interaction of the disease with pregnancy is uncertain.

Epilepsy

The highest proportion (10) of the 19 deaths from epilepsy recorded in this Report occurred in the last trimester: almost all were unwitnessed sudden deaths with the diagnosis made at autopsy. Although there was usually a strong clinical history of epilepsy, two patients had not had fits in the previous two years and one other had only had her first minor fits during her pregnancy.

An epileptic patient who was under the care of her GP only, was having increasing numbers of fits from 30 weeks' gestation onwards. Admission to hospital was sought after one series of fits but was refused. Later that day she was found dead on the floor of her bedroom.

Following autopsy, death was attributed to aspiration of vomit due to epilepsy but there was no toxicology undertaken and there was no histology of any tissues. It is therefore unknown whether or not the patient was taking the prescribed medication, whether or not this produced adequate therapeutic levels and whether or not there was any other potential causes of death.

A primipara was a known epileptic with a history of depression and multiple attempts at suicide by drug overdose. She had her therapy changed from phenytoin to carbamazepine during her pregnancy. Drug levels were within the therapeutic range three weeks before death. When 36 weeks pregnant she had a stomach upset and she was found dead in bed the next day.

None of the previous medical history or features of drug therapy were identified in the post-mortem report and, although histology was performed on the heart, lungs, liver and kidneys there was no histology on the placental bed or placenta. More particularly, no drug levels were ascertained either to establish therapeutic levels at the time of death or to exclude overdose as a cause of death even though death was attributed to epilepsy.

Another primipara had been epileptic from a very early age and averaged one grand mal fit a month. When she became pregnant she halved her dose of anticonvulsants with the agreement of her GP. Even on this reduced regime she was having fewer fits whilst pregnant. She was found dead in the bath one day when her husband returned from work.

Autopsy confirmed features consistent with death by drowning. There was no trace of any anticonvulsant in the post-mortem toxicological analysis, suggesting that the patient had inadvertently, or deliberately, not taken her medication. Detailed neuropathological examination of the brain failed to demonstrate any specific seizure-related pathology and there was no evidence of any other specific pathological process on histological examination of many organs.

A known poorly controlled insulin-dependent diabetic who also had epilepsy was found dead in bed at 12 weeks' gestation. Epilepsy was suspected as the cause of death.

At autopsy, post-mortem blood levels of valproate (37.3 µg) were only marginally below the therapeutic range (40-100 µg) whereas vitreous humour glucose was significantly reduced, being less than 1 mmol/l. Hypoglycaemia was therefore clearly incriminated in the cause of death.

Obviously the role of the autopsy is crucial in establishing the cause of death in these patients. It must therefore be considered an essential part of the autopsy to exclude other causes of fits, such as eclampsia, and to establish the therapeutic drug levels in the body at the time of death.

Phaeochromocytoma

Phaeochromocytoma must be considered here, particularly as some of these deaths presented as other obstetric catastrophes. There were five deaths from phaeochromocytoma and these were almost invariably autopsy-based diagnoses:

A woman who had been normotensive during her pregnancy had an elective caesarean section at term, but shortly afterwards had a cardiorespiratory arrest. Pulmonary embolus from DVT or amniotic fluid was suspected.

At autopsy an infarcted tumour of the right adrenal was found. Viable portions of the tumour were histologically identifiable as a phaeochromocytoma.

Another woman had an episode of sudden, severe headache lasting about two hours during spontaneous vaginal delivery of a full-term infant. Eight days later she was admitted with a 12-24-hour history of very severe, sudden onset headache. She was normotensive and investigations were inconclusive. Another four days later she had another episode of severe headache, and some hours after this she had a witnessed tonic convulsion and went into a deep coma. Evidence of DIC and renal failure developed and she died in Intensive Care a day later.

A thorough, detailed autopsy showed changes consistent with DIC and revealed a 3 cm diameter infarcted left adrenal tumour. Neuropathological examination of the brain showed acute hypoxic/ischaemic changes consistent with an episode of severe hypertensive encephalopathy.

Other causes

There were a small number of deaths from infections that may be regarded as opportunistic. Some complicated diseases which are known to reduce immunocompetence but in other instances the fulminating infection may have been the result of the pregnancy.

A woman suffered from diabetes and renal failure with nephrotic syndrome. She also had cataracts, peripheral neuropathy and asthma. During the last trimester she was investigated for anaemia, tachycardia, pericardial effusions and a peripheral blood eosinophilia: a diagnosis of Churg-Strauss syndrome had been made. At 34 weeks she went into spontaneous labour and had a normal vaginal delivery. Eight days postpartum she collapsed with abdominal pain and then had a cardiac arrest. She was transferred to ICU and a subarachnoid haemorrhage was found on CT scan. She developed peritonitis through her peritoneal dialysis and died 27 days postpartum.

There was an excellent and very detailed autopsy with macroscopic evidence of peritonitis and an abscess in the myocardium. Papillary necrosis of the kidneys was also present. There was no growth on culture of swabs and tissues but hyphae of Candida were found in the heart and peritoneum on histology. There were extensive diabetic complications but no evidence of a vasculitis was found.

A young primipara was in the last trimester of her pregnancy when she presented to A&E with fever, a distressing cough and a three-day rash. A diagnosis of varicella pneumonitis was made. Intravenous acyclovir was started together with broad spectrum antibiotics. Evidence of DIC rapidly developed, she continued to deteriorate and emergency caesarean section was undertaken. Cardiac arrest occurred during the procedure and attempts at resuscitation were unsuccessful.

An excellent detailed autopsy confirming the disseminated lesions of varicella virus was supplemented by extensive microbiological and virological culture of tissues as well as histology. Varicella virus was confirmed by immunocytochemistry and by PCR techniques on the histological samples. Lymphoid tissues were reactive with no evidence of any underlying immunodeficiency.

These cases illustrate the positive contribution of an autopsy to assessing the circumstances giving rise to a maternal death.

Chapter 13: Fortuitous deaths

Whilst many of these deaths are of no consequence to the Enquiry, it is important that the survey is comprehensive. The difficult "borderline Indirect" cases, particularly infections and tumours, have already been highlighted. Even when death is Fortuitous, useful information relevant to pregnancy may be available for analysis. Deaths by suicide are an obvious problem area as puerperal depression may have been a factor. There are also several deaths from road traffic accidents in this report and many of these are from acceleration/deceleration injuries. What is often not known is whether or not a seat belt was being worn and whether some of the injuries, particularly to the large gravid uterus, were seat belt-associated or not.

As the circumstances of these Fortuitous deaths are so varied it is difficult to formulate precise recommendations, except to re-emphasise the importance of the College Guidelines in the conducting and reporting of autopsies.

For this Report, however, we would recommend that the information on the use of seat belts is recorded as accurately as possible.

Reference

1. Guidelines for Post Mortem Reports, 1993. Royal College of Pathologists, London.

ANNEX 1 TO CHAPTER 16

RECOMMENDATIONS FOR PATHOLOGISTS

The following abbreviated guide to the requirements for a maternal death autopsy should be regarded as supplementing, and not replacing the Guidelines issued by the Royal College of Pathologists¹. If in doubt, advice and help should be sought from the local CEMD Regional Pathology Assessor : the current list of Assessors can be found at the end of this Report.

In general:

Maternal deaths are still under-reported and pathologists performing autopsies on women who are pregnant or are known have been pregnant within a year of their death should contact their local Director of Public Health (DPH) to check that the case has been reported.

Clinical information for the Enquiry is sometimes incomplete and pathologists should provide a review of the clinical history in autopsy reports, including height and weight of the woman.

Specific disorders

Hypertensive diseases

Check and note existence of local guidelines for the management of hypertensive disorders of pregnancy.

Identify fluid balance; minimum histology of lungs, liver, kidney, brain, placental site; exclude previous hypertension.

Thromboembolism

Identify any local risk factors protocol, note family history, significance of chest symptoms, heparin prophylaxis.

Describe the nature and distribution of emboli; site of origin; evidence of previous episodes; histology.

Haemorrhage: APH and PPH

Macroscopic: Identify the site and severity of bleeding; location of placenta; detail genital tract trauma.

Histology: placental histology; search for DIC; exclude AFE; review other tissues resected.

Early pregnancy

Ectopics: diagnostic awareness, ultrasound monitoring and diagnosis; Location and size of ectopic; estimate blood loss; review other tissues resected.

Abortions: sites and locations of bowel perforations; culture of tissues.

Amniotic fluid embolism

Macroscopic: detailed examination of genital tract for trauma.

Histology: detailed histology of both lungs; immunocytochemistry for cytokeratins if in doubt.

Women dying after labour of causes other than suspected amniotic fluid embolism should have their lungs examined for amniotic squames to check whether amniotic fluid can enter the circulation without a fatal outcome.

Hyperemesis

Exclude acute Wernicke's encephalopathy.

Epilepsy

Macroscopic: exclude specific brain pathology; establish anticonvulsant drug levels.

Histology: exclude eclampsia as cause of fits.

Cardiac deaths

Macroscopic: full description of heart; weigh/measure ventricles separately.

Histology: both ventricles; assess conducting system; seek cardiac pathology opinion if in doubt.

Aneurysms

Macroscopic: nature and site of aneurysm.

Histology: distribution of arterial pathology.

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