In 1998, 41 cases of maternal death, due to acute collapse or embolism, were reported. These cases comprise 11.5% of all direct obstetric deaths and 7.3% of all reported maternal deaths. There were 32 cases of acute collapse with the majority of these women dying from cardiac arrest. Thrombotic pulmonary embolism occurred in seven cases, whilst amniotic fluid embolism occurred in two cases. The patients were approximately one decade older than the other maternal deaths forming the database (p=0.0146). There was no difference between levels of care at the time of death. The proportion of patients dying as a result of acute collapse and embolism is similar in all provinces, except for KwaZulu-Natal where less patients were reported to have died in this category. In approximately half of these patients no avoidable factors could be identified, but poor record keeping inhibited analysis. Other problems included an inability to identify serious underlying conditions and poor monitoring of sick patients in hospital. Recommendations include improved clinical evaluation and examination of patients, better record keeping, the categorisation of acute collapse as well as acute respiratory decompensation as emergency conditions and recognition of a painful and/or swollen leg as a danger sign for possible deep vein thrombosis and embolism. If the survey could be extended to include more home deaths, more cases of thrombotic pulmonary embolism would probably be recorded.
Key Recommendations
Acute collapse signifies an acute event involving the brain, heart or lungs and may ultimately result in death. It is, however, a non-specific term given that adequate clinical investigation will uncover the actual event leading to death. For example, cerebral haemorrhage or cardiac arrhythmia may result in heart arrest. It is therefore not surprising that in many cases inadequate documentation may lead to the diagnosis of "acute collapse". Furthermore, poor monitoring of patients results in the early warning signs for disaster not being detected. When the problem is detected, the patient is often in extremis and little can be done. In these cases the collapse has not really been acute, but is the end result of a long process.
Embolism, on the other hand, is a specific diagnosis resulting in most cases from deep vein thrombosis. The quality of documentation in this case is superior to that of acute collapse even though the diagnosis is often made retrospectively and based on clinical evidence with no confirmation by special investigation. Approximately 22% of patients with pulmonary embolism were confirmed by special investigation. Deep vein thrombosis usually involves the large pelvic veins or veins of the leg resulting in a painful, swollen leg. Embolism usually manifests as pulmonary embolism with acute respiratory decompensation or arrest.
Amniotic fluid embolism is an interesting condition whereby the amniotic fluid in the uterus enters the large veins ultimately causing pulmonary embolism. The same situation may be caused by, for example, air entering the opened uterus during a caesarean section. These conditions are rare and although no case of air embolism was reported in 1998, two cases compatible with amniotic fluid embolism were noted.
In addition to the large proportion of cases with inadequate clinical notes, many Assessor forms were not completed. Furthermore, the relatively small number of cases documented as "acute collapse" or "embolism", make meaningful data analysis difficult. The 41 documented cases comprise only 11.5% of all direct obstetric causes of death and only 7.3% of all reported maternal deaths occurring in 1998.
These 41 cases consist of 32 cases of acute collapse and 9 cases of embolism (7 of pulmonary embolism and 2 of amniotic fluid embolism).
The distribution of cases is shown in Table 8.1. It is interesting to note that although the largest number of deaths were reported in KwaZulu-Natal (n = 188), only 3% of these were said to be the result of Acute collapse/Embolism. It may be that women, suddenly falling ill at home are not able to reach a hospital before dying and that these deaths are therefore not reported as institutionalised maternal deaths. Both transport problems and limited reporting of deaths could have played a major role.
The reported number of women who died does not allow for detailed demographic analysis. The relevant data is shown in Tables 8.2 to 8.4.
The maternal deaths primarily diagnosed as acute collapse or embolism seemed to be relatively older when compared to the rest of the maternal deaths. A statistically significant proportion of these women were aged 30 years or older (Chi squared - 5,97, p = 0,0146). Approximately 55% of women with acute collapse or embolism were either P0 or P1 at time of death, comparable to the remainder of the sample (57%).
There is no statistical difference between the 3 levels of care.
The number of women delivering at the various levels of care is unknown, but it is assumed that a large number of deliveries are managed by level 1 and level 2 hospitals. Therefore, the proportion of deaths due to acute collapse/embolism per number of deliveries may be higher in level 3 hospitals than level 1 or 2. Limited reporting, however, adversely affects these results.
A definite diagnosis of embolism was made in only 22% of cases. It is probable that embolism was the primary cause of death in a significant proportion of the remainder of cases. Acute pulmonary oedema from undiagnosed cardiac lesions or myocardial infarcts could also have played a part. Autopsies in cases of acute collapse would be very useful. Poor monitoring resulting in the early warning signs of disaster not being detected can also lead to patients being discovered extremely ill, leading to the erroneous diagnosis of acute collapse.
Cardiac arrest was the final cause of death in the majority of cases. However, with optimal investigation and adequate recording, the actual cause may have been identified in most of the cases admitted to hospitals or clinics, for example hypovolaemia or multi-organ failure. (See Table 8.5).
The numbers are small and deductions made from the table are to be evaluated accordingly:
Short case descriptions and general problems that were encountered are illustrated below. These cases highlight commonly encountered problems.
Case 1: A note on the diagnostic evaluation reads as follows: "Correct diagnosis was made in referring hospital, proper management and transfer arranged but specialist unit misdiagnosed the condition and sent the patient back without treatment, resulting in a repeat massive embolus, this time fatal. Poor facilities in the ambulance (no suction, oxygen) and poor communication between specialists and referring doctor (who consulted appropriately throughout and behaved most responsibly) were the most obvious problems".
Comment: Referrals should always be seen in a serious light although unnecessary referrals do occur.
Case 2: A patient in labour (cervix 4 cm dilated) with normal vital signs and meconium stained liquor, collapsed and fell from the examination couch, lost consciousness, had a cardiac arrest and could not be resuscitated.
Comment: A possible diagnosis could have been cerebral haemorrhage, myocardial infarct or amniotic fluid embolism, but as no autopsy was done, a definite diagnosis was not made. This case stresses the need for autopsies to be done where possible.
Case 3: A grande multiparous woman (G5P4) who was a known diabetic (on insulin) and who was not referred for specialist care, had a complicated delivery (shoulder dystocia) and died 15 hours after delivery during which time minimal observations and care were provided.
Comment: Clear indications for referral to a higher level of care should be available to all facilities conducting deliveries. The lack of an accurate diagnosis here is probably related to poor monitoring. Death could have been due to postpartum haemorrhage, embolism or diabetic ketoacidosis. Appropriate monitoring could have detected a problem much sooner, and allowed for appropriate intervention.
Case 4: Patient was admitted hours after an uneventful delivery with complaints of a painful leg and pyrexia. She was discharged after 3 days but suddenly collapsed on Day 18 after complaining of chest pain. Was dead on arrival.
Comment: Deep vein thrombosis could have been missed during hospitalisation. Evaluation of postpartum patients for possible risk factors (such as a painful leg) before discharge is imperative.
Case 5: The patient had a caesarean section and subsequently died. The assessor could find no record of daily postnatal assessment. Comment: Several assessors mentioned that clinical notes were either incomplete or not available. The writing of complete, up to date clinical notes is very important.
Case 6: The patient had a caesarean section after prolonged labour. Intrapartum foetal death had already occurred. Post-caesarean section she had persistent tachycardia and tachypnoea (HIV positive). She was first seen by doctor 14 hours post surgery and a single dose of metronidazole was prescribed. No proper clinical assessment appears to have been made.
Comment: This patient was not regarded as being high risk and therefore, she was managed at a substandard level. In obstetrics, it is of particular importance to identify risk factors followed by timely and appropriate management.
Case 7: Post-caesarean section patient who complained of pain and nausea 18 hours after surgery. Pethedine was administered and the patient died. According to her documentation she had been shocked for 6 hours prior to death, but the staff had not identified the problem.
Comment: Unprofessional conduct regarding management of intrapartum and postpartum events are apparent here. A patient who was shocked and probably hypoxic was given a sedative, which suppressed respiration and caused death. The shock was not recognised and no appropriate treatment was given.
Case 8: G4P3 was in obstructive labour and could not be referred to secondary hospital for caesarean section as it was supposedly full. While waiting to be transferred to another hospital, she suddenly collapsed and died. Assessment and monitoring during the intrapartum period was not optimal.
Comment: Clarity on the doctor's responsibility in the receiving hospital is urgently required. Emergency cases should always be admitted, even if a hospital is supposedly full. This woman probably died of uterine rupture.
Since "acute collapse" is a collective term for several ill-described conditions resulting in death, it is not surprising that in this section many cases of poor documentation have been identified. The clinical notes as well as the assessors' notes were in many cases incomplete. However, this diagnosis was made in 7.3% of all maternal deaths. Although it may seem to be a minor cause of death, poor documentation has been mentioned in several other chapters and should therefore be regarded as an important problem.
Although it was reassuring to note that in approximately half of the cases, no avoidable factors could be identified, this finding must be evaluated against the background of poor recording. Avoidable factors occurred more often in the emergency room or during an emergency event, particularly during the intra- and postpartum periods. In addition, poor monitoring of acutely ill patients was identified as a major problem. The biggest problem, however, was a lack of identification of a severe underlying condition, particularly shock or other forms of decompensation. Insufficient resuscitation or treatment was administered, leading to acute collapse and death. This occurred at all three levels of service. The solution lies not only in better training, but also in improved examination of patients, documentation of the findings and subsequent monitoring.
Thrombotic pulmonary embolism occurred infrequently, although, the United States of America has reported embolism to be the major cause of death in women whose pregnancies resulted in a live birth1. Limited reporting may be a major factor since pulmonary embolism tends to occur several days after birth usually when the patient has already been discharged. Following surgery, it tends to occur between the 15th and 30th postoperative day1. Since this enquiry into maternal deaths is almost exclusively limited to institutional deaths, home deaths are most probably inadequately reported.
Several cases have been recorded where possible deep vein thrombosis was not diagnosed, even given a painful and/or swollen leg. This should be viewed as an emergency condition necessitating adequate treatment. Some of these cases duly led to pulmonary embolism and death.
Acute collapse and embolism occurred infrequently and meaningful conclusions are difficult. However, these conditions can be minimised by examining patients more thoroughly, by improving clinical documentation and close monitoring of patients in hospital. Obtaining autopsies would also help significantly in making accurate diagnoses. Furthermore, an attempt should be made to investigate home deaths as pulmonary embolism may be a particular problem in this respect.