The following section outlines the recommendation and the corresponding narrative summary of evidence for the prioritized question. The GRADE table is presented in Annex 5. The evidence-to-decision table, summarizing the balance between desirable and undesirable effects and the overall quality of the supporting evidence, values and preferences of stakeholders, resource requirements, cost-effectiveness, acceptability, feasibility and equity that were considered in determining the strength and direction of the recommendation, is presented in Annex 4.
The following recommendation was adopted by the GDG. Evidence on the effectiveness of the intervention was derived from one systematic review and was summarized in GRADE tables (Annex 5). The quality of the supporting evidence was rated as ‘moderate’ for most critical outcomes. To ensure that the recommendation is correctly understood and appropriately implemented in practice, additional ‘remarks’ reflecting the summary of the discussion by GDG are included under the recommendation.
Early use of intravenous tranexamic acid (within 3 hours of birth) in addition to standard care is recommended for women with clinically diagnosed postpartum haemorrhage following vaginal birth or caesarean section.(Strong recommendation, moderate quality of evidence)
Remarks
Based on the dosing regimen used in the WOMAN trial, the GDG supports the administration of tranexamic acid (TXA) at a fixed dose of 1 g (100 mg/ml) intravenously (IV) at 1 ml per minute (i.e. administered over 10 minutes), with a second dose of 1 g IV if bleeding continues after 30 minutes, or if bleeding restarts within 24 hours of completing the first dose.
The WOMAN Trial defined “clinically diagnosed postpartum haemorrhage” as clinically estimated blood loss of more than 500 ml after a vaginal birth or 1000 ml after caesarean section, or any blood loss sufficient to compromise haemodynamic stability.
Based on evidence from the WOMAN trial, the reference point for the start of the 3-hour window for starting TXA administration is time of birth. If time of birth is unknown, the best estimate of time of birth should be used as the reference point. As most deaths due to postpartum haemorrhage occur within the first 2 to 3 hours after birth, it is critical that TXA is given as soon as possible to achieve clinical benefits.
Analysis of the effects of timing of administration in the WOMAN trial, as well as an individual participant data (IPD) meta-analysis of 40 138 bleeding patients (including WOMAN trial participants), indicates that TXA administration beyond 3 hours does not confer any clinical benefit. Furthermore, the point estimates of effect of TXA use beyond 3 hours on death for trauma or after PPH were both in the direction of harm, albeit not statistically significant for women with PPH. In view of this evidence, the GDG does not support the use of TXA more than 3 hours after birth.
Administration of TXA should be considered as part of the standard postpartum haemorrhage treatment package. Standard care in the context of this recommendation includes routine care for PPH treatment, including fluid replacement, medical (uterotonics), monitoring of vital signs, nonsurgical (e.g. bimanual compression, intrauterine balloon tamponade, nonpneumatic antishock garment, aortic compression) and surgical interventions (e.g. brace sutures, arterial ligation or hysterectomy) in accordance with WHO guidelines or adapted local PPH treatment protocols.
TXA should be used in all cases of PPH regardless of whether the bleeding is due to genital tract trauma or other causes.
The use of TXA should be avoided in women with a clear contraindication to antifibrinolytic therapy (including TXA) (e.g. a known thromboembolic event during pregnancy).
This recommendation applies only to IV use. The evaluation of benefits and potential harms of other routes of TXA administration is a research priority.
Regardless of the level of health system resources, TXA should be recognized as a life-saving intervention and be made readily available for the management of postpartum haemorrhage in settings where emergency obstetric care is provided.
B. Assessment
Effects of the intervention
What are the anticipated effects of administration of TXA in addition to standard care for PPH treatment?
Research evidence
Evidence on the use of TXA for treatment of PPH was extracted from a forthcoming Cochrane systematic review of two trials (20 212 women).12 This review included trials that compared the use of any fibrinolytic drug with no treatment in women with PPH. However, no evidence was identified for interventions other than TXA.
One multicentre trial was conducted in eight obstetric units in France with recruitment between 2005 and 2008.14 This trial randomized 152 women with PPH > 800 ml after a vaginal birth. The intervention group received a loading dose of 4 g TXA mixed with 50 ml saline, administered IV over 1 hour, followed by a maintenance dose of 1 g/hour for 6 hours. Women in the control group were given standard care only, as per the routine practice in participating facilities. The primary outcome was blood loss between randomization and 6 hours.
The second (WOMAN trial) was a multicountry, multicentre, placebo-controlled randomised trial of 20 060 women in 193 hospitals, across 21 high-, middle- and low-income countries conducted between March 2010 and April 2016.9 The trial randomized women with clinically diagnosed PPH, defined as clinically estimated blood loss after a vaginal birth of > 500 ml, or > 1000 ml following a caesarean section, or any blood loss sufficient to compromise haemodynamic stability and where the clinician responsible for care was uncertain as to whether or not to use TXA. In addition to usual care, women in the experimental group were initially given 1 g TXA IV in a 10 ml solution, at an approximate rate of 1 ml/minute, as soon as possible after randomization. A second dose was used if bleeding continued after 30 minutes or if it stopped and restarted within 24 hours after the first dose. The control arm received placebo (normal saline) using the same regimen. When the trial protocol was registered, the primary outcome was a composite of death from all causes or hysterectomy within 42 days. During the course of the study (but before results were available or any unblinding), the primary outcome was revised to maternal death due to bleeding, and the sample size increased.
Evidence regarding this intervention is almost entirely derived from the WOMAN trial.
Comparison: TXA (in addition to standard care) versus standard care alone
The effects of TXA on critical outcomes for all women with PPH, regardless of how PPH was defined, the mode of birth or timing of PPH administration, are described below.
Maternal mortality (all causes): Moderate certainty evidence suggests slightly fewer deaths in the group receiving TXA although this difference was not statistically significant (two studies, 20 172 women; 227/10 113 (2.2%) vs 256/10 059 (2.5%); RR 0.88, 95% CI 0.74 to 1.05).
Maternal mortality due to PPH: In both trials, clinicians were asked to record the primary cause of death. Moderate certainty evidence suggests that deaths that were considered to be due to bleeding were probably reduced in the TXA group (two studies, 20 172 women, 155/10 113 (1.5%) vs 191/10 059 (1.9%), RR 0.81, 95% CI 0.65 to 1.00). The number needed to treat (NNT) to prevent one maternal death due to bleeding is 258 (95% CI 133.2 to 4051.8).
Severe maternal morbidity: The French trial reported multiple organ failure; there were no events in either arm and very few admissions to intensive care (one study, 152 women, 3/77 (3.9%) vs 5/74 (6.8%), RR 0.58 (95% CI 0.14 to 2.33). The number of women suffering any severe morbidity was not reported in the WOMAN trial report, but specific morbidities were reported. Moderate certainty evidence suggested little or no difference between groups for any of morbidity outcomes reported (respiratory failure: RR 0.87, 95% CI 0.67 to 1.12; seizure: two studies; RR 0.76, 95% CI 0.49 to 1.20; hepatic failure RR 0.96, 95% CI 0.58 to 1.60; cardiac failure: RR 0.95, 95% CI 0.73 to 1.23; renal failure: two studies; RR 1.09, 95% CI 0.85 to 1.39).
Blood products transfusion (all): Moderate certainty evidence suggests there is very little or no difference between groups for transfusion of blood products, with more than half of the women in both arms of the WOMAN trial receiving a transfusion (two studies; RR 1.00, 95% CI 0.97 to 1.03).
Additional blood loss: The French trial reported additional blood loss > 500 ml or > 1000 ml. Low-quality evidence suggests TXA probably reduces blood loss > 500 ml (RR 0.50, 95% CI 0.27 to 0.93, 151 women). Although the direction of effect was the same for loss > 1000 ml, the study had insufficient power to demonstrate a difference between groups (4/77 women versus 8/74).
Additional uterotonics: The vast majority of women in the WOMAN trial received uterotonics (99.3% vs 99.1%, two studies; RR 1.00, 95% CI 1.0 to 1.0).
Surgical interventions: High or moderate certainty evidence suggests there is probably little difference between groups for most surgical interventions to control bleeding (hysterectomy (all): two studies; RR 1.01, 95% CI 0.88 to 1.17; ligature: RR 0.88, 95% CI 0.74 to 1.05; embolization: RR 0.82, 95% CI 0.42 to 1.62). High certainty evidence suggests laparotomy to control bleeding is reduced for women in the TXA group (0.8% vs 1.3%) (RR 0.64, 95% CI 0.49 to 0.85) while brace sutures are increased (RR 1.19, 95% CI 1.01 to 1.41).
Invasive nonsurgical interventions: High certainty evidence suggests there is probably little or no difference in intrauterine tamponade (one study; RR 0.96, 95% CI 0.87 to 1.06) or manual removal of placenta: (one study; RR 0.95, 95% CI 0.87 to 1.04).
Procedure-related complications: Moderate certainty evidence suggests there is probably little or no difference between groups for thromboembolic events (any maternal thromboembolic event: RR 0.88, 95% CI 0.54 to 1.43; deep venous thrombosis: two studies; RR 0.62 95% CI 0.20 to 1.88; pulmonary embolism RR 0.85, 95% CI 0.44 to 1.61; myocardial infarction: RR 0.66, 95% CI 0.11 to 3.97; stroke: RR 1.33, 95% CI 0.46 to 3.82).
Neonatal adverse effects: Available neonatal outcome data were limited (data from WOMAN trial only). There were no neonatal thromboembolic events and no clear differences in deaths in breastfed neonates (eight deaths with TXA vs seven deaths with placebo) in the WOMAN trial.
Longer-term outcomes: Available data on longer-term outcomes was limited (data from the WOMAN trial only). Outcomes in the WOMAN trial were measured up to hospital discharge or 42 days if still in hospital. There was no information on longer-term outcomes in women or babies.
Subgroup analysis examining treatment effect by mode of birth (vaginal or caesarean) suggests no clear difference in effect on maternal death (all causes) and maternal death due to PPH for type of birth (moderate certainty of evidence).
Comparison: TXA (in addition to standard care) versus standard care alone, by timing of TXA administration
Evidence for this subgroup comparison was derived from a pre-planned subgroup analysis of the WOMAN trial.
Maternal mortality due to PPH: There are subgroup differences for the timing of drug administration. Women receiving TXA less than 1 hour after birth had reduced risk of death from bleeding, but the confidence interval crossed the line of no effect (less than 1 hour: RR 0.80, 95% CI 0.55 to 1.16). Women receiving TXA 1 to 3 hours after birth were at reduced risk of death from bleeding (1 to 3 hours: RR 0.60, 95% CI 0.41 to 0.88) compared with women where more than 3 hours had elapsed before TXA was administered (more than 3 hours: RR 1.07, 95% CI 0.76 to 1.51).
Maternal mortality (all cause): Compared to the control group, women receiving TXA less than 1 hour after birth had similar risks of death (any cause) (less than 1 hour: RR 0.98, 95% CI 0.72 to 1.33), as did women receiving TXA more than 3 hours after birth (more than 3 hours: RR 1.00, 95% CI 0.75 to 1.33). However, women receiving TXA 1 to 3 hours after birth were at reduced risk of death from all causes (1 to 3 hours: RR 0.69, 95% CI 0.49 to 0.96).
Death or hysterectomy: Compared to the control group, women receiving TXA less than 1 hour after birth had similar risks of death or hysterectomy (less than 1 hour: RR 1.08, 95% CI 0.91 to 1.28), as did women receiving TXA more than 3 hours after birth (more than 3 hours: RR 1.01, 95% CI 0.82 to 1.25). However, women receiving TXA 1 to 3 hours after birth were at reduced risk of death or hysterectomy (1 to 3 hours: RR 0.80, 95% CI 0.63 to 1.00).
Laparotomy for bleeding: Compared to the control group, women receiving TXA less than 1 hour after birth had reduced risk of laparotomy for bleeding (less than 1 hour: RR 0.48, 95% CI 0.29 to 0.79), as did women receiving TXA at 1 to 3 hours after birth (1 to 3 hours: RR 0.54, 95% CI 0.31 to 0.95). Women receiving TXA more than 3 hours after birth were not at reduced risk of laparotomy for bleeding (more than 3 hours: RR 0.89, 95% CI 0.59 to 1.35).
Desirable effects
How substantial are the desirable anticipated effects of TXA + standard care vs standard care alone?
Undesirable effects
How substantial are the undesirable anticipated effects TXA + standard care vs standard care alone?
Certainty of the evidence
What is the overall certainty of the evidence of effects?
Additional considerations
Additional evidence was obtained from a forthcoming individual patient data (IPD) on the impact of treatment delay on the effectiveness and safety of antifibrinolytics in acute, severe haemorrhage.15 The IPD meta-analysed 40 138 bleeding patients (with 3 558 deaths recorded) who received TXA or placebo from WOMAN and CRASH-2 trials combined. The authors reported that deaths from PPH peaked at 2 to 3 hours after childbirth, and immediate treatment improved bleeding survival. Treatment delay appears to reduce benefit - the benefit appears to decrease by 10% for every 15 minutes’ delay, with no benefit seen after 3 hours. The point estimates of effect of TXA use beyond 3 hours on death for trauma or after PPH were both in the direction of harm, albeit not statistically significant for women with PPH.
Values and preferences
Is there important uncertainty about, or variability in, how much women value the main outcomes?
Typically, women, healthcare providers and policy-makers place a higher value on avoiding a maternal death, even when potentially associated with an increase in invasive surgical interventions, such as brace sutures. Therefore, women, healthcare providers and policy-makers in all settings are likely to place a high value on the reduction in the risk of maternal death due to bleeding. The GDG is confident that women, healthcare providers and policy-makers in any setting will invariably place a higher value on this benefit, compared to any inconvenience (or drawbacks) that TXA use might cause to the woman, her baby or the health system. Stakeholders with different values in different contexts are unlikely to make different decisions when presented with these choices.
View in own window
| Judgement |
|---|
□
Important uncertainty or variability | □
Possibly important uncertainty or variability | □
Probably no important uncertainty or variability | ☒
No important uncertainty or variability |
Balance of effects
Does the balance between desirable and undesirable effects favour use of TXA in addition to standard care (intervention) or standard care alone (comparison)?
There is evidence that TXA is probably beneficial in reducing maternal deaths due to bleeding and reducing the need for laparotomy to stop bleeding. Early treatment appears to optimize benefit. There does not appear to be evidence of maternal or newborn harms, or significant side-effects. While no difference in newborn thromboembolic events were seen, in the WOMAN trial most women and babies were followed until discharge from the health facility, thus this evidence is more likely representative of the first few days after birth.
View in own window
| Judgement |
|---|
□
Don’t know | □
Varies | □
Favours the standard care alone | □
Probably favours the standard care alone | □
Does not favour TXA + standard care or standard care alone | □
Probably favours TXA + standard care | ☒
Favours TXA + standard care |
Resources required
What are the resource requirements for administering TXA in addition to standard care for PPH treatment?
Research evidence
None of the studies included in the Cochrane systematic review conducted a formal cost-effectiveness analysis.
Main resource requirements
The use of TXA in addition to standard PPH treatment requires the existence of healthcare providers who have been trained in how to administer intravenous drugs.
View in own window
| Resource | Description |
|---|
| Training | 2 to 3 day practice-based training/practice drills for PPH management |
| Supplies | 1 to 2 g of TXA (varies between settings, with an approximate range of $1.00 to $5.70 per g) 16
IV infusion set
Syringe/needle/swabs = approximately US$0.08 to $0.10 |
| Equipment | None required. |
| Time | Average time needed is 10 to 15 minutes for gaining IV access and administration of the drug (depending on other factors such as provider skills). However, sufficient time is needed for monitoring the response of the woman to treatment as required for all cases of PPH. |
| Supervision and monitoring | Regular supervision and review by labour ward lead, especially when first introduced. |
Additional considerations
TXA is relatively cheap in most contexts, easy to administer, and it is often available in healthcare settings due to its use in trauma and surgery. Research evidence on cost-effectiveness can be extrapolated from cost-effectiveness analysis of TXA for bleeding trauma patients.16 The study found that administering TXA to bleeding trauma patients within 3 hours of injury saved an estimated 372, 315 and 755 life-years (LYs) per 1 000 trauma patients in Tanzania, India and the UK respectively. The cost of giving TXA to 1 000 patients was $17 483 in Tanzania, $19 550 in India and $30 830 in the UK. The incremental cost of giving TXA versus not giving TXA was $18 025 in Tanzania, $20 670 in India and $48 002 in the UK. The estimated incremental cost per LY gained of administering TXA is $48, $66 and $64 in Tanzania, India and the UK respectively. Early administration of TXA to bleeding trauma patients is likely to be highly cost-effective in low-, middle- and high-income settings.
The use of TXA may also reduce subsequent costs related to surgical procedures for PPH treatment (such as laparotomy) as well as any complications associated with surgery.
Out-of-pocket costs to individual women might be higher when TXA is added to standard care for PPH in settings where women incur financial costs for births.
Resource requirements
How large are the resource requirements for administering TXA in addition to standard care for PPH treatment compared to standard care alone?
View in own window
| Judgement |
|---|
□
Don’t know | □
Varies | □
Large costs | □
Moderate costs | ☒
Negligible costs or savings | □
Moderate savings | □
Large savings |
Certainty of evidence on required resources
What is the certainty of the evidence on costs?
Cost-effectiveness
Does cost-effectiveness favour TXA + standard care or standard care alone?
View in own window
| Judgement |
|---|
□
Don’t know | □
Varies | □
Favours the standard care alone | □
Probably favours the standard care alone | □
Does not favour either the TXA + standard care or the standard care alone | ☒
Probably favours TXA + standard care | □
Favours TXA +standard care |
Equity
What would be the impact on health equity of TXA administration in addition to standard care for PPH treatment?
Research evidence
No direct evidence of the impact of the TXA administration in addition to standard care for PPH treatment on equity was found. However, indirect evidence from a review of barriers and facilitators to facility-based birth indicates that poor quality of care, as evident by poor birth outcomes, is probably a significant barrier to the uptake of facility birth by women in LMICs.
17
Additional considerations
The 2015 WHO State of Inequality report indicates that women who are poor, least-educated, and reside in rural areas have lower health intervention coverage and worse health outcomes than more advantaged women.
18 Therefore, reducing maternal deaths due to bleeding through scaling up of TXA for PPH treatment could have a positive impact on health equity and improve outcomes among disadvantaged women, especially in LMICs where these women are at significantly higher risk of PPH-related maternal deaths.
Reducing the need for expensive, life-saving surgical interventions (such as laparotomy to stop bleeding in women with vaginal birth) through an IV medication would probably reduce inequities, especially in contexts where health services are covered through out-of-pocket means.
View in own window
| Judgement |
|---|
□
Don’t know | □
Varies | □
Reduced | □
Probably reduced | □
Probably no impact | ☒
Probably increased | □
Increased |
Acceptability
Is TXA (in addition to standard care) acceptable to key stakeholders (women and healthcare providers) for PPH treatment?
The intervention is likely to be acceptable to both women and healthcare providers. TXA is administered in adequately equipped health facilities (providing emergency obstetric care) by a skilled healthcare provider via a standard IV infusion over a short period of time. There is no evidence of adverse maternal or neonatal effects. The balance between benefits and harms suggests that TXA will be acceptable to key stakeholders (women, providers and policy makers) across settings. An incremental cost with substantial benefits in terms of saving lives would be generally acceptable.
Feasibility
Is TXA feasible to implement in addition to standard care for PPH treatment?
The use of IV TXA for treatment of PPH in healthcare facilities was regarded by the GDG as feasible. Standard IV infusion equipment is required, as well as healthcare providers with sufficient training to safely administer IV bolus infusions (similar to oxytocin infusion). Many hospitals already have access to TXA due to its common use for trauma and surgery. Available preparations are compatible with recommended dosing regimens for PPH treatment. In many healthcare facilities (including in LMICs) no (or minimal) additional resources, infrastructure or training is required to commence using TXA for this indication. Administration of TXA should be relatively easy to integrate into standard PPH treatment packages. It is listed on the WHO Model List of Essential Medicines under medicines affecting coagulation.
The successful implementation of the WOMAN trial in 193 hospitals in 21 countries, which recruited over 20 000 women, in itself can be considered a potential demonstration of the feasibility of implementing this intervention.9 The pragmatic nature of the trial, coupled with the variations in the capacities of participating institutions (from low to very high) also supports feasibility across low-, middle- and high-income settings. These hospitals are likely to implement a recommendation of TXA easily.
However, given that evidence currently supports IV TXA for treatment, the intervention may not be feasible in settings where IV administrations are restricted to doctors working in high-level or referral facilities.
