Safe obstetric anaesthesia in low- and middle-income countries—a perspective from Africa

Learning objectives.

By reading this article, you should be able to:

• •Discuss the impact of resource limitation on the practice of obstetric anaesthesia.
• •Identify patients at high risk for adverse maternal outcomes in these settings.
• •Outline key areas where context-sensitive management might improve maternal outcomes.

Key points.

• •Perioperative maternal mortality rates are significantly higher in low- and middle-income countries than in high-income countries.
• •Evidence-based practice from these settings is limited and data are lacking.
• •The lack of trained anaesthetists is a key barrier to safe obstetric anaesthesia.
• •Identifying high-risk patients in advance may enable better use of limited resources.
• •Context-sensitive management may prevent complications in an accessible and safe manner.

Maternal and neonatal outcomes in low- and middle-income countries (LMICs) remain poor, despite gradual improvement. The global maternal mortality ratio (MMR) has declined from 339 to 223 deaths per 100,000 live births during the period 2000 to 2020, an annual reduction of 2.1% per year.¹ However, the MMR in sub-Saharan Africa in 2020 was 545 (from 807 in 2000), with a lifetime risk of maternal death of 1 in 40 (compared to 1 in 16,000 in Australia and New Zealand). In 2020, sub-Saharan Africa still accounted for 70% of the global mortality. These poor outcomes are the result of complex and multifactorial health systems deficiencies. A recent review in this journal outlined the causes and contributors to higher MMRs in LMICs.² It discussed elements that are still relevant today, including the 3-delays model (transport, training and treatment), challenges related to systems and infrastructure, and the choice of anaesthesia for Caesarean section (CS), and methods for labour analgesia.²

The safe provision of anaesthesia remains a key goal to improve maternal and neonatal outcomes, although it is just one aspect within a complex, interconnected and interdependent obstetric healthcare system. There may be marked differences between the levels of accessible care even within LMIC. For example, in South Africa major referral hospitals provide suitable care for complex obstetric anaesthesia, but in peripheral areas, there are often inadequate anaesthesia skills and resources to provide routine care during CS. Labour epidural rates remain unacceptably low in state hospitals, even in referral centres.³ Similarly, there may be considerable differences in resources between countries classified as LMIC. This article focuses on specific anaesthesia practices that may contribute to improved outcomes, rather than the systems-level issues covered previously in this journal.² Although we concentrate on Africa, we believe that the underlying principles and contextual similarities make these suggestions globally applicable to resource-limited settings.

There is a lack of evidence from LMIC contexts

Literature informing practice in LMICs originates from vastly different contexts.⁴ Although this is to be expected given resource constraints that may preclude labour-intensive research, it highlights the need for context-sensitive solutions developed in conjunction with clinicians in that setting. We will discuss clinical anaesthetic approaches to LMIC concerns, based on the limited research available. We will also provide an update on anaesthesia-related morbidity and mortality in LMICs, including data from the South African National Committee for Confidential Enquiry into Maternal Deaths (NCCEMD).^5,6

Maternal outcomes are poor in LMIC

Many high-income countries (HICs) have processes that evaluate maternal outcomes, through confidential enquiries into maternal deaths. Until recently, South Africa was the only country that conducted this process in Africa. Kenya completed its first report in October 2016 (relating to deaths in 2014), although <15% of the actual deaths were reported and only half of these were assessed. The South African and the Kenyan processes highlighted similar problems: poor antenatal care, and a high number of avoidable deaths, particularly driven by obstetric haemorrhage and hypertensive disorders. Limited analysis of anaesthesia-related outcomes was available in the Kenyan report.⁷ In contrast the 2017–2019 South African data suggested that anaesthesia contributed to 205/921 (22%) of maternal deaths after CS. In 66/921 (7%) of patients, deaths resulted primarily from complications of anaesthesia, whereas in a further 139/921 (15%), anaesthesia was contributory (mortality from another cause, such as obstetric haemorrhage, but action or inaction by the anaesthetist might have contributed significantly to the death or events ultimately leading to death).⁶ Thus anaesthesia contributed to one in five perioperative maternal deaths, often including an avoidable component such as delayed or poor resuscitation, delay in recognition of complications and inappropriate method of anaesthesia.

Poor outcomes are driven by low workforce density and resource constraints

The African Surgical Outcomes Study (ASOS) showed that many regions in Africa have inadequate workforce density, and maternal and neonatal mortality remain unacceptably high, driven largely by haemorrhage and anaesthesia complications.^8,9 Whereas the complication rate of CS is only two to three times higher in women in Africa, maternal mortality is 50 times higher than in their HIC counterparts. This illustrates the concept of ‘failure-to-rescue’: when perioperative complications occur, the ‘rescue’ of these patients requires resources that are lacking, such as monitoring, skilled practitioners, blood products and intensive care support. The prevention and early detection of complications, in tandem with context-sensitive management solutions, are therefore a key priority.

Poor maternal outcomes are compounded by inadequate CS rates, which should be up to 19%,¹⁰ but are <10% in more than half of African countries, indicating a deficiency in access to surgery.¹¹ There is also an increased morbidity and mortality that occurs in the absence of adequate numbers of trained anaesthesia providers.¹² In LMICs, obstetric anaesthesia deaths occur at rate of 1.2 per 1000, and represent 13.8% of all deaths after CS—compared with 3.8–6.5 per million in the USA. Anaesthesia is responsible for 2.8% of all maternal deaths in LMICs, with general anaesthesia (GA) resulting in a five-fold higher mortality rate than spinal anaesthesia.¹² In Malawi, poorly trained anaesthesia providers have been found to have almost three times higher maternal mortality rates compared with trained anaesthesia providers.¹³ A lack of robust data makes it impossible to determine whether non-physician anaesthesia providers (NPAPs) are associated with worse maternal outcomes than physician anaesthesia providers (PAPs) within any particular jurisdiction.¹² An inflection point regarding optimal number of anaesthetists has been described, below which maternal mortality increases precipitously.¹⁴ The authors suggest that a minimum of four specialist anaesthetists per 100,000 of the population are required to achieve a minimum standard of healthcare. Only five African countries have more than one PAP per 100,000 of the population (Fig. 1), and HICs have 90 times the number of PAPs compared with low-income countries (LICs).¹⁵

Fig 1.

The global distribution of physician anaesthesia providers per 100,000 population.¹⁵ White indicates no data. Reproduced with permission.

Despite the fact that this proportion is exceeded in South Africa, the South African confidential enquiry process confirmed that in some areas, unqualified junior doctors and clinical associates were providing anaesthesia care in unsupervised settings, and that in 2.5% of deaths the designated anaesthetist was also an assistant at surgery.⁶

Workforce deficiencies usually coexist with poor operative capacity. One study in LMICs, including three African countries, suggested that one-third of hospitals do not have access to reliable electricity, one-quarter do not have access to oxygen, three-quarters do not have pulse oximetry for routine monitoring and half of hospitals do not have a recovery area.¹⁶ A survey conducted in referral hospitals in East Africa showed that none of the hospitals had all the necessary requirements to provide safe obstetric anaesthesia (based on the availability of drugs, equipment, monitoring and anaesthesia machines), and only 7% reported adequate staffing.¹⁷ Providing guidance to inexperienced anaesthetists is challenging in the absence of these minimum standards.

Specific anaesthesia practices

Spinal anaesthesia

Single-shot spinal anaesthesia remains the safest choice for CS in most patients in LMICs. However, it is probably true that spinal anaesthesia is performed inappropriately in many patients in variable resource countries, resulting from concerns about the ability to perform a safe general anaesthetic, because of the lack of availability of trained staff or basic equipment. Data from Africa have shown that spinal anaesthesia is often performed in patients with conditions such as abnormal placentation or even a ruptured uterus, or eclampsia associated with a low Glasgow Coma Scale (GCS).⁹ Most practitioners are comfortable with performing spinal anaesthesia in healthy patients, but often cannot manage complications should they arise, whether in healthy or compromised patients.

Clinical scenario 1.

In a district hospital in Uganda, a woman is scheduled for CS as a result of obstructed labour. She is assessed by an NPAP, who decides to perform spinal anaesthesia. The patient has a mild tachycardia and is experiencing uterine contractions while the spinal anaesthetic is performed. Soon after commencement of surgery, the patient complains of numbness in the hands, and subsequently becomes drowsy. The NPAP notices a slow heart rate, but the blood pressure is undetectable, and there is no pulse oximetry available. After checking the blood pressure cuff and restarting the machine, the NPAP notices that the patient is unconscious and is not breathing. Mask ventilation is commenced, but the patient quickly progresses to a cardiac arrest. The surgeon (a general practitioner) assists with a full resuscitation. Following treatment with adrenaline (epinephrine) and fluids there is a return of spontaneous circulation, but both the mother and neonate have suffered significant neurological injury.

Spinal hypotension

Spinal hypotension is usually a consequence of sympathectomy, and rarely may result from a high spinal anaesthetic (manifesting as hypotension, bradycardia and a high spinal sensory level to the cervical dermatomes). Guidelines for the management of high spinal anaesthesia are available, which include suggestions for emergency inotrope infusions in the absence of a syringe pump (adrenaline [epinephrine] 1 mg in 1 L modified Ringer's lactate solution).¹⁸ Pattern recognition should be taught of the haemodynamic presentations of all the causes of spinal hypotension. This is particularly important where anaesthesia providers with minimal experience are practicing without supervision by a senior anaesthetist.¹⁹

It is crucial that spinal anaesthesia is given only to appropriate patients, avoiding patients in whom a sympathectomy could cause rapid decompensation, such as those with hypovolaemia, or when major haemorrhage is present or expected. Unexplained tachycardia (heart rate >120–140 beats min⁻¹) should be a relative contraindication to spinal anaesthesia and requires discussion with the best qualified anaesthesia provider, and further investigation as indicated.

The management of spinal hypotension should include fluids and early initiation of appropriate vasopressor drugs, optimally phenylephrine if available, or other agents such as ephedrine, etilefrine, metaraminol or adrenaline.²⁰ In the setting of precipitous hypotension unresponsive to phenylephrine or ephedrine, adrenaline boluses may be required (10–50 μg), in conjunction with airway management if needed.

Hypotension during routine spinal anaesthesia is both likely and predictable and should be managed proactively if possible. Given that hypotension occurs in >70% of patients and is severe in >20%, the use of prophylactic vasopressor infusions is advised. A titrated phenylephrine infusion remains the gold standard,²¹ but fixed-rate, low-dose infusions have been shown to be effective in variable resource settings in the hands of junior doctors.²² Where infusion pumps are unavailable, adding vasopressor (phenylephrine 500 μg) to the first litre of fluid after spinal anaesthesia may also be safe and effective.²³ A recent review of these practices in LMICs deals with the subject in more detail.¹⁹ Regardless of how they are given, the early use of vasopressors is recommended. Recently, there has been considerable research in HICs into noradrenaline (norepinephrine) for both prophylaxis and treatment of spinal hypotension, predominantly motivated by the mild beta-adrenergic activity of noradrenaline. However, research has been conducted predominantly in resource-rich contexts. Current evidence is unconvincing even in HICs for any clinically significant benefits of potent agents such as noradrenaline. In LMICs this agent should be avoided as clinicians are unfamiliar with the drug and unable to monitor for adverse effects. In addition, syringe pumps are often unavailable, and the risk of catastrophic dose error is significant. In the rare situation in which phenylephrine is associated with bradycardia and hypotension, small doses of anticholinergic agents or ephedrine would suffice.

Conversion to general anaesthesia

Spinal anaesthesia may require conversion to GA, either because of complete or partial failure of the block, protracted surgery, or haemodynamic instability. The decision to convert to GA because of pain should be guided by the risk associated with GA in that context. This is likely increased in proportion to the shortage of equipment available to anaesthesia providers, and their lack of training. Recent guidelines have been published for the management of intraoperative pain and include modifications for variable resource countries.²⁴

Conversion to GA either for protracted surgery or haemodynamic instability requires a higher level of anaesthesia expertise, particularly with respect to airway management. Tracheal intubation may be technically more challenging in a patient draped and prepared for surgery. Whereas exact details of the practice of GA for CS in Africa are unavailable, in Sierra Leone, the use of ketamine without intubation is commonly performed.²⁵ It is likely that i.v. or i.m. ketamine is often used without definitive airway management. Recently, a prospective case-series from Kenya analysed 401 emergency CS that were conducted by 54 non-anaesthesia providers using ketamine-based anaesthesia (ESM-Ketamine), in circumstances where no anaesthetist was available.²⁶ There were no serious adverse maternal events and all mothers survived to hospital discharge. Further data related to the prevalence and safety of the various practices used for GA in LMICs are urgently required, and currently under investigation in the ASOS studies.

Intrathecal tranexamic acid injection

Following the World Health Organization's recommendation to use tranexamic acid early in the management of obstetric haemorrhage at CS, there has been a concerning increase in the number of intrathecal tranexamic acid injections, associated with a high mortality.²⁷ Clinical presentation typically follows initial failure of the spinal anaesthesia, proceeding to a seizure and sympathetic hyperactivity, which may be mistaken for eclampsia. This arises because of a drug error caused by similarities in the size and appearance of tranexamic acid and bupivacaine ampoules, compounded by storage of the two drugs in close proximity to each other. Systems are necessary that allow for timely use of tranexamic acid, but reduce the risk of drug error. Improved procedures for checking of ampoules, and ultimately re-design of the ampoules so that they are clearly distinguishable, are recommended. Tranexamic acid should be stored in a separate location to bupivacaine, either in a sealed box containing drugs used specifically for obstetric haemorrhage, or just outside the operating theatre, where the drug can be accessed quickly for use when required.²⁷

Obstetric airway management

With the increased use of spinal anaesthesia for CS, it is possible that junior clinicians have become less proficient in the management of the obstetric airway, as a consequence of a lack of regular exposure. Hypoxaemia (Spo₂ <90%) is a relatively common event during the management of the obstetric airway in anaesthesia departments in South Africa, occurring in one in six patients.²⁸ Guidelines exist for the management of the difficult airway in obstetric anaesthesia.²⁹ Calls for the availability of a videolaryngoscope in all obstetric theatres are unlikely to be realised because of cost restraints. In addition, training in their use is in short supply. However, many of the basic principles of modern airway management are available and accessible in variable resource countries through education and training. These include the following practices²⁹:

• •Adequate preoperative assessment of the airway
• •Elevation of the head by 25°
• •Ramping (horizontal alignment between the external auditory meatus and the sternal notch)
• •Preoxygenation (end-tidal oxygen fraction >0.9)
• •Apnoeic insufflation by face mask before tracheal intubation
• •Considering early mask ventilation at pressures ≤20 cmH₂O during induction of anaesthesia, if there is a high risk of hypoxaemia, or in the event of difficult intubation

Amendments to traditional rapid-sequence induction should be taught routinely, including the early release of cricoid pressure if laryngoscopy is difficult, and the use of low-pressure mask ventilation.²⁹ Appropriate teaching of the use of the Macintosh blade is essential, and should be supplemented with online teaching and simulation where necessary. These principles are essential in ensuring that an ‘easy airway’ is not converted to a ‘difficult airway’ through deficiencies in technique and positioning. Early use of a rescue airway such as the supraglottic airway device should be encouraged, as indicated in the most recent difficult airway algorithm.²⁹

Hypertension

Hypertensive disorders of pregnancy are a leading cause of mortality in LMICs.³⁰ Late presentation and delayed treatment often compound the severity of this multisystem disorder. In preeclampsia with severe features, rapid clinical deterioration is possible, particularly where cardiovascular involvement is present. Point-of-care ultrasound evaluation of the heart and lungs has utility in the assessment of disease severity and strategic anaesthesia choices. However, although handheld devices are becoming more affordable and increasingly used in variable resource countries, this equipment and experience are unlikely to be routinely available in many LMICs. In the absence of a reliable measurement of stroke volume responsiveness, fluid restriction is likely the safer general strategy. Adequate control of the blood pressure on admission and before CS is crucial. Two further areas warrant more detailed discussion.

Is spinal anaesthesia safe in preeclampsia?

Spinal anaesthesia is safe and effective in uncomplicated severe preeclampsia, and is routinely performed in LMICs. Research has shown that during spinal anaesthesia there is usually only a modest reduction of afterload, and that preload is not the major concern in the absence of haemorrhage and other comorbidities.³¹ There are two common dilemmas for clinicians in variable resource countries related to the decision to perform spinal anaesthesia. The first is concern regarding potential thrombocytopaenia in patients with preeclampsia in the absence of recent laboratory results. Most guidelines recommend a recent (<6 h) platelet count that is ≥70×10⁹ L⁻¹, but this is often unavailable. Clinicians are then required to make a risk–benefit decision. A recent study summarises anaesthesia for preeclampsia in the setting of thrombocytopenia, and presents a decision algorithm which includes a decision aid (Fig. 2) for the choice of spinal anaesthesia or GA in an individual case.³² The authors concluded that there may be circumstances in which the clinician may justifiably opt for spinal anaesthesia when a platelet count is indicated but unavailable.

Fig 2.

Decision algorithm for spinal anaesthesia versus general anaesthesia in patients with hypertensive disorders of pregnancy and thrombocytopenia.³² HELLP, haemolysis, elevated liver enzymes, and low platelets. Reproduced with permission.

A second common concern is the choice of anaesthesia for CS after eclamptic seizures. Most guidelines recommend GA if the GCS is ≤14, with careful attention to the principles of neuroanaesthesia and a brief period of postoperative ventilation. In many LMICs, expertise for GA and facilities for postoperative ventilation are unavailable. A recent audit of anaesthesia practice in a referral hospital in South Africa showed that seven of 89 (7.9%) eclamptic patients received spinal anaesthesia and 82/89 (92.1%) GA. A significant proportion of patients whose clinical presentation would have allowed spinal anaesthesia, received GA for CS. By contrast, in an African study, most patients received spinal anaesthesia despite the presence of eclampsia,⁹ and >50% of women with eclampsia received spinal anaesthesia in the recent ASOS study. This is likely to result from the reluctance in resource-limited settings to perform GA, particularly in such high-risk patients. A major audit of anaesthesia for patients with eclampsia and their perioperative management in LMICs is needed.

Airway management in women with hypertensive disorders

Recently, an airway management registry has been established with a view to improving GA practice in obstetrics, including hypertensive disorders of pregnancy.³³ The data have highlighted that the latter patients are at twice the risk for desaturation during tracheal intubation compared with healthy patients, probably a result of interstitial pulmonary oedema. Although the airway management principles discussed earlier still apply, the identification of the higher risk patient is invaluable, to ensure that the most experienced anaesthesia practitioner is present for tracheal intubation.³⁴

Clinical scenario 2.

In a district hospital in South Africa, a 23-year-old primigravida at 32 weeks' gestation presents in the evening, after three witnessed seizures at home. Her blood pressure is 196/122, oxygen saturation is 92% on room air, and her GCS is 14/15. The attending doctor gives magnesium and labetalol i.v. for blood pressure control, but there are no laboratory services until the next morning. The referral hospital accepts the patient, but transport is likely to be delayed and the doctors are advised to perform a CS after reducing the blood pressure to ≤160/110. In light of the absence of blood results, the attending anaesthetist elects to perform a GA. Following induction of anaesthesia, there is rapid desaturation and maternal bradycardia develops. However, the doctor is able to intubate the trachea and the patient does not suffer a cardiac arrest. Following CS, the mother's GCS does not recover sufficiently for tracheal extubation. The patient is sedated and ventilated in the operating theatre overnight and a transfer is organised for the following morning.

Inability to perform general anaesthesia: guidance and training

The lack of confidence to perform a safe GA is a significant concern in LMICs. GA is often required for patients who have suffered complications or are haemodynamically unstable, requiring considerable expertise during the induction of anaesthesia. In clinical scenarios such as a ruptured ectopic pregnancy, definitive management involves laparotomy to control haemorrhage, yet in South Africa, largely in district hospitals, 38% of these patients die without surgery.³⁵ The primary reason for attempting to transfer such a patient to a referral hospital rather than stabilising her in the operating theatre at district level, is the lack of ability to provide GA.

Training deficits

Effective education and training, both for doctors and allied staff such as NPAPs, lies at the heart of improvement in anaesthesia services. The shortage of skilled staff and appropriate trainers is a significant barrier to care. Models that use referral hospitals to provide this training to district hospitals may be limited by the staff shortages at the higher tier of care. A hybrid model of outreach (where an anaesthetist visits a district hospital on a monthly basis and provides hands-on teaching in anaesthesia during a dedicated operating theatre list) and in-reach (where doctors from the district hospital attend the referring hospitals for a limited period in order to be exposed to many anaesthesia cases), may be used. This model is then supplemented by workshops and online training courses, such as developed by the SAFE-OB team, which has shown a sustained impact on knowledge and skills in Ethiopia.³⁶ In addition, teaching can be enhanced by mobile-based platforms and applications, and virtual consultations. Systems such as these, aimed at change management, have been successfully developed and implemented by organisations such as Kybele, and include a comprehensive package of support.³⁷ At a national level, the development of national surgical, obstetric and anaesthesia plans (NSOAPs) have allowed for a framework for equitable surgery to be developed in countries such as Senegal, Ethiopia, Nigeria, Madagascar, Zambia, Rwanda and Tanzania. In Tanzania, NSOAP aims to increase the number of anaesthesia providers from 0.09 per 100,000 to 2.23 per 100,000 by 2025.³⁸ They will perform training in a 1:2 ratio of anaesthetists to nurse anaesthetists, at a cost of 2 US dollars per person in the population. Implementation of these frameworks has yet to be assessed.

Allied anaesthesia skills

In addition to the clinical skills required, there are other aspects of cross-cultural healthcare that may be problematic. Language barriers may decrease understanding and increase perioperative anxiety in parturients, which could have both psychological and physiological consequences. The use of mobile-based explanatory videos has been shown to assist in the perioperative period by providing a consistent explanation in the patient's first language, improving the consent process, and decreasing maternal anxiety.³⁹

A second key area is the issue of perioperative documentation of anaesthesia. Both the South African and the Kenyan confidential enquiry processes have commented on deficiencies in this area, and a recent publication describes the shortfalls in details recorded during spinal anaesthesia for CS in South Africa.⁴⁰ There are several consequences. The first is that effective communication is hindered between multiple clinicians involved in the management of complications. Secondly, in the audit and investigation of complications and poor outcomes, it is difficult to conduct a meaningful analysis and develop appropriate guidelines and recommendations. Finally, the medicolegal defence of perioperative complications is very challenging, as the absence of notes is interpreted as inaction.⁴¹ Indefensible litigation arising from poor documentation leads to consequences for staff and the healthcare system, a situation that is crippling healthcare in South Africa. Guidelines are available on the nature and extent of documentation required and should be introduced wherever possible.²⁴

Guidelines

It is apparent that recommendations for clinical practice should be tailored to the specific context and should consider the availability and level of skill of staff, and appropriate equipment and infrastructure required for healthcare. Three recent consensus guidelines have made specific recommendations for variable resource countries, in the areas of spinal hypotension,²¹ the use of oxytocic agents²⁰ and the prevention and management of intraoperative pain during CS under neuraxial anaesthesia.^24,42 Existing guidelines are increasingly including the context-specific recommendations wherever possible, although research informing these modifications remains scarce. Most research occurs in resource-rich, controlled environments, and it is difficult to extrapolate safety data to vastly different settings.⁴³

Providing guidance is a complex matter in situations where acceptable international standards are not achieved. For example, the use of intrathecal morphine during anaesthesia for CS is considered the gold standard for postoperative analgesia in HICs. At lower doses, it has been suggested that no additional postoperative monitoring is required compared with intramuscular opioids.^44,45 However, in LMICs, several modifications might be required. Intrathecal morphine should be preservative-free, and dose errors are possible given the need for careful dilution, or poor availability of tuberculin syringes. In some areas, postoperative monitoring may not occur at all, and it is possible that patients will receive both intrathecal and intramuscular morphine unintentionally, especially when postoperative pain protocols require flexibility, given the unreliability of supplies. Preoperative clinical detection of patients with comorbidities such as obstructive sleep apnoea may be inadequate, potentially resulting in poor outcomes when dosing errors occur. In these settings, one is faced with the choice of prioritising either safety or effective analgesia; achieving both may be difficult.

Conclusions

Despite global improvements in maternal outcomes, there are still significant deficiencies in LMICs. Key priorities should include achieving an adequate number of trained staff who are competent to perform obstetric GA. Safe, context-sensitive clinical practices should be prioritised. Systems to train and support these clinicians should be developed in the local environment. There should be an opportunity to receive additional training in referral hospitals, including simulation of tracheal intubation and maternal critical incidents, and support through online methods. The early detection and appropriate management of complications are areas that require improvement, and the identification of the high-risk patient is essential, ensuring that the most experienced assistance is provided early to the patients most in need. Further research into context-sensitive practices in LMICs remains vital.

Declaration of interests

The authors declare that they have no conflicts of interest.

MCQs

The associated MCQs (to support CME/CPD activity) will be accessible at www.bjaed.org/cme/home by subscribers to BJA Education.

Biographies

David Bishop FCA (SA), PhD is a specialist anaesthesiologist and associate professor at the University of KwaZulu-Natal. His main research interests are obstetric anaesthesia, burns anaesthesia and perioperative outcomes in resource-limited settings.

Associate Professor Dominique van Dyk FCA (SA) is a specialist anaesthetist with a particular interest in high-risk obstetric anaesthesia. She leads anaesthesia-related clinical services at Groote Schuur Hospital in Cape Town, South Africa.

Robert Dyer FCA (SA), PhD is emeritus professor in the Department of Anaesthesia and Perioperative Medicine of the University of Cape Town (UCT), and Chairperson of the WFSA obstetric anaesthesia committee. His main research interest has been obstetric anaesthesia, and he completed a PhD on the haemodynamics of spinal anaesthesia for Caesarean section.

Matrix codes: 1H02, 1I03, 2B02, 2B03, 2B04, 3B00, 3J02