Active management of the third stage of labour is advised by national and international third stage of labour practice guidelines and recommendations (World Health Organization (WHO), 2012; 2018; Royal College of Obstetrics and Gynaecology (RCOG), 2016; National Institute for Health and Care Excellence (NICE), 2017; Royal College of Midwives (RCM), 2018). This is a result of evidence provided by research studies conducted in obstetric-led units, which found evidence for a reduction in blood loss (primary postpartum haemorrhage) and treatment of this excessive blood loss after the birth of the baby with active management compared to expectant management (Prendiville et al, 1988; Thilaganathan et al, 1993; de Groot et al, 1996; Rogers et al, 1998; Prendiville et al, 2000; Begley et al, 2010; 2011a; 2015). However, research studies have reported many beneficial outcomes for healthy women at low risk of obstetric complications, who plan to give birth away from hospital obstetric-led units, particularly for women who choose to birth in midwife-led units (Brocklehurst et al, 2011; Hollowell et al, 2011; Hodnett et al, 2012; Christensen and Overgaard, 2017). Consequently, it has been suggested that the generalisability of these research studies and the guidelines for which they provide evidence may not be generalisable to women who have a normal physiological birth and choose to birth away from a hospital obstetric-led unit (Baker and Stephenson, 2022).
Literature review
A literature review by Baker and Stephenson (2022) previously published in the British Journal of Midwifery reported nine published research papers directly or indirectly comparing the incidence of postpartum haemorrhage in women at low risk of postpartum haemorrhage giving birth in midwife-led units, who receive either active or expectant management.
The cohort study by Kataoka et al (2018) found a statistically significantly higher incidence of severe postpartum haemorrhage with expectant management compared to active management. The studies by Begley et al (2011a, b) and Monk et al (2014) also compared the incidence of severe postpartum haemorrhage in midwife-led units compared to the obstetric-led units, finding no statistically significant difference in incidence levels between unit types, despite an increased use of expectant management in the midwife-led units compared with increased use of active management in the obstetric-led units. However, Fahy et al (2010) and Davies et al (2012) found a statistically significantly higher prevalence of postpartum haemorrhage in women receiving active management conducted at a tertiary unit, consisting of an obstetric-led unit and an alongside midwife-led unit, compared with expectant management conducted at a freestanding midwife-led unit. Dixon et al (2013) also reported that women who had expectant management and then received treatment (a uterotonic drug) for perceived excessive blood loss were less at risk of having a postpartum haemorrhage than women who recieved active management. Studies by Dixon et al (2009), Laws et al (2014) and Grigg et al (2017) revealed a higher prevalence of postpartum haemorrhage in obstetric-led units compared with midwife-led units despite an increased use of active management in the obstetric-led units, compared to an increased use of expectant management in the midwife-led units. Davis et al (2012) also found that women who had active management had a statistically significantly increased incidence of severe postpartum haemorrhage compared with women who received expectant management. Dixon et al (2009) also found an increase in postpartum haemorrhage in active compareed with expectant management.
None of the nine studies identified by Baker and Stephenson (2022) were conducted in the UK and only two (Fahy et al, 2010; Davis et al, 2012) directly examined the incidence of postpartum haemorrhage and active versus expectant management in women at low risk of postpartum haemorrhage, giving birth in midwife-led units. Davis et al (2012), a large national study, only examined the incidence of severe postpartum haemorrhage (defined as blood loss of more than 1000 ml). Fahy et al (2010) was a small-scale study in which the low numbers of women may limit the reliability, validity and generalisability of this study.
The aim of this present study was to address this gap in knowledge, using a retrospective cohort study to examine the relationship between active and expectant third stage of labour management approaches and the incidence of postpartum haemorrhage in women who had a normal birth in one of two midwife-led units. This study controlled for maternal body mass index, maternal age and baby's birthweight, as these are variables additionally identified as risk factors for postpartum haemorrhage. Significant maternal morbidity and mortality can occur because of excessive bleeding during the third stage of labour or shortly after, from the uterus not contracting strongly enough after the birth of the baby. Reducing the incidence of postpartum haemorrhage during the third stage of labour or shortly thereafter is an important issue that needs to be addressed to improve women's wellbeing.
Methods
Setting
The study was conducted between 1 July 2015 and 30 December 2016 within an NHS Foundation Trust in northwest England. The Trust provided maternity care (community and hospital-based services) for women at high and low risk of complications during pregnancy, birth and the postnatal period. The hospital-based services included one antenatal and postnatal unit, two antenatal day units, a maternity assessment centre, an obstetric-led unit and two midwife-led units. The Trust's birth centres (midwife-led units) consisted of an alongside and a freestanding midwife-led unit.
Women defined as at low risk of obstetric complications received antenatal and postnatal care from their community midwife. Women defined as at high risk of obstetric complications received shared care by midwives and the obstetric team and were advised to birth at the hospital's obstetric-led unit. Women at low risk of obstetric complications were given the option to birth at the obstetric-led unit, midwife-led units or at home. Occasionally, women at high risk of obstetric complications chose to birth at one of the midwife-led units or at home. Although the Trust advised these women to birth at the hospital obstetric-led unit, if the woman made an informed choice to birth at one of the midwife-led units or at home, the Trust supported her choice.
Women who laboured and birthed on the midwife-led units received care in labour and during the birth by a midwife. If any complications occurred during labour or postnatally, the woman was then transferred to the obstetric-led unit for assessment and further treatment by the obstetric staff and cared for by them and the rest of the maternity care team. The midwives who provided care for the women on the midwife-led units were not known to the women professionally before labour started.
Third stage of labour management
Care provided by midwives during the third stage of labour was based on the Trust's third stage of labour guideline, based on national and international guidelines (WHO, 2012; NICE, 2014) at the time the study was conducted. These guidelines recommended active management of labour for all women. The Trust's guideline defined active management of the third stage of labour as administering a uterotonic drug with the anterior shoulder or as soon as possible after the birth of the baby and before the cord was clamped and cut. The uterotonic drug consisted of syntometrine given by intramuscular injection. However, if the woman had raised blood pressure or the midwife was unable to monitor the woman's blood pressure, oxytocin by intramuscular injection should be administered. In active management, the administration of a prophylactic uterotonic drug is given to accelerate the contractility of the uterus and to prevent excessive blood loss.
After administering the uterotonic drug the cord should be clamped and cut. The cord should not be clamped and cut earlier than 1 minute after the birth of the baby, unless there were concerns about the integrity of the cord or the baby's heart rate was below 60 beats per minute and not getting faster. Ideally, the cord should be clamped and cut within 5 minutes of the birth of the baby. However, if the woman wanted the cord to be clamped and cut later than 5 minutes, she should be supported in her choice. Controlled cord traction to deliver the placenta should be carried out after signs of placental separation.
Expectant management was defined in the Trust's guideline as no routine use of uterotonic drugs, no clamping of the cord until pulsation has stopped and delivery of the placenta by maternal effort. In addition, women should be advised to convert to active management if their third stage blood loss becomes excessive, the placenta is not birthed within 60 minutes, or if there are concerns about the baby or the integrity of the umbilical cord or maternal request.
Regardless of management approach, the Trust guidelines stated that once the placenta has been delivered, it along with any blood loss from the third stage of labour should be collected in a receiver. If there are any pads or sheets underneath the woman that are blood-stained, they should be removed and replaced. Any blood loss during the third stage of labour, including blood-stained sheets and pads, should be weighed, to give an estimated blood loss. Weighing this blood loss is not always possible; for example, if the woman has a pool birth, then the blood loss in the pool must be estimated by the midwife.
Postpartum haemorrhage was defined as blood loss of 500 ml or over, but under 1000 ml, and severe postpartum haemorrhage was defined as blood loss of 1000 ml or more.
Eligibility criteria
All women who had a normal vaginal birth (unassisted vaginal birth following a spontaneous labour and birth at term between 37- and 42-weeks' gestation with a cephalic presentation of a single live baby) at the midwife-led units were included in the study.
Data collection
Anonymised data for this study were collected from the computer-based password-protected maternity data records held by the Trust. Entries also included the woman's intended third stage management approach, identified by the midwife providing her care, and the third stage management approach they used (treatment received). Any changes to intended management approaches and any blood loss volume experienced by the woman during the third stage of labour were documented. Blood loss was assessed by midwives providing care for the woman by weighing any blood-stained sheets and pads and by visual estimation, as per Trust guidelines. Any deviation in care given from Trust guidelines was also documented.
Data were stored in line with the sponsoring University's recommendations and the Data Protection Act 2018 (Mullock and Leigh-Pollet, 2000), and in accordance with the NHS Trust's Research and Development Department protocols.
Data analysis
The sample was summarised descriptively by intended management style and as a full cohort. Maternal antenatal characteristics recorded included proportion with previous retained placenta, previous postpartum haemorrhage as a result of hypertonic uterus, previous caesarean section and existing uterine abnormalities, body mass index, maternal age and parity (number of previous pregnancies reaching viable gestational age of 24 weeks, including live births and stillbirths). Maternal intrapartum characteristics recorded included duration of first, second and third stages of labour, birth weight of baby and incidences of trauma (first, second and third degree tears).
Adjusted (controlled) logistic regression analyses were used to assess the effect of management approach on the outcomes of postpartum haemorrhage and severe postpartum haemorrhage, controlling for variables additionally identified as risk factors for postpartum haemorrhage, including maternal body mass index (categorised as BMI of 35 kg/m2 or above or BMI up to 35 kg/m2), maternal age (categorised as aged over 40 years or aged up to 40 years) and baby's birth weight (categorised as over 4.0 kg or 4.0 kg or under). Parallel analyses were conducted on the two outcome measures. The intention-to-treat approach was used for both analyses. The amount of missing data was negligible and there was no evidence that missing data were not missing completely at random. Hence complete case analysis was used.
Ethical approval
Approval for the study was given by the University of Huddersfield School Research Ethics Panel (Subject: 0764615, SREP/2016/093). Permission to conduct the study was given by the NHS Trust's Research and Development Department and the Trust's clinical governance lead and the head of midwifery. The Trust's Caldicott Guardian was made aware of the study protocol and that the necessary approval had been given.
Results
A total of 1268 women were included in the study: 765 (60.1%) who birthed at the midwife-led units intended to have active management while 508 (39.9%) intended to have expectant management. A small number of women subsequently converted from expectant to active management; however, following the intention-to-treat paradigm, these women were analysed as per intended treatment. A summary of the characteristics of the women and of risk factors for postpartum haemorrhage partitioned by intended management approach are presented in Table 1.
Table 1. Descriptive summary of sample
| Variable | Frequency, n=1268 (%) | |||
|---|---|---|---|---|
| Active management | Expectant management | Total | ||
| Previous retained placenta | 1 (0.13) | 0 (0.0) | 1 (0.08) | |
| Previous postpartum haemorrhage because of hypotonic uterus | 1 (0.13) | 0 (0.0) | 1 (0.08) | |
| Previous caesarean section | 1 (0.13) | 0 (0.0) | 1 (0.08) | |
| Existing uterine abnormalities | 2 (0.26) | 0 (0.0) | 2 (.016) | |
| Body mass index (kg/m2) | <35 | 761 (99.5) | 495 (98.4) | 1256 (99.1) |
| ≥35 | 4 (0.5) | 8 (1.6) | 12 (0.9) | |
| Maternal age (years) | <20 | 21 (2.7) | 8 (1.6) | 29 (2.3) |
| 20–29 | 340 (43.9) | 221 (43.9) | 561 (43.9) | |
| 30–34 | 268 (34.7) | 182 (36.2) | 450 (35.2) | |
| 35–39 | 122 (15.7) | 76 (15.1) | 198 (15.5) | |
| ≥40 | 24 (3.1) | 16 (3.2) | 40 (3.1) | |
| Parity | 0 | 248 (32.4) | 187 (36.9) | 435 (34.2) |
| 1–3 | 503 (65.8) | 320 (63.1) | 823 (64.7) | |
| 4 | 14 (1.8) | 0 (0.0) | 14 (1.1) | |
| >4 | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Mean maternal age (years) (standard deviation) | 29.8 (5.2) | 29.9 (5.24) | 29.8 (5.23) | |
| Mean duration of stage of labour (minutes) (standard deviation) | First | 153.0 (142) | 142 (140) | 148 (141) |
| Second | 25.1 (32.5) | 19.7 (19.7) | 23.0 (29.3) | |
| Third | 21.4 (34.6) | 35.1 (35.0) | 26.8 (35.3) | |
| Birth weight of baby (g) | 3478 (440) | 3501 (39.2) | 3487 (423) | |
| Birth weight of baby ≥4kg | 162 (20.9) | 137 (27.0) | 299 (23.3) | |
| Birth weight of baby <4kg | 603 (79.1) | 370 (73.0) | 973 (77.3) | |
| Birth trauma | None | 320 (41.5) | 216 (42.7) | 536 (41.9) |
| First degree tear | 175 (22.7) | 116 (22.9) | 291 (22.8) | |
| Second degree tear | 251 (32.5) | 167 (33.0) | 418 (32.7) | |
| Third degree tear | 23 (3.0) | 9 (1.8) | 32 (2.5) | |
| Outcome | No postpartum haemorrhage | 692 (90.4) | 437 (86.0) | 1129 (89.7) |
| Postpartum haemorrhage (including severe) | 59 (9.54) | 71 (14.0) | 130 (11.3) | |
| Severe postpartum haemorrhage | 14 (1.83) | 16 (3.66) | 30 (2.38) | |
Multiple logistic regression analysis revealed that management approach was significantly associated with postpartum haemorrhage with an effect of moderate magnitude (P=0.015). None of the controlling variables were significantly associated with postpartum haemorrhage. Table 2 summarises the regression parameters for this analysis.
Table 2. Multiple logistic regression parameters for postpartum haemorrhage
| Variable | Odds ratio (95% confidence interval) | P value | |
|---|---|---|---|
| Management approach | Active (ref) | 1.54 (1.09–2.19) | 0.015 |
| Expectant | |||
| Maternal body mass index (kg/m2) | ≤35 (ref) | 0.785 (0.181–3.41) | 0.746 |
| >35 | |||
| Maternal age (years) | <40 (ref) | 0.639 (0.194–2.11) | 0.462 |
| ≥40 | |||
| Birth weight (kg) | ≤4 (ref) | 0.907 (0.557–1.68) | 0.907 |
| >4 | |||
A multiple logistic regression analysis revealed that management approach was not significantly associated with severe postpartum haemorrhage (P=0.134). None of the controlling variables were significantly associated with severe postpartum haemorrhage. Table 3 summarises the regression parameters for this analysis.
Table 3. Multiple logistic regression parameters for severe postpartum haemorrhage
| Variable | Odds ratio (95% confidence interval) | P value | |
|---|---|---|---|
| Management approach | Active (ref) | 1.74 (0.843–3.61) | 0.134 |
| Expectant | |||
| Maternal body mass index (kg/m2) | ≤35 (ref) | 2.24 (0.288–17.4) | 0.441 |
| >35 | |||
| Maternal age (years) | <40 (ref) | 2.41 (0.549–10.5) | 0.244 |
| ≥40 | |||
| Birth weight (kg) | ≤4 (ref) | 0.537 (0.126–2.29) | 0.400 |
| >4 | |||
Discussion
The study revealed that the incidence of postpartum haemorrhage was higher in the expectant management group compared with the active management group. This difference was statistically significant and of moderate magnitude, with a raised odds of 54.3% in the expectant management group compared with the active management group. This agrees with the findings from Cochrane systematic reviews (Begley et al, 2010; 2011a; 2011b; 2015; 2019).
Although the incidence of severe postpartum haemorrhage was higher in the expectant management group compared with the active management group, this effect was not statistically significant. This is in line with the findings from the Cochrane systematic reviews (Begley et al, 2010; 2011a; 2011b; 2015; 2019). The findings from the present study regarding severe postpartum haemorrhage contrast with those of other researchers investigating women at low risk of postpartum haemorrhage giving birth in midwife- and obstetric-led units (Fahy et al, 2010; Davis et al, 2012; Laws et al, 2014). The present study's findings are also in contrast with the findings from a study by Monk et al (2014), which revealed a trend towards a higher incidence of severe postpartum haemorrhage in obstetric-led units. These units had an increased use of active management compared to increased use of expectant management in midwife-led units. Similarly, Kataoka et al (2018) found a significantly higher incidence of severe postpartum haemorrhage with expectant management compared to active management in women classified as low risk of postpartum haemorrhage, giving birth in midwife- and obstetric-led units.
No evidence was revealed that any of the variables included in the analysis additionally identified as potential risk factors for postpartum haemorrhage were substantively or significantly associated with postpartum haemorrhage or severe postpartum haemorrhage. However, the relative clinical rarity of the outcomes, particularly severe postpartum haemorrhage, may have limited the power of the analysis to detect significant effects.
While the present study and other research studies have revealed significantly higher incidences of postpartum haemorrhage with expectant management, as compared to active management, these findings are based on a low baseline, and in absolute terms, the raised risk of postpartum haemorrhage and severe postpartum haemorrhage in expectant management approaches is low. It has also been commented that well-nourished, healthy women are able to compensate for a blood loss of up to 1000 ml (Cunningham and Williams, 2001; Blackburn, 2008; Oishi et al, 2017). As a result, blood loss up to 1000 ml may be considered physiological in a woman, depending on their physiological response to that loss (WHO, 1997). Therefore, the prevalence of postpartum haemorrhage up to 1000 ml in women at low risk of postpartum haemorrhage with no clinical symptoms of excessive blood loss may be of limited clinical importance.
The present study found that a significant minority (38.4%) of women chose to have expectant third stage of labour management rather than active management, suggesting that this approach seemed acceptable for them. Numerous other studies have also shown that when women are offered expectant management as a reasonable option, they will choose it (Rogers et al, 1998; Dixon et al, 2009; 2013; Fahy et al, 2010; Begley et al, 2011b; Gottvall et al, 2011; Davis et al, 2012; Laws et al, 2014; Monk et al, 2014; de Jonge et al, 2015; Grigg et al, 2017; Kataoka et al, 2018).
Conclusions
This study revealed a statistically significant increase in the incidence of postpartum haemorrhage with expectant management compared with active management, but no evidence for a statistically significant difference in incidence of severe postpartum haemorrhage across groups defined by management style. Blood loss up to 1000 ml may be considered physiological in a woman. Therefore, the prevalence of postpartum haemorrhage up to 1000 ml in women at low risk of postpartum haemorrhage with no clinical symptoms of excessive blood loss may be of limited clinical importance. Expectant management is seen as a reasonable choice for some women.
Findings from this study could be used to provide evidence to inform practice guidelines and recommendations for midwife-led units. Having separate practice guidelines and recommendations for midwife-led units is important, as practices and outcomes during the third stage of labour are influenced by the healthcare professional, the woman they provide care for and the setting in which they provide care. National and international third stage of labour practice guidelines and recommendations that are based on research studies of varying quality and conducted in obstetric-led-units may have limited applicability to midwives practising in midwife-led units.
Expectant management is supported by the findings of the present study, as well as other research studies, as being a reasonable option for women at low risk of postpartum haemorrhage who want to birth with minimal intervention at a midwife-led unit. Therefore, midwives practising in this setting should be given the opportunity to gain the knowledge and skills to conduct both active and expectant third stage of labour management approaches. Trusts and higher education institutions should provide education and study days to facilitate this. Additionally, student midwives should be equally exposed to both third stage management approaches during their training, either in the clinical setting or simulated, so they are confident and skilful in both third stage approaches on qualification.
Key points
- This large cohort study revealed a significant minority of women at low risk of postpartum haemorrhage birthing in midwife-led units choose to receive expectant, rather than active management of the third stage of labour.
- The risk of postpartum haemorrhage (blood loss 500–1000 ml) and severe postpartum haemorrhage (blood loss>1000 ml) is low both in women receiving active management and in women receiving expectant management.
- The incidence of postpartum haemorrhage is significantly lower in women intending to receive active, management than in women intending to receive expectant management.
- The incidence of severe postpartum haemorrhage is not statistically significantly lower in women intending to receive active management than in women intending to receive expectant management.
CPD reflective questions
- Do you think the risk of postpartum haemorrhage in women birthing in midwife-led units is different than women birthing in obstetric-led units? Why do you think this is?
- Do you feel women are able to make an informed choice regarding their third stage care approach?
- How confident are you in providing women with information to make an informed choice regarding the third stage of labour care?