Safety and efficacy of tranexamic acid for prevention of obstetric haemorrhage: an updated systematic review and meta-analysis

Abstract

Background

A number of clinical systematic review and meta-analysis have been published on the use of tranexamic in the obstetric setting. The aim of this meta-analysis was to evaluate the safety and effectiveness of tranexamic acid in reducing blood loss when given prior to caesarean delivery.

Materials and methods

We searched the Cochrane Wounds Specialized Register, Cochrane Central, MEDLINE (through PUBMED), Embase, and SCOPUS electronic databases. We also searched clinical trials registries for ongoing and unpublished studies, and checked reference lists to identify additional studies. We used no restrictions with respect to language and date of publication. Two review authors independently performed study selection, “Risk of bias” assessment, and data extraction. Initial disagreements were resolved by discussion, or by including a third review author when necessary.

Results

We found 18 randomised controlled trials (RCTs) that met our inclusion criteria. Overall, 1,764 women receiving intravenous tranexamic acid for prevention of bleeding following caesarean sections and 1,793 controls receiving placebo were enrolled in the 18 RCTs evaluated. The use of tranexamic acid compared to controls (placebo or no intervention) reduces post-partum haemorrhage >400 mL (risk ratio [RR] 0.40, 95% confidence interval [CI] 0.24–0.65; 5 trials with a total of 786 participants), severe post-partum haemorrhage >1,000 mL (RR 0.32, 95% CI: 0.12–0.84; 5 trials with a total of 1,850 participants), and need for red blood cell transfusion (RR 0.30, 95% CI: 0.18–0.49; 10 trials with a total of 1,873 participants). No particular safety concerns on the use of this antifibrinolytic agent emerged from the analysis of the 18 RCTs included.

Discussion

Overall, the results of this meta-analysis support the evidence of a beneficial effect of tranexamic acid in reducing blood loss and need for blood transfusion in pregnant women undergoing caesarean section.

Introduction

Obstetric haemorrhage is a leading cause of premature maternal mortality, accounting for at least 100,000 deaths each year worldwide^1–4. Although post-partum haemorrhage (PPH) may be unpredictable, the most common causes include uterine atony, abnormal placentation, retained placental tissue, and lacerations of the lower genital tract⁵. In addition, obesity, multiple pregnancies, and previous caesarean section have been recognised as risk factors for PPH⁶. Considering the health and social burden of PPH, it is not surprising that a number of recommendations and guidelines have been issued from national and international scientific societies and health authorities to optimise the use of obstetric interventions and uterotonic drugs in this critical clinical setting^7–9.

As far as the pathophysiology is concerned, recent evidence has linked the activation of the fibrinolytic pathway with the onset of severe haemorrhage in different settings, including trauma, heart and orthopaedic surgery, and obstetrics^10,11. Following these observations, several randomised controlled trials (RCTs) have assessed the impact of tranexamic acid (TA), a lysine analogue that inhibits plasmin-mediated fibrin degradation¹², on decreasing bleeding complications and mortality in such clinical conditions at increased haemorrhagic risk^13–21. The aim of this paper is to provide an up-dated review, through a systematic analysis of the existing literature on the published RCTs on the safety and efficacy of TA for prevention of postpartum blood loss.

Material and methods

This systematic review was conducted according to the recommended Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) checklist guidelines²².

Search strategy

A computer-assisted literature search of the MEDLINE (through PUBMED), EMBASE, SCOPUS, OVID and Cochrane Library electronic databases was performed to identify RCTs on the use of TA for prevention and treatment of PPH. A combination of the following text words was used to maximise search specificity and sensitivity: “tranexamic acid” AND “TXA” AND “antifibrinolytic agent” AND “post-partum haemorrhage” AND “PPH” AND “obstetric haemorrhage” AND “caesarean” AND “vaginal” AND “randomized” AND “prevention” AND “treatment”. In addition, we checked the reference lists of the most relevant items (original studies and reviews) in order to identify potentially eligible studies not captured by the initial literature search. Abstracts from relevant conferences or scientific meetings were hand-searched for additional studies.

Study selection and inclusion criteria

Study selection was performed independently by two reviewers (MF and MC), with disagreements resolved through discussion and on the basis of the opinion of a third reviewer (CM). Eligibility assessment was based on the title or abstract and on the full text if required. Articles were eligible if they reported either in the title or in the abstract the use of TA for the prevention of PPH. Only RCTs published in full in English between January 1970 and December 2017 were included in this systematic review and meta-analysis. As we found only two RCTs evaluating the efficacy of TA for the prevention of PPH after vaginal delivery^23,24, the present analysis was limited to RCTs evaluating this antifibrinolytic agent after caesarean delivery.

Data extraction and outcome analysis

For each study included in the systematic review, the following data were extracted by two reviewers (MF and MC) independently: publication date, sample size (TA and control groups), and protocol (TA dose administered). The primary outcome was the incidence of PPH (i.e. blood loss more than 400 mL) and severe PPH (i.e. blood loss >1,000 mL). Secondary outcomes included mean blood loss volume (mL), need for blood transfusion, and overall severe side effects related to TA (including thromboembolic events). Disagreement was resolved by consensus and by the opinion of a third reviewer, if necessary.

Assessment of risk of bias in included studies

Two review authors (MF and MC) independently assessed the risk of bias of each included study following the domain-based evaluation described in the Cochrane Handbook for Systematic Reviews of Interventions²⁵. They discussed any discrepancies and achieved consensus on the final assessment. The Cochrane “Risk of bias” tool addresses six specific domains: sequence generation, allocation concealment, blinding, incomplete data, selective outcome reporting, and other issues relating to bias. We have presented our assessment of risk of bias using two “Risk of bias” summary figures: 1) a summary of bias for each item across all studies; and 2) a cross tabulation of each trial by all of the “Risk of bias” items.

“Summary of findings” tables

We used the principles of the GRADE system to assess the quality of the body of evidence associated with specific outcomes, and constructed a “Summary of findings” table using REVMAN 5²⁶. These tables present key information concerning the certainty of the evidence, the magnitude of the effects of the interventions examined, and the sum of available data for the main outcomes²⁷. The “Summary of findings” tables also include an overall grading of the evidence related to each of the main outcomes using the GRADE approach, which defines the certainty of a body of evidence as the extent to which one can be confident that an estimate of effect or association is close to the true quantity of specific interest (see Online Supplementary Content, Table SI). The certainty of a body of evidence involves consideration of within-trial risk of bias (methodological quality), directness of evidence, heterogeneity, precision of effect estimates, and risk of publication bias²⁸.

We have presented the following outcomes in the “Summary of findings” table: i) PPH; ii) severe PPH; iii) need for blood transfusion (Table SI).

When evaluating the “Risk of bias” domain, we down-graded the GRADE assessment only when we classified a study as being at high risk of bias for one or more of the following domains: selection, attrition, reporting, and other bias; or when the “Risk of bias” assessment for selection bias was unclear (this was classified as unclear for either the generation of the randomisation sequence or the allocation concealment domain). For the outcomes PPH, need for blood transfusion, and post-partum blood loss, we did not down-grade for high risk of bias in performance and detection domains since we judged that the outcomes considered are not likely to be influenced by lack of blinding, and for unclear “Risk of bias” assessments in other domains.

Data analysis

All calculations were made using Stata 15.1, R v.3.4.3, and REVMAN 5. The effect size measures were the risk ratio (RR) and the risk difference (RD) between the treated arm and the control study arm. The study weight was calculated using the Mantel-Haenszel method. The heterogeneity χ² was calculated as the I² for the variation due to heterogeneity²⁵. If significant heterogeneity was detected, a random effect method of study weight calculation was followed (DerSimonian-Laird method)²⁹; otherwise, the fixed effect procedure was used. We also calculated the number needed to treat (NNT), which is the average number of patients who need to be treated in order to avoid (or to harm) an event.

Results

Literature search and study characteristics

In total, 4,567 articles were initially identified after the initial electronic and manual search, which was concluded on 7 January 2018 (Figure 1). Of them, 4,502 were excluded because they were focusing on other topics. Thus, 65 potentially relevant articles were selected and the next screening led to the exclusion of 47 additional studies (reviews, protocols of RCTs, not RCTs, studies containing no informative data). The remaining 18 randomised studies^30–47 were finally included in the systematic review and meta-analysis (see Table I for main characteristics and results of the included studies). Overall, 1,764 women receiving intravenous TA for prevention of bleeding following caesarean sections and 1,793 controls receiving placebo were enrolled in the 18 RCTs that went forward for evaluation.

Figure 1.

Flow chart of study inclusion criteria.

Table I.

Characteristics and main results of the 18 randomised controlled trials on the use of tranexamic acid for the prevention of obstetric haemorrhage.

First Author yearref.	Cases (IVTA/C)	Age, years1	IVTA dose	Controls	Post-partum blood loss, mL1	PPH onset	RBC transfusion	TA-related SAEs
Gai 200425	91/89	IVTA: 29.71 (4.18)	1 g 10 min before CS	No treatment	IVTA: 359.29 (152.02)2	IVTA: 22/913	NR	IVTA: 0/91
		C: 29.75 (4.01)			C: 439.36 (191.48)2	C: 35/893		C: 0/89
Sekhavat 200926	45/45	IVTA: 26.2 (4.7)	1 g 10 min before CS	5% glucose	IVTA: 28.02 (5.53)4	NR	NR	IVTA: 0/45
		C: 27.1 (4.1)			C: 37.12 (8.97)4			C: 0/45
Gungorduk 201127	330/330	IVTA: 26.3 (3.5)	1 g 10 min before CS	5% glucose	IVTA: 499.9 (206.4)5	IVTA: 7/3306	IVTA: 2/330	IVTA: 0/330
		C: 26.6 (3.6)			C: 600.7 (215.7)5	C: 19/3306	C: 7/330	C: 0/330
Movafegh 201128	50/50	IVTA: 27.0 (3.4)	10 mg/kg 20 min before SA	Normal saline	IVTA: 262.5 (39.6)2	NR	NR	IVTA: 0/50
		C: 27.6 (4.1)			C: 404.7 (94.4)2			C: 0/50
Xu 201329	88/86	IVTA: 26.7 (3.7)	10 mg/kg 20 min before SA	Normal saline	IVTA: 379.2 (160.1)2	IVTA: 19/887	IVTA: 8/88	IVTA: 2/88
		C: 27.1 (4.1)			C: 441.7 (189.5)2	C: 28/867	C: 19/86	C: 2/86
Shahid 201330	38/36	IVTA: 24.18 (3.93)	1 g 10 min before CS	Normal saline	IVTA: 356.44 (143.2)2	NR	IVTA: 3/38	IVTA: 0/38
		C: 24.89 (4.16)			C: 710.22 (216.72)2		C: 12/36	C: 0/36
Goswami 201331	60/30	IVTA: 23.6 (2.5)	10 or 15 mg/kg 20 min before skin incision	5% glucose	IVTA: 261.17 (56.78)–376.83 (31.96)8	IVTA: 0/606	IVTA: 0/60	IVTA: 0/60
		C: 24.3 (2.6)			C: 527.17 (88.67)8	C: 0/306	C: 2/30	C: 0/30
Senturk 201332	101/122	IVTA: 30.2 (6.83)	1 g 10 min before CS	5% glucose	IVTA: 272.05 (143.23)	NR	IVTA: 0/101	IVTA: 0/101
		C: 29.22 (6.93)			C: 346.87 (189.49)		C: 0/122	C: 0/122
Abdel-Aleem 201333	373/367	IVTA: 26.34 (5.16)	1 g 10 min before CS	No treatment	IVTA: 241.61 (126.02)2	IVTA: 2/3736	NR	IVTA: 0/373
		C: 26.62 (5.05)			C: 510.66 (144.52)2	C: 2/3676		C: 0/367
Ghosh 201434	70/70	IVTA: 25.94 (3.78)	1 g 10 min before CS	Normal saline	IVTA: 48.06 (8.20)4	NR	IVTA: 0/70	IVTA: 0/70
		C: 26.04 (3.39)			C: 76.01 (6.21)4		C: 3/70	C: 0/70
Singh 201435	100/100	IVTA: 25 (1.46)	1 g 20 min before CS	No treatment	IVTA: 270.05 (30.88)9	NR	NR	IVTA: 0/100
		C: 30 (1.24)			C: 510.45 (30.34)9			C: 0/100
Yehia 201436	106/106	IVTA: 28.4 (4.9)	1 g 20 min before CS	Normal saline	IVTA: 369.5 (198.0)9	IVTA: 33/1063	IVTA: 0/106	IVTA: 0/106
		C: 28.6 (4.7)			C: 606.8 (193.0)9	C: 67/1063	C: 2/106	C: 0/106
Gobbur 201437	50/50	IVTA: 23.62 (3.43)	1 g 20 min before CS	Normal saline	IVTA: 360.9 (110.3)2	IVTA: 6/507	NR	IVTA: 0/50
		C: 24.5 (3.98)			C: 443.0 (88.55)2	C: 15/507		C: 0/50
Ramani 201438	60/60	IVTA: 24.9 (3.9)	1 g 10 min before CS	Normal saline	IVTA: 222.07 (97.02)2	NR	IVTA: 2/60	IVTA: 0/60
		C: 24.4 (3.7)			C: 274.5 (179.2)2		C: 6/60	C: 0/60
Ahmed 201539	62/62	IVTA: 28.6 (5.9)	10 mg/kg 5 min before CS	Normal saline	IVTA: 391.0 (48.5)2	NR	NR	IVTA: 0/62
		C: 26.9 (5.2)			C: 596.7 (38.02)2			C: 0/62
Maged 201540	100/100	IVTA: 24.9 (4.6)	1 g 20 min before CS	Normal saline	IVTA: 459.4 (75.4)	IVTA: 0/1006	NR	IVTA: 0/100
		C: 25.3 (4.7)			C: 700.3 (143.9)	C: 6/1006		C: 0/100
Lakshmi 201641	60/60	IVTA: 26.77 (2.81)	1 g 20 min before CS	No treatment	IVTA: 347.17 (108.6)9	IVTA: 2/607	IVTA: 0/60	IVTA: 0/60
		C: 26.82 (2.8)			C: 517.72 (150.0)9	C: 36/607	C: 0/60	C: 0/60
Sujata 201642	30/30	IVTA: 29.4 (4.16)	1 g 15 min before CS	Normal saline	IVTA: 432 (337–497)5,10	IVTA: 0/306	IVTA: 1/30	IVTA: 0/30
		C: 30.27 (4.31)			C: 819 (663–1001)5,10	C: 7/306	C: 4/30	C: 0/30

IVTA: intravenous tranexamic acid; C: controls; SD: standard deviation; min: minutes; CS: caesarean section; SAEs: severe adverse events; RBC: red blood cell transfusion; NR: not reported; PPH: post-partum haemorrhage; SA: spinal anaesthesia.

¹Mean (standard deviation);

²measured from placental delivery to 2 hours post-partum;

³defined as blood loss >400 mL;

⁴measured from the end of cesarean section to 2 hours post-partum;

⁵measured from the skin incision to 48 hours post-partum;

⁶defined as blood loss >1,000 mL;

⁷defined as blood loss >500 mL;

⁸measured from placental delivery to 24 hours post-partum;

⁹measured from placental delivery to the end of caesarean section;

¹⁰median (interquartile range).

Risk of bias in included studies

Eleven studies were at high risk of bias for one or more domains, and 12 studies were at unclear risk of bias for one or more domains (Figure 2 A and B).

Figure 2.

Risk of bias graph and summary.

(A) Review authors’ judgements about each risk of bias item presented as percentages across all included studies. (B) Review authors’ judgements about each risk of bias item for each included study.

Allocation

We assessed two studies as being at high risk of selection bias, as randomisation was by alternation of the two treatments, so the intervention allocations could have been foreseen in advance^31,40. The reports of six studies were unclear for random sequence generation and/or allocation concealment, while ten studies were at low risk of selection biases.

Blinding

There were nine studies reported as open label, and they were graded as high risk of performance bias (blinding of participants and personnel). Eight studies were reported as double blind^{32–37,39,41,47}, and one⁴⁵ as single blind; one of these studies³⁷ did not provide any information on the blinding procedures. Eleven studies were graded at unclear risk of detection bias due to the fact that it did not provide information to permit judgement about “high” or “low” risk of bias related to the blinding of outcome assessors; one study⁴⁴ was graded at high risk of bias because it stated that the clinical care team was aware of the administered treatment.

Incomplete outcome data

One study⁴⁵ was judged at high risk of attrition bias because it reported only per protocol analysis. Two studies^31,44 were judged at unclear risk of bias. The remaining studies were judged at low risk of bias.

Selective reporting

Selective reporting was low in all included studies, and graded as “unclear” in one study³⁵ where some data were not presented in detail (e.g. some side effects) although the report states that there were no differences between groups.

Other potential sources of bias

We judged three studies at high risk of other source of bias: two because of imbalance at baseline^37,38, and one³⁶ because it did not mention PPH, enrolled anaemic women, and because there was a significant difference in the duration of surgery between groups.

Effects of interventions

The overall incidence of PPH was 71 cases on 395 treated patients (17.9%) and 172 cases on 391 control patients (43.9%). Using the average treatment effect from a random-effects model, the use of TA reduces significantly the episodes of PPH when compared with control group: five trials, 786 patients; RR 0.40, 95% CI: 0.24–0.65; p=0.0003 for overall effect; I²=68% (Figure 3A). The overall incidence of severe PPH was 9 cases of 893 treated patients (1.0%) and 34 cases of 857 control patients (3.9%). Using the average treatment effect from a random-effects model, the use of TA significantly reduces the episodes of severe PPH when compared with control group: five trials, 1, 750 patients; RR 0.32, 95% CI: 0.12–0.84; p=0.02 for overall effect; I²=19% (Figure 3B).

Figure 3.

Forest plot demonstrating effects of tranexamic acid (TA) on the incidence of (A) post-partum haemorrhage and (B) severe post-partum haemorrhage.

M-H: Mantel-Haenszel; CI: confidence interval.

As far as secondary outcomes are concerned, using the average treatment effect from a fixed-effects model, the use of TA reduces the need of red blood cell transfusion compared to controls: ten trials, 1,873 patients; RR 0.30, 95% CI: 0.18–0.49; p=0.00001 for overall effect; I²=0% (Figure 4). The overall incidence for blood transfusion was 16 cases of 943 treated patients (1.69%) and 55 cases of 930 control patients (5.91%). The NNT was calculated as 25.6.

Figure 4.

Forest plot of the effect of tranexamic acid (TA) on red blood cell (RBC) transfusion need. M-H: Mantel-Haenszel; CI: confidence interval.

Tranexamic acid reduces the amount (mL) of post-partum blood loss: mean difference −155.14 (95% CI: −192.69 to −157.58; p=0.00001 for overall effect) (Figure 5). The direction of the effect was consistent across studies, with a beneficial effect of TA, but there was a considerable heterogeneity (I²=100%). The heterogeneity observed can be ascribed to different intervals of blood loss recorded (from placental delivery^{30,33–35,38,42–44} or caesarean section³⁹ to two hours post-partum, from skin incision to 48 hours post-partum^32,47, from placental delivery to 24 hours post-partum³⁶, from placental delivery to the end of caesarean section^40,41,46); to different populations of patients enrolled (e.g. several studies excluded anaemic women, but one study³⁶ included only anaemic women); to different dosage of TA used; and also to methods used to quantify the loss (e.g. weight of gauze pads, and/or sponges, mops, drapes; bloods in suction bottles; pre- and post-intervention haemoglobin values).

Figure 5.

Forest plot of tranexamic acid (TA) on post-partum blood loss (mL).

CI: confidence interval.

One study³⁴ reported severe side effects (deep vein thrombosis) in 2 of 88 TA recipients and in 2 of 86 controls (RR, 0.98, 95% CI: 0.14–6.78). The remaining 17 studies did not report severe adverse events, including thromboembolic events, in either group.

Based on GRADE assessment, all these comparisons were graded as moderate certainty evidence, and down-graded once due to inconsistency or to risk of biases (see Online Supplementary Content, Table SI).

Discussion

The antifibrinolytic agent TA is routinely used in cardiac, orthopaedic and oral surgeries to reduce perioperative blood loss^13–15. Recent evidence from RCTs also indicate that TA usage results in a significant reduction of obstetric bleeding^20,48,49. A recent randomised, placebo-controlled trial (WOMAN, WOrld Maternal ANtifibrinolytic) on 20,060 women with PPH found that TA reduces deaths due to bleeding (RR 0.81, 95% CI: 0.65–1.00; p=0.045) with no adverse effects, especially when given early after bleeding onset⁵⁰. A Cochrane systematic review evaluating TA for preventing post-partum haemorrhage was recently published. After the analysis of 12 RCTs involving 3,285 women, the authors concluded that TA decreases postpartum blood loss and prevents PPH and blood transfusion requirements⁵¹. The results of our meta-analysis are in line with these observations and indicate that, in women undergoing caesarean delivery, the prophylactic use of TA significantly reduces the incidence of PPH, including severe PPH, total blood loss and transfusion requirements without increasing the risk of thromboembolic complications. The main strength of our study is that it represents a very large, up-to-date and comprehensive analysis which involved overall 4,557 women enrolled in 18 RCTs. In addition, most of the included studies are of high quality and with a low risk of bias according to the Cochrane risk of bias tools. Nevertheless, our systematic review has some limitations which are related to the different study designs of the RCTs evaluated that generate a substantial heterogeneity across studies. For instance, there were differences among the various RCTs in the definition of PPH (blood loss >400 or >500 mL) and post-partum blood loss assessment (2 hours, 24 hours or 48 hours post-partum). In particular, the abbreviated time for data collection may mean that total blood loss and the true incidence of PPH have been under-estimated. In addition, we were not able to detect the effect of TA on maternal death due to the lack of fatal events in the studies evaluated, which was probably related to the small size of the population of women enrolled. Larger RCTs on this clinical setting are, therefore, needed to assess this important outcome.

In summary, the results of our meta-analysis document the safety and efficacy of prophylactic administration of TA in reducing post-partum blood loss, PPH incidence and need for blood transfusion in women undergoing caesarean delivery. Therefore, given its efficacy in preventing one of the most common and serious complications of pregnancy, we recommend the use of TA in this clinical setting. The use of this drug in the framework of patient blood management (PBM) programmes^52–55 can play a key role as a strategy to save blood loss but, to be really beneficial in improving patient outcome and efficient for health systems, it should be part of an adequate management of perioperative anaemia^{56, 57}. In fact, the benefits of a PBM programme⁵⁸ are greater when it includes optimisation of patient’s haemoglobin level^59–65. This approach is very important also for the perinatal care of women, a setting where the management of anaemia and haematinic deficiencies should be obligatory both for clinicians and policymakers in charge of decision making processes aimed at up-dating clinical practice in health care⁶⁶.