Identifying the risk: a prospective cohort study examining postpartum haemorrhage in a regional Australian health service

The third-stage of labour is known as the time from the birth of the baby to the complete expulsion of the placenta and membranes [1]. The most common complication of the third stage of labour is postpartum haemorrhage (PPH) [2]. Both within Australia, and the international context, a primary PPH is defined as blood loss greater than 500 ml within the first 24 h following a vaginal birth [3‐5]. Globally, PPH is the leading cause of maternal mortality [6], and a significant cause of morbidity. During 1994 less than 5% of women birthing experienced a PPH in Australia, and this increased to just under 7% in 2005 [7]. Associated factors which have also increased during this time are maternal age, caesarean birth, multiple pregnancies and induced labours [7]. Between 2008 and 2012, 12 (11%) of the Australian maternal deaths were attributable to obstetric haemorrhage [8], with the incidence of severe PPH (> 1000 ml) in Queensland doubling between 2001 (n = 1229; 2%) and 2016 (n = 1491; 3%) [3]. Internationally, the incidence is also on the rise, especially in industrialised countries, but reasons for this are not clear [9].

Despite the routine use of active management of the third stage of labour (current recommendation aimed at reducing PPH), PPH rates in high resource countries continue to rise [7]. Active management of the third stage of labour involves three components: administration of a utertonic, cord clamping (previously this was immediate, however now is often deferred to allow for placental transfusion in line with emerging evidence) [10, 11], and application of controlled cord traction to deliver the placenta [2]. Expectant management allows for physiological delivery of the placenta by maternal effort and gravity, and is usually offered in industrialised countries upon an individual woman’s choice and the preferred absence of risk factors for PPH.

Whilst active management is known to reduce the risk of PPH for many women [2], other contributing factors have also been associated with the upward trend in PPH throughout industrialised countries, including: increases in maternal age (≥35 years old) [12, 13], caesarean birth [13, 14], multiple pregnancies [15], labour inductions [14] and augmentation of labour [14, 16]. As exact aetiology is unknown, all women should be monitored for excessive blood loss following childbirth. Other international studies, based on retrospective data, have indicated that nulliparity [9, 17], episiotomy [17], retained placenta [17], high neonatal body weight [17], low antenatal haemoglobin [17], physiological management of the third stage in mixed-risk women [2], prior caesarean section [14], placenta previa [9], transverse lie [14, 18], and labour induction and augmentation [9] are the major independent risk factors for PPH. In addition, the increasing use of synthetic oxytocin to induce and augment labour has been significantly associated with severe PPH due to uterine atony [16]. However, changes in risk factors, such as increasing maternal age, accounted for only 5.6% of the increase in severe PPH in Kramer and colleagues’ [14] large study (n = 8.5 million), and further studies are warranted to understand the changing trends.

A substantial limitation of the retrospective data from which the previously mentioned studies are based upon, is the inaccuracy of health care professionals’ estimation of blood loss [19, 20]; the primary outcome of these studies. Standard practice within birthing units (including the study setting) is to estimate postpartum blood loss and commence more accurate measurement in dishes, or through weighing, once the midwife is concerned. Whilst this has been accepted practice, there are concerns regarding the inaccuracy of this practice and the speed at which a postpartum haemorrhage can progress. Glover [21] conducted a small study (n = 21) examining the accuracy of obstetricians, obstetric registrars and registered midwives in estimating blood loss through simulated ‘blood’ on pads, draw sheets and absorbent pads. Consistently the health care professionals underestimated the blood loss, and this underestimation increased proportionate to the blood loss. Similarly, another more recent Australian study (n = 88), aimed to determine the accuracy of estimating blood loss using simulated examples [22], by Registered Midwives. Participants were more accurate in their estimations with smaller loss and if it was contained in a measurable dish, yet wide variation in estimation was apparent, for example at a measured 600 ml loss, estimations varied from 50 ml through to 400 ml. Internationally, studies have confirmed this underestimation. Al Kadri and colleagues [19] conducted a prospective cohort study in Saudi Arabia and found a significant difference between the gravimetric (weighed) calculated blood loss and health care providers’ estimation, with a tendency to underestimate the loss by around 30%. This underestimation was not affected by the seniority of the health care provider. Overestimation has also been found to occur. Yoong and colleagues’ study conducted in London found that smaller volumes of estimated blood loss were especially inaccurate, being overestimated by up to 540% [23]. This study also found that level of experience did not appear to have a confounding effect on accuracy. Gravimetric measurement of postpartum blood loss has been shown to be a much more accurate method of quantifying blood loss than estimation alone [24]; while previous studies have relied on estimation alone to determine which risk factors are then associated with a PPH. Gravimetric measurement involves a clear quantification of blood loss through use of weighing blood soaked pads, linen and/or drapes [19].

Another important variable which has not been well accounted for in studies predicting risk for PPH are the changes which have evolved specifically within the traditional components of the actively managed third-stage of labour. The most recent Cochrane review comparing active with physiological management of the third stage of labour [2] determined that for women of ‘mixed risk’, active management reduced the risk of severe PPH by 34% . However, modifications have been made to active management in recent years (such as delayed cord clamping), and no studies to date have assessed the impact these modifications have had on PPH. The review identified this as a specific area for further research.

In 2015, 4672 (7.7%) women birthing in Queensland, Australia experienced a primary PPH [25]. At the time of this study the local area had a relatively high rate of women without known risk factors experiencing a severe PPH (6.8%) and it was important for the local health services to understand the reasons for this. So, within the context of rising incidence internationally and locally, this study aimed to prospectively explore variables associated with PPH in a cohort of women birthing vaginally in coastal Queensland, Australia, using the gravimetric method to measure blood loss.

Postpartum haemorrhage rates continue to rise throughout industrialised countries, and whilst key risk factors have been identified [9, 14], the rate at which PPH has increased is not explained by the contemporaneous changes seen in known risk factors. This study aimed to identify if variables such as third-stage management practices and other specific factors which may be associated with PPH were contributing to the higher than average PPH rate in the study location within Queensland, Australia. The overall rate of PPH (defined as blood loss ≥500 ml following vaginal birth) was 28.1% (n = 128), and severe PPH (defined as blood loss ≥1000 ml following vaginal birth) was 6.8% (n = 31), which is higher than reported rates in other Queensland hospitals [25]. Several variables were significantly associated with PPH in this cohort, including high neonatal body weight, perineal injury (requiring suturing), complications arising during labour, separation of mother and baby during the first hour after birth, and a gush of blood observed as a sign of placental separation prior to the placenta being delivered. Nulliparity, labour induction and augmentation, and synthetic use of syntocinon did not arise as significant risk factors in this study, contrary to other literature [14, 17].

Birth weight is reported consistently in the literature as a predictive factor associated with postpartum haemorrhage [7, 17, 27] and subsequent postpartum anaemia [27]; with a positive correlation found between increasing risk of PPH and increasing birth weight. Mean birth weight for sample infants was 3487 (± 448) grams with fetal macrosomia (≥4000 g) contributing 11.7% (n = 61) to this sample; consistent with figures reported for infants born 4000 g or more in Queensland (11%) during 2015 [25]. In this sample, increasing birth weight remained a significant factor associated with PPH in the GLM model and the reduced linear model, consistent with previous reports [14, 27]. However, the average birth weight of neonates within Queensland has not increased significantly over the past 15 years (3413 g in 2000 compared with 3407 g in 2015), thus not explaining the increasing trend for PPH within the local area. Perineal injury (including episiotomy) is also a known contributing factor to increased blood loss after birth [9, 27], and was supported as a contributing factor to PPH in this study.

Complications arising during labour and birth occurred in 185 women (35.4%) of the cohort, and were significantly associated with PPH (p = 0.012). Complications were defined in Table 2 and were aggregated for the GLM model, due to low numbers amongst specific complications. The rate of any complications arising during labour and birth has steadily risen over the past 15 years throughout Queensland, with 27.7% of women birthing without complication in 2000, decreasing to 21.3% in 2015. This is likely to be associated with the increasing PPH rates.

For maternity health professionals this finding is relevant, in that whilst women may initially present in labour as ‘low-risk’ without known risk factors, complications may arise during the labour and birth which should be considered and inform third-stage management practices. Specifically, an actively managed third-stage of labour should be recommended [6], within a shared decision-making framework with the woman and her support people, to enhance safety and reduction of risk of PPH [2]. Ideally this conversation should occur prior to labour and birth and is best situated in a continuity of care model [28, 29].

Conversely women in spontaneous labour, without complications arising during labour and birth have demonstrated less likelihood of PPH with physiological management of the third stage of labour [30‐32]. Whilst the mechanisms regarding this are not well understood, it has been hypothesized that interruption of endogenous oxytocin at any stage during labour and birth (including placental delivery) may effect optimal release and uptake of oxytocin by the uterine muscle [33], essential in reducing blood loss after birth. One such interruption to the physiological process is separation of mother and baby in the first hour after birth, which was significantly associated with PPH in our study. This confirms findings from Saxton and colleagues study [34] which found women who did not have uninterrupted skin to skin or initiate breastfeeding within the first hour were twice as likely to experience a PPH after adjustment for covariates (aOR 0.55, 95% CI 0.41–0.72, p < 0.001). Whist Saxton and colleagues study was limited by the retrospective, estimated primary outcome measure, our study substantiates their findings with a prospective, gravimetrically measure primary outcome measure.

Debate continues within the wider literature on inaccuracy of postpartum blood loss estimation, with experts suggesting that improving assessment of blood loss after birth is a critical step in early detection, and thus early management, of PPH [35]. Hancock and colleagues [36] conducted a comprehensive integrative review evaluating the various methods of assessing maternal blood loss during childbirth. Key findings from this review were that health professionals were highly inaccurate in the estimation of blood loss as volume and that training afforded some short-term improvement, but this was not sustained and did not eventuate in improvements to clinical outcomes [36]. Other aspects of assessment were identified by health professionals which aided their clinical decision making, including speed of blood flow and the woman’s clinical condition, which have received less attention in the literature and academic discussion. Other novel approaches to aid accurate blood loss measurement are also being more widely utilized, such as under-buttocks drapes – which are effective if the mother is birthing supine or semi-recumbent on a bed [37], but challenging in other positions.

Within our study, midwives were asked to identify which signs of placental separation were observed. Observation of a ‘gush of blood,’ which is well recognised as a sign of placental separation [1], was significantly associated with a PPH following vaginal birth in our study. Given this unique finding could be consistent with Hancock and colleagues’ [36] argument that not only is blood loss volume important to assess, but also the type and speed at which blood loss occurs, further scientific exploration is warranted. Earlier recognition and appropriate treatment of abnormal blood loss has potential for both human and fiscal resource savings [38].

As the midwives and obstetric doctors were requested to both estimate and then gravimetrically measure the blood loss of women included in this study, interesting comparisons were made. The primary outcome in this study was gravimetric blood loss, which was strongly correlated with the estimated blood loss (r = 0.88; p < 0.001). On average, the estimated blood loss was lower than the gravimetric blood loss; about 78% of the measured value. However, the majority of estimations were relatively accurate; 76% of estimated values were within 100 ml of measured values (either side). This too is an important finding as the sample size of this study was considerably larger than previous simulated example studies [22, 23] and Al Khadri and colleagues’ prospective study [19]. More accurate diagnosis of emerging PPH, and prompt and effective management may explain the comparatively high rate of non-severe PPH (blood loss of 500-999 ml) (128; 28.1%), in the study sites, and reflects positively on the health professionals’ involved ability to identify early and effectively manage PPH following vaginal birth. Natrella and colleagues [39] found that blood loss estimation becomes more inaccurate as the blood loss volume increases. Routine gravimetric measurement has been introduced into practice without evidence of its effectiveness at improving the management of PPH. It can be argued that the act of collecting and measuring blood loss may distract the mother and disrupt the release of endogenous oxytocin [40]; in addition to redirecting the health professional’s focus from the woman. In this study health professionals’ estimations of blood loss were relatively accurate, and discrepancies were not clinically significant. These findings support the use of gravimetric measurement in response to a variance in blood loss rather than as a routine practice.

This study identified several variables associated with PPH that support the findings of previous studies. High neonatal weight, perineal injury, complications during labour, and separation of mother and baby in the first hour after birth were risk factors for PPH. In contrast to previous study findings, nulliparity, labour induction and augmentation, and use of syntocinon were not independently associated with PPH in this study. This study also found that a health professional’s estimation of blood loss was relatively accurate in comparison to gravimetric assessment. This finding raises questions about routine gravimetric assessment of blood loss following uncomplicated births. The observation of a ‘gush of blood’ prior to the birth of the placenta was a significant risk factor for PPH in this study. This unique finding suggests a need for further research into the type and speed of blood loss associated with PPH.