The process of labour and delivery remains an unnecessary and preventable cause of death of women and babies around the world. Although the rates of maternal and perinatal death are declining, there are large disparities between rich and poor countries, and sub-Saharan Africa has not seen the scale of decline as seen elsewhere. In many areas, maternity services remain sparse and under-equipped, with insufficient and poorly trained staff. Priorities for reducing the mortality burden are provision of safe caesarean section, prevention of sepsis and appropriate care of women in labour in line with the current best practices, appropriately and affordably delivered. A concern is that large-scale recourse to caesarean delivery has its own dangers and may present new dominant causes for maternal mortality. An area of current neglect is newborn care. However, innovative training methods and appropriate technologies offer opportunities for affordable and effective newborn resuscitation and follow-up management in low-income settings.
Highlights
- •
There are still large rich–poor global disparities in labour-related mortality rates.
- •
Better access to safe caesarean section carries great promise, with risks attached.
- •
Regular emergency obstetric drills are indispensable elements of in-service training.
- •
Low-resource services must prioritise functional maternity units to save lives.
- •
Training and practice innovations can reverse the current neglect of newborn care.
Introduction
Until very recently in the evolution of our species, labour was a necessary event at the start of every human’s life. Even with anticipation of the birth of a healthy baby to a happy mother, the possibility of serious hazards and death always lurked. This reality remains, even in modern childbirth facilities, but is especially pertinent in low- and middle-income countries (LMICs).
Maternal and perinatal deaths related to labour and delivery
Burden of maternal deaths due to labour and delivery complications
Labour and delivery play a role in just over one-third of maternal deaths worldwide . In a recent systematic analysis for 2003–2009, the World Health Organisation (WHO) estimated that 901,000 deaths were labour-related (37% of maternal deaths globally), the causes being intrapartum haemorrhage (3%), postpartum haemorrhage (53%), sepsis (29%), obstructed labour (8%) and complications of labour (8%) . Over 99% of labour-related deaths occurred in the developing countries. The overwhelming burden remains in sub-Saharan Africa and South Asia . It is however encouraging that the Global Burden of Disease Study 2010 showed a 34% decline from 137,700 labour-related deaths in 1990 to 91,100 in 2010 .
Clinical causes of maternal deaths
There are few studies with clinical details on deaths due to obstetric haemorrhage and sepsis. Possibly the best available data are South African triennial reports on confidential enquiries into maternal deaths . The report for 2011–2013 states that there were 553 deaths due to intrapartum and postpartum haemorrhage. The most frequent cause was bleeding at or after caesarean section in 40% of the deaths, followed by ruptured uterus with previous caesarean section (9%), ruptured uterus without previous caesarean section (9%), uterine atony after vaginal birth (9%) and retained placenta (8%) ( Table 1 ). There were 205 deaths due to puerperal sepsis, of which 43% had caesarean sections. Caesarean section, traditionally the answer to the problem of labour-related death and injury, has become the chief cause of delivery-related maternal mortality in South Africa.
| Maternal deaths | Perinatal deaths | ||
|---|---|---|---|
| Cause | Cause-specific MMR ∗ | Cause | Cause-specific PNMR † |
| Haemorrhage: | |||
| Bleeding during or after CS | 8.8 | Non-specific ‘intrapartum asphyxia’ | 29.4 |
| Uterine rupture – previous CS | 2.1 | Cord around the neck | 5.8 |
| Uterine rupture – no previous CS | 2.0 | Cord prolapse | 4.2 |
| Atonic uterus after vaginal birth | 2.0 | Traumatic breech birth | 2.0 |
| Retained placenta | 1.7 | Traumatic assisted delivery | 0.5 |
| Cervical trauma | 0.8 | Precipitate labour | 0.4 |
| Vaginal trauma | 0.2 | ||
| Uterine inversion | 0.2 | ||
| Sepsis: | |||
| After vaginal birth | 4.6 | ||
| After CS | 3.4 | ||
∗ MMR = facility-based maternal mortality ratio per 100,000 live births in 2011–2013; denominator = 2,526,387.
† PNMR = perinatal mortality rate per 10,000 births 2000–2003; denominator = 462,348.
Avoidable factors in maternal deaths
Quality-of-care audits of adverse events reveal patient, system and health-worker deficiencies. Merali et al. recently conducted a systematic review of audits of maternal deaths in LMICs . Most of the audits were from sub-Saharan Africa. Recurring avoidable factors for labour-related maternal mortality were patient delay, blood transfusion problems, delay in care on admission to health facility, transport delay, delayed operative delivery, inadequate initial clinical management and unavailability of health workers. Patient delay as an avoidable factor raised the issue not only of access and finance but also of cultural factors and the need to involve families and communities in helping women to present for skilled care. New lessons on access emerged from a verbal autopsy study in Bangladesh, in an area where most births are still conducted by unskilled attendants . The study found that primary-care facilities were increasingly used by women, but were unable to respond with appropriate life-saving care. Health systems, previously underutilised, will need to meet the legitimate demand of women and families for emergency care that can prevent maternal deaths.
Burden of perinatal deaths due to labour and delivery complications
As with maternal mortality, about one-third of perinatal deaths (stillbirths and neonatal deaths) are related to labour and delivery . Intrapartum-related perinatal deaths are always tragic and often avoidable, occurring mostly in well-grown term babies who are alive when labour starts . Lawn et al. estimated that in 2008, there were 2.65 million stillbirths worldwide, of which 1.19 million (45%) were intrapartum-related; 99.6% of intrapartum stillbirths occurred in LMICs. The intrapartum stillbirth rate for LMICs was 22 times greater than in high-income countries (9.4 vs. 0.43 per 1000 births) . A similar 24-fold difference was found for neonatal deaths between the richest and the poorest countries (11.8 vs. 0.5 per 1000 live births) . Liu et al. found that of 3.07 million neonatal deaths in 2010, 0.72 million (23%) were intrapartum related. The good news was that the numbers of neonatal deaths fell between 2000 and 2010, by 2.4% annually worldwide. But Africa lagged behind the global average with only a 1.1% annual reduction in intrapartum-related deaths .
Clinical causes of perinatal deaths
There are different clinical pathways to intrapartum foetal hypoxia: 1) a condition predating labour, e.g., pre-eclampsia, putting the baby at risk for hypoxia; 2) placental separation as in placental abruption or uterine rupture; 3) umbilical cord compression; 4) uterine contractions in prolonged or augmented labour; and 5) foetal entrapment causing hypoxia and/or physical injury . Death may follow from hypoxic organ damage, meconium aspiration or infection . A problem with the published perinatal death data is a lack of clinical detail. A systematic review of 142 studies reporting the causes of stillbirth found only patchy information . A South African national audit from 102 facilities classified intrapartum-related deaths by clinical cause, the most frequent being ‘ intrapartum asphyxia’ (67%), followed by umbilical cord accidents (23%), traumatic breech birth (5%) and uterine rupture (3%) ( Table 1 ) .
Avoidable factors in perinatal deaths
The systematic review by Merali et al. also included avoidable factor audits of perinatal deaths . The most frequent labour-related avoidable factors for perinatal deaths were patient delay, delay in care on admission to birth facility, delayed operative delivery, unavailability of health workers and inadequate intrapartum monitoring of the foetus. To these can be added misuse of oxytocin and neonatal resuscitation failures. Service provision assessments in six African countries in 2007 found that less than a quarter of hospital-born babies had access to neonatal resuscitation. There were serious equipment deficiencies, with only a minority of staff appropriately trained .
Maternal and perinatal deaths related to labour and delivery
Burden of maternal deaths due to labour and delivery complications
Labour and delivery play a role in just over one-third of maternal deaths worldwide . In a recent systematic analysis for 2003–2009, the World Health Organisation (WHO) estimated that 901,000 deaths were labour-related (37% of maternal deaths globally), the causes being intrapartum haemorrhage (3%), postpartum haemorrhage (53%), sepsis (29%), obstructed labour (8%) and complications of labour (8%) . Over 99% of labour-related deaths occurred in the developing countries. The overwhelming burden remains in sub-Saharan Africa and South Asia . It is however encouraging that the Global Burden of Disease Study 2010 showed a 34% decline from 137,700 labour-related deaths in 1990 to 91,100 in 2010 .
Clinical causes of maternal deaths
There are few studies with clinical details on deaths due to obstetric haemorrhage and sepsis. Possibly the best available data are South African triennial reports on confidential enquiries into maternal deaths . The report for 2011–2013 states that there were 553 deaths due to intrapartum and postpartum haemorrhage. The most frequent cause was bleeding at or after caesarean section in 40% of the deaths, followed by ruptured uterus with previous caesarean section (9%), ruptured uterus without previous caesarean section (9%), uterine atony after vaginal birth (9%) and retained placenta (8%) ( Table 1 ). There were 205 deaths due to puerperal sepsis, of which 43% had caesarean sections. Caesarean section, traditionally the answer to the problem of labour-related death and injury, has become the chief cause of delivery-related maternal mortality in South Africa.
| Maternal deaths | Perinatal deaths | ||
|---|---|---|---|
| Cause | Cause-specific MMR ∗ | Cause | Cause-specific PNMR † |
| Haemorrhage: | |||
| Bleeding during or after CS | 8.8 | Non-specific ‘intrapartum asphyxia’ | 29.4 |
| Uterine rupture – previous CS | 2.1 | Cord around the neck | 5.8 |
| Uterine rupture – no previous CS | 2.0 | Cord prolapse | 4.2 |
| Atonic uterus after vaginal birth | 2.0 | Traumatic breech birth | 2.0 |
| Retained placenta | 1.7 | Traumatic assisted delivery | 0.5 |
| Cervical trauma | 0.8 | Precipitate labour | 0.4 |
| Vaginal trauma | 0.2 | ||
| Uterine inversion | 0.2 | ||
| Sepsis: | |||
| After vaginal birth | 4.6 | ||
| After CS | 3.4 | ||
∗ MMR = facility-based maternal mortality ratio per 100,000 live births in 2011–2013; denominator = 2,526,387.
† PNMR = perinatal mortality rate per 10,000 births 2000–2003; denominator = 462,348.
Avoidable factors in maternal deaths
Quality-of-care audits of adverse events reveal patient, system and health-worker deficiencies. Merali et al. recently conducted a systematic review of audits of maternal deaths in LMICs . Most of the audits were from sub-Saharan Africa. Recurring avoidable factors for labour-related maternal mortality were patient delay, blood transfusion problems, delay in care on admission to health facility, transport delay, delayed operative delivery, inadequate initial clinical management and unavailability of health workers. Patient delay as an avoidable factor raised the issue not only of access and finance but also of cultural factors and the need to involve families and communities in helping women to present for skilled care. New lessons on access emerged from a verbal autopsy study in Bangladesh, in an area where most births are still conducted by unskilled attendants . The study found that primary-care facilities were increasingly used by women, but were unable to respond with appropriate life-saving care. Health systems, previously underutilised, will need to meet the legitimate demand of women and families for emergency care that can prevent maternal deaths.
Burden of perinatal deaths due to labour and delivery complications
As with maternal mortality, about one-third of perinatal deaths (stillbirths and neonatal deaths) are related to labour and delivery . Intrapartum-related perinatal deaths are always tragic and often avoidable, occurring mostly in well-grown term babies who are alive when labour starts . Lawn et al. estimated that in 2008, there were 2.65 million stillbirths worldwide, of which 1.19 million (45%) were intrapartum-related; 99.6% of intrapartum stillbirths occurred in LMICs. The intrapartum stillbirth rate for LMICs was 22 times greater than in high-income countries (9.4 vs. 0.43 per 1000 births) . A similar 24-fold difference was found for neonatal deaths between the richest and the poorest countries (11.8 vs. 0.5 per 1000 live births) . Liu et al. found that of 3.07 million neonatal deaths in 2010, 0.72 million (23%) were intrapartum related. The good news was that the numbers of neonatal deaths fell between 2000 and 2010, by 2.4% annually worldwide. But Africa lagged behind the global average with only a 1.1% annual reduction in intrapartum-related deaths .
Clinical causes of perinatal deaths
There are different clinical pathways to intrapartum foetal hypoxia: 1) a condition predating labour, e.g., pre-eclampsia, putting the baby at risk for hypoxia; 2) placental separation as in placental abruption or uterine rupture; 3) umbilical cord compression; 4) uterine contractions in prolonged or augmented labour; and 5) foetal entrapment causing hypoxia and/or physical injury . Death may follow from hypoxic organ damage, meconium aspiration or infection . A problem with the published perinatal death data is a lack of clinical detail. A systematic review of 142 studies reporting the causes of stillbirth found only patchy information . A South African national audit from 102 facilities classified intrapartum-related deaths by clinical cause, the most frequent being ‘ intrapartum asphyxia’ (67%), followed by umbilical cord accidents (23%), traumatic breech birth (5%) and uterine rupture (3%) ( Table 1 ) .
Avoidable factors in perinatal deaths
The systematic review by Merali et al. also included avoidable factor audits of perinatal deaths . The most frequent labour-related avoidable factors for perinatal deaths were patient delay, delay in care on admission to birth facility, delayed operative delivery, unavailability of health workers and inadequate intrapartum monitoring of the foetus. To these can be added misuse of oxytocin and neonatal resuscitation failures. Service provision assessments in six African countries in 2007 found that less than a quarter of hospital-born babies had access to neonatal resuscitation. There were serious equipment deficiencies, with only a minority of staff appropriately trained .
Approaches and interventions to prevent deaths related to labour and delivery
The commonality of avoidable factors in labour-related deaths illustrates that overall improvements in intrapartum care will prevent maternal deaths, intrapartum stillbirths and neonatal deaths .
Safe caesarean delivery
Improved accessibility to caesarean section is clearly needed, in line with recent ‘Global Surgery 2030’ advocacy . But, as shown in the South African data, large-scale recourse to caesarean delivery is not appropriate as the intervention itself carries significant harms . By contrast, in high-resource settings the safety of caesarean delivery is similar to that of vaginal birth, allowing guidelines to include ‘maternal request’ as an indication for surgery following appropriate counselling . In low-resource settings, the challenge is to extend access to caesarean delivery beyond the low levels currently found in population surveys, for example, 4.6% for Malawi , while avoiding the excess morbidity and mortality seen in referral hospitals where caesarean delivery rates now approach 50% . Key clinical considerations for safe caesarean delivery are discussed below.
Gestational age
Timing of the planned surgery depends on accurate knowledge of the gestational age. Where early-pregnancy sonography confirms the dates, it is possible almost completely to avoid newborn respiratory distress by scheduling the procedure for 39 weeks . As the cost of ultrasound continues to fall, there is a need to develop systems to incorporate routine early-pregnancy scanning into antenatal care. Implementation research is needed on training, staffing, referral systems, procurement and maintenance of obstetric ultrasound for low-resource settings, as well as a critical study of the clinical impact. Correct dating also affects the management of post-term pregnancy, where induction of labour may avoid late intrauterine foetal death and labour-related perinatal death .
Safe anaesthesia
Anaesthetic complications accounted for a substantial proportion of maternal deaths reported in the United Kingdom until concerted efforts were made to ensure adequacy of equipment, staffing and preoperative preparation. A commentator noted a decade ago that ‘anaesthesia for caesarean section in the UK is now 30 times safer than it was in the 1960s’ . An insight into the multiple components that need to be in place to assure optimally safe obstetric anaesthesia is provided by the exhaustive listing of the Royal College of Anaesthetists . Many facilities in LMICs will find these standards difficult to meet.
Subsequent pregnancies
Concern about increasing caesarean section rates includes subsequent pregnancy outcome, such as placenta praevia or accreta leading to major haemorrhage and risk of obstetric hysterectomy. It is estimated that 359 caesarean deliveries at the first birth result in one additional case of placenta praevia among the next pregnancies . With regard to uterine scar rupture in subsequent trials of labour, while the guidelines typically quote risks of around 1:200 after one and 1:50 after two caesarean sections, there is concern about the ability to undertake adequate intrapartum monitoring and mount a rapid response if scar dehiscence is suspected in low-resource settings .
Surgical workforce
The cadre of surgeons is probably less important than the specific competencies of individual practitioners. The role of surgical technicians or clinical officers has been the subject of systematic review, and while there were no differences in maternal and perinatal mortality following procedures done by clinical officers versus physicians, there was a greater risk of wound sepsis and dehiscence among the former . The possible link between surgical skill and wound sepsis is hard to pin down in formal studies, but the excess of wound dehiscence may be explained by frequent use of midline vertical sub-umbilical incisions by non-physician surgeons. Surgical training should emphasise the transverse approach to mitigate this problem. Non-physician providers remain critically important for extending access to surgery . Regarding the cadre of anaesthesia providers, a Cochrane review could not generate informative conclusions . In any event, the comparison is theoretical, considering the small numbers of physician anaesthetists practicing in the developing world, especially in sub-Saharan Africa. Where present, such practitioners are likely to be in leadership roles or supporting complex surgical and critical care and would not be available as primary obstetric anaesthesia providers.
Sepsis related to labour and delivery
Common challenges include failure to use the established best practice, whether antibiotics for pre-labour rupture of the membranes or routine prophylactic antibiotics during caesarean section are used. While the recent widespread uptake of antiretrovirals in regions of high-human immunodeficiency virus prevalence has reduced the burden of sepsis associated with immunosuppression, there is a real risk of recrudescence of puerperal sepsis owing to deficient hygiene and antibiotic resistance . Effective care for women with severe sepsis requires early recognition, often a challenge in previously healthy women who are much sicker than is realised. Microbiology and rapid clinical chemistry backup is absent in many settings, so indirect tools such as the white cell count may be the limit of the available investigations.
Foetal intrapartum hypoxia and birth trauma
Life-threatening foetal hypoxia in labour is likely preceded by a period of detectable foetal distress, with foetal heart decelerations associated with tachycardia or later bradycardia. Routine foetal monitoring should allow detection of foetal distress. Cardiotocography (CTG) is not the answer, as it increases the likelihood of caesarean delivery while providing no benefit to the baby . Intermittent auscultation every 15 min in the active phase of the first stage, and every 5 min in the second stage, is currently recommended in the International Federation of Gynecology and Obstetrics 2015 guidelines . CTG should be reserved for high-risk labour, for example, in association with pre-eclampsia, induction of labour, prolonged labour or where foetal distress is suspected on auscultation. The response to foetal distress may be intrauterine resuscitation, assisted vaginal birth or caesarean section. Clear systematic documentation, for example, presented on a partograph, cannot be overemphasised .
Intrapartum foetal emergencies include severe foetal distress (from cord prolapse, placental abruption, uterine tachysystole and ruptured uterus) as well as entrapment disorders, such as shoulder dystocia and difficult breech birth. Randomised trials do not exist to guide the management of these conditions in terms of saving babies’ lives, but numerous guidelines, based mainly on clinical experience, are available to assist midwives and physicians with management algorithms. The important point is that all obstetric clinicians must be ready to deal with these catastrophic events at any time. Current evidence suggests that the best way to stay skilled and up to date is in performing regular drills (simulation training) in the labour ward . All that is required in the maternity facilities is leadership to ensure that the drills are done, with the necessary simulation models at hand.
Risks associated with childbirth and approaches to mitigation in low-resource settings are summarised in Table 2 . Risks related to haemorrhage and hypertensive disease are covered in Chapters 6 and 7, respectively.
