Study protocol. TRAAPREVIA—TRAnexamic acid for preventing blood loss following a cesarean delivery in women with a placenta pREVIA or low-lying placenta: a multicenter randomized, double blind, placebo controlled trial

Abstract

Background

Placentas that are previa or low-lying are a major cause of severe postpartum hemorrhage (PPH). Tranexamic acid, by inhibiting the fibrinolytic pathway and protecting blood clots from degradation, is a promising drug for preventing blood loss after childbirth, especially in high-risk conditions. It remains unclear whether tranexamic acid would decrease the incidence of severe PPH among women with placentas that are previa or low-lying.

Methods and design

This multicenter, double-blind, randomized controlled trial with two parallel groups will include 1380 women with placenta previa or a low-lying placenta and a cesarean delivery at a term ≥ 32 weeks, modeled on our previous study of tranexamic acid administered after cesarean deliveries (TRAAP2). Women with high antenatal suspicion of placenta accreta spectrum will not be included. Treatment (either tranexamic acid 1 g or placebo) will be administered intravenously just after birth. All women will also receive a prophylactic uterotonic agent. The primary outcome will be the incidence of a red blood cell transfusion before discharge. This study will have a 90% power to show a 33% reduction in the incidence of transfusion, from 20.0% to 13.4%.

Discussion

This large multicenter, randomized placebo-controlled trial aims to determine with adequate power if the prophylactic use of tranexamic acid among women with cesarean delivery and a placenta that is previa or low-lying would decrease the incidence of transfusion.

Trial registration

ClinicalTrials.gov NCT04304625 (March 9, 2020).

Background

Rationale

Placenta previa: a major cause of severe postpartum hemorrhage

Definition

Before 2014, the term placenta previa referred to a placenta that overlies or is proximal to the internal cervical os. The placenta normally implants in the upper uterine segment. Among women with placenta previa, the placenta either totally or partially lies within the lower uterine segment. In the past, previas were categorized into 4 types [1]: (i) complete placenta previa, where the placenta completely covers the internal os; (ii) partial placenta previa, where the placenta partially covers the internal os; (iii) marginal placenta previa, which just reaches the internal os, but does not cover it; and (iv) low-lying placenta, which extends into the lower uterine segment, does not reach the internal os but is within 2 cm of the internal os. In response to the widespread use of transvaginal ultrasonography, which allows precise localization of the placental edge and the cervical os, the nomenclature was modified in 2014 to eliminate the terms “partial” and “marginal”. Instead, all placentas overlying the os (to any degree) are called previa and those near but not overlying the os (within 2 cm of the internal os) low-lying [2].

Mode of delivery

A placenta that covers the internal os partially or completely (placenta previa) requires cesarean delivery [3, 4]. The optimal delivery of low-lying placentas that do not overlie the internal os (are within 2 cm of it) remains unknown, however. Some experts propose an elective cesarean for all such women, others only for those with an internal os distance of 1–10 mm, and still others recommend a trial of labor for all women with asymptomatic third-trimester low-lying placenta, depending on their obstetric history and preferences [1, 3–7]. In a retrospective multicenter study, we showed that the rate of cesarean delivery for women with trial of labor was 50% for women with an internal os distance of 11–20 mm and 81.5% in those with a distance of 1–10 mm. Most women with low-lying placenta ultimately have cesarean deliveries [6].

Incidence and maternal morbidity

The estimated incidence of placenta previa in the United States (USA) has been estimated at 1 in 200 pregnancies at term [1, 2, 5, 7–15], mainly based on an old population-based study covering the years 1979–1987 that found its overall annual incidence to be 4.8 per 1000 deliveries (0.48%) [9]. More recent data demonstrate that the placenta previa rate has continued increasing [8, 16], accounting for 1.3% of US pregnancies in 2007 [10]. This increase in placenta previa frequency has accompanied the rise in cesarean deliveries, with some lag time [12–15]. One study used a decision-analytic model to calculate that should primary and secondary cesarean rates continue to rise as they have in recent years in USA, the annual incidence of placenta previa, placenta accreta, and maternal death will also rise substantially with 130 additional maternal deaths yearly by 2020, mainly related to abnormal placentation [12]. Deliveries complicated by placenta previa are at high risk of obstetric hemorrhage, in particular after delivery [1, 16]. Morbidities associated with placenta previa include need for hysterectomy (relative risk [RR] 33.26, 95% confidence interval [CI] 18.19–60.89), placenta accreta spectrum (RR 17.45, 95% CI 7.06–43.14), intrapartum hemorrhage (RR 2.48, 95% CI 1.55–3.98), postpartum hemorrhage (PPH) (RR 1.86, 95% CI 1.46–2.36), blood transfusion (RR 10.05, 95% CI 7.45–13.55), septicemia (RR 5.5, 95% CI 1.31–23.54), and thrombophlebitis (RR 4.85, 95% CI 1.50–15.69) [15]. In the USA, mothers die in 0.03% of cases of placenta previa [1, 9]. Maternal hemorrhagic morbidity (defined as death, blood product transfusion, atony requiring uterotonics, uterine or hypogastric artery ligation, hysterectomy, coagulopathy, and/or exploratory laparotomy) is more common in women with than without previa (19% versus 7%, aRR 2.6, 95% CI 1.9–3.5) [14]. A large prospective US study estimated the risk of intraoperative or postoperative blood transfusion in women with placenta previa at 15% in primary and 32.2% in repeat cesarean deliveries [17]. Thus, the incidence of placenta previa is rising with the cesarean delivery rate. Placenta previa is a major cause of severe PPH and consequently an important contributor to PPH-related severe maternal morbidity and a condition at high risk of death from bleeding.

Tranexamic acid

Outside of obstetrics

Tranexamic acid has antifibrinolytic effects due at least in part to its promotion of hemostasis by blocking lysine-binding sites on plasminogen molecules [18]. In clinical trials outside of obstetrics, tranexamic acid has been found to reduce both transfusions in elective surgery [19, 20] and mortality among patients with extracranial [21] or mild-to-moderate intracranial trauma [22].

Treatment of postpartum hemorrhage

Tranexamic acid was shown to reduce bleeding-related mortality among women with PPH [23], especially when administered shortly after delivery [23]. This finding suggested that tranexamic acid might be considered as an intervention not only to treat but to prevent postpartum coagulopathy [24].

Prevention of postpartum hemorrhage

General population of women with vaginal and cesarean delivery.

In a previous trial, we showed that prophylactic administration of tranexamic acid in addition to prophylactic oxytocin in women with vaginal delivery tended to reduce blood loss of at least 500 mL (RR, 0.83; 95% CI, 0.68–1.01; P = 0.07) and significantly reduced provider-assessed clinically significant PPH and the use of an additional uterotonic due to bleeding, compared to placebo [25]. In a subsequent trial, we also showed that among women who underwent cesarean delivery and received a prophylactic uterotonic agent, tranexamic acid treatment resulted in a significantly lower incidence of calculated estimated blood loss greater than 1000 mL than placebo (mean between-group difference of approximately only 100 mL). But the incidence of clinical hemorrhage-related secondary outcomes did not decline [26]. These results are consistent with those of a subsequent large US multicenter trial, in which tranexamic acid did not lead to a significantly lower rate of maternal death or blood transfusion (primary outcome) compared with placebo [27]. We also reported that prophylactic tranexamic acid for the prevention of blood after vaginal or cesarean delivery is likely cost-effective but with a very small effect size for both cost and effectiveness [28, 29]. Thromboembolic events did not differ significantly between the two groups in any of these trials, although the power to detect differences in the incidence of thromboembolic events was limited. Nevertheless, in women with moderate and severe anemia delivering vaginally, administration of tranexamic acid in the 15 min after the umbilical cord was clamped did not reduce the risk of clinically diagnosed PPH [30]. Finally, an individual patient data meta-analysis including these 4 large trials failed to show that prophylactic tranexamic acid improved maternal outcome after birth, although it did provide reassuring but underpowered data that the rate of thromboembolic events did not increase [31]. These equivocal results raise questions about the relevance of the prophylactic administration of tranexamic acid to all women undergoing cesarean and especially vaginal deliveries [18, 32–35].

High-risk population

Because the magnitude of the potential effect of a given preventive medication on maternal outcomes may vary greatly between low- and high-risk populations, the question of the effect of prophylactic tranexamic acid in decreasing the incidence of PPH among high-risk populations remains unanswered. We have shown that in women with a multiple pregnancy and cesarean delivery, prophylactic tranexamic acid did not reduce the incidence of any blood loss-related outcomes [36]. But no data are available regarding very high-risk populations, such as women with placenta previa or a low-lying placenta. We can speculate that the magnitude of the effect of tranexamic acid is lower in the general population than in a very high-risk population. On the other hand, the benefit-harm ratio of tranexamic acid may be unfavorable if the risk of thromboembolism, which is inherent in massive bleeding [4, 37–39], is significantly higher in this high-risk population and if it simultaneously shows no significant benefits in reducing blood loss-related outcomes.

Our hypothesis is that a low dose of tranexamic acid administered in women with placenta previa or a low-lying placenta who have cesarean deliveries and prophylactic uterotonics decreases the incidence of postpartum blood transfusion.

Objectives

The primary aim of our study is to compare the effect of a low dose of tranexamic acid (1 g), versus placebo, administered within 3 min after the infant's delivery along with a prophylactic uterotonic in women with placenta previa or a low-lying placenta who have cesarean deliveries, on the incidence of red blood cell (RBC) transfusion between the delivery and maternal discharge.

To facilitate the reading of this protocol, the term “placenta previa” used hereafter will refer to both placenta previa (i.e., placenta overlies the os) and low-lying placenta (in which the edge is within 2 cm of the internal os), unless clarity appears to require distinguishing them.

The secondary objectives, similar to but modified from those for the TRAAP trial [25, 40] and particularly the TRAAP2 trial [26, 41], which included women with cesarean deliveries but without placenta previa, are as follows:

• Secondary outcome measures assessing postpartum blood loss

  • Proportion of women who received between delivery and maternal discharge a transfusion of blood products, defined as any of the following: packed RBCs, fresh frozen plasma, and platelets
  • Incidence of quantitative (i.e., gravimetrically estimated) postpartum blood loss > 1000 mL, assessed by measuring the suction volume and swab weight [42, 43]
  • Incidence of quantitative blood loss > 1500 mL, assessed by measuring the suction volume and swab weight [42, 43]
  • Incidence of quantitative blood loss > 2000 mL, assessed by measuring the suction volume and swab weight [42, 43]
  • Mean quantitative blood loss, assessed by measuring the suction volume and swab weight [42, 43]
  • Incidence of calculated estimated blood loss > 1000 mL. Calculated blood loss = estimated blood volume × (preoperative hematocrit-postoperative hematocrit)/preoperative hematocrit (where estimated blood volume = body weight (kg) × 85) [26, 44]. Preoperative hematocrit will be the most recent hematocrit in the 7 days before delivery. Postoperative hematocrit will be that measured at day 2 postpartum. For women who receive a transfusion before the blood sample is taken, the value of the postoperative hematocrit will be calculated as the hematocrit on day 2 (%) − (3 × number of units of RBCs transfused) [26, 37]
  • Incidence of calculated estimated blood loss > 1500 mL [26, 44]
  • Incidence of calculated estimated blood loss > 2000 mL [26, 44]
  • Mean calculated estimated blood loss
  • Incidence of provider-assessed clinically significant PPH
  • Mean shock index measured at 15, 30, 45, 60, and 120 min after birth [45], or if and when PPH occurs
  • Proportion of women requiring supplementary uterotonic treatment, including sulprostone
  • Proportion of women who receive RBC transfusion between the child's birth and 24 h postpartum
  • Mean number of RBC units transfused during the postpartum hospital stay
  • Incidence of intrauterine hemorrhage-control devices including intrauterine balloon tamponade and vacuum-induced hemorrhage control
  • Incidence of arterial embolization or emergency surgery for PPH including vessel ligation, uterine compression suture or peripartum hysterectomy
  • Incidence of peripartum hysterectomy
  • Incidence of maternal transfer to intensive care unit
  • Maternal death from any cause
  • Mean change in peripartum hemoglobin and hematocrit (difference between most recent hemoglobin and hematocrit in the 7 days before the cesarean and at day 2 postpartum)
  • Proportion of women who receive an iron sucrose perfusion between the child's birth and maternal discharge
  • Proportion of breastfeeding at hospital discharge (as PPH may impact the woman’s ability to establish or maintain breastfeeding [46]).
• Outcomes related to potential adverse effects of tranexamic acid (1 g) after cesarean delivery:

  • Hemodynamic parameters (heart rate, blood pressure) 15, 30, 45, 60, and 120 min after delivery [25, 26, 40, 41].
  • Proportion of women with nausea, vomiting, phosphenes, dizziness (these events will be identified in the operating room as well as during the postpartum hospital stay).
  • Proportion of women with thromboembolic events (deep vein thrombosis or pulmonary embolism confirmed by radiological exams), myocardial infarction, seizure, renal failure necessitating dialysis, and other severe unexpected adverse reactions (these events will be assessed up to 12 weeks after delivery by a telephone interview at 12 weeks postpartum).

• Outcomes related to women’s satisfaction and psychological status at day 2 and 8 weeks postpartum [25, 26, 47–50].

Methods and design

The methodology and design are very similar to those of the TRAAP2 trial conducted in women with cesarean deliveries but without placenta previa, modified slightly where appropriate [26, 41].

Recruitment and allocation

Recruitment and allocation, including inclusion and exclusion criteria, are essentially identical to those reported previously for TRAAP1 [25, 40] and TRAAP2 [26, 41], modified for a population with cesarean deliveries with placenta previa instead of cesarean or vaginal deliveries without placenta previa.

Inclusion criteria

• Age ≥ 18 years

• Placenta previa, defined by a placenta covering the cervical os at the most recent transvaginal ultrasound examination before delivery, as per French guidelines [4, 6] or low-lying placenta, defined as a distance between the cervical os and the placenta (internal os distance) less than 20 mm [2]

• Cesarean delivery before or during labor

• Gestational age at delivery ≥ 32 weeks gestation

• Signed informed consent

Exclusion criteria

• History of venous (deep vein thrombosis and/or pulmonary embolism) or arterial (angina pectoris, myocardial infarction, or stroke) thrombosis

• History of epilepsy or seizure

• Any chronic or acute cardiovascular disease, chronic or acute renal disease (including chronic or acute kidney failure with glomerular filtration rate < 90 mL/min), chronic or acute liver disorder with hemorrhagic or thrombotic risk

• Active autoimmune disease with thromboembolic risk (including lupus, antiphospholipid syndrome, and Crohn’s disease)

• Sickle cell disease (homozygous)

• Severe hemostasis disorder

• High antenatal suspicion of placenta accreta spectrum disorder

• Placenta previa, diagnosed during delivery

• Placental abruption

• Acute bleeding of more than 500 mL in the 12 h before the cesarean

• Eclampsia or HELLP (Hemolysis Elevated Liver Low Platelet) syndrome

• In utero fetal death

• Administration of low-molecular-weight heparin or antiplatelet agents in the week before delivery

• Known hypersensitivity to tranexamic acid or concentrated hydrochloric acid

• Women under legal protection

• Poor understanding of the French language.

In France, a free routine ultrasound between 30–34 weeks is offered to all women for the assessment of fetal biometry, morphology, and placental localization. In case of placenta previa (including a low-lying placenta), another ultrasonography is performed at about 36 weeks of gestation to confirm its persistence, if delivery has not occurred [2, 6]. Information about the study will be provided from 32 weeks'gestation to all women who meet the inclusion criteria by obstetricians and midwives during prenatal visits or by anesthetists during the systematic anesthesia visit, or both. Women will confirm their participation and provide signed informed consent before delivery even if a trial of labor is planned for women with a low-lying placenta. This will facilitate recruitment in the trial, particularly for cesareans during labor, which may be decided upon and performed in emergency situations, and cesareans performed before the planned date because of the occurrence of an acute event.

Women will be randomized to receive either 1 g of tranexamic acid (Sanofi-Aventis, Paris, France; Marketing authorization number: 3400931157618 [1974, RCP rev 06.08.2018]) or placebo (normal saline, Fresenius Kabi, Sèvres, France; Marketing authorization number: 3400936751293, RCP rev 02.01.2018), at a 1:1 ratio. The Angers Clinical Research Unit will create and securely hold the randomization list for allocations to the tranexamic acid and placebo groups. As in the previous TRAAP and TRAAP2 trials [25, 26, 40, 41], a list of treatment units, containing the type of product (tranexamic acid or placebo) and the corresponding treatment number, will be forwarded to the pharmacy department of the Angers University Hospital, a member of the PPRIGO hospital pharmacists’ consortium (Production Pharmaceutique pour la Recherche Institutionnelle du Grand Ouest). The blinded products will be supplied by the Angers CHU pharmacy according to GMP for hospital pharmacies (BPP 2007), which will provide for each participant a 10-mL glass vial of the study drug (1 g of tranexamic acid or placebo according to the randomization list) in a numbered, labeled box. All vials and boxes will be identically labeled, and the drug packs will be differentiated only by the treatment number. Accordingly, all participants, including caregivers, will be masked and the women’s safety simultaneously guaranteed.

The randomization will be centralized and stratified by center. Randomization numbers will be attributed through a web platform (Ennov Clinical Software). Once a woman has been included and a randomization number assigned to her, she will retain this number even if she withdraws from the study.

Intervention

The intravenous administration of a 10-mL blinded vial of the study drug (either 1 g of tranexamic acid or placebo, depending on the randomization group) will be the intervention, as in the TRAAP2 trial [26, 41]. The anesthesiologist or nurse-anesthetist will administer it slowly (over 30–60 s), within three minutes after the birth, after injection of the routine prophylactic and cord clamping.

As previously reported [26, 41], all aspects of the management of cesarean delivery, including the third stage, will be identical in both arms, except for the study drug vial content. It will therefore comprise the routine prophylactic intravenous injection of 5 or 10 IU oxytocin or 100 µg carbetocin (according to the hospital's policy) at delivery within the three minutes after birth, as the national clinical practice guidelines issued by the French College of Gynecologists and Obstetricians recommend [4].

Other aspects of the cesarean procedure

• Surgeons will be requested to use a standardized procedure for the surgery:

  • ◦ The association of the Joel-Cohen technique and its variants (such as Misgav-Ladach technique) with reduced blood loss compared to the Pfannenstiel abdominal incision (and its variants) warrant a preference for the application of the Joel-Cohen and its variant approaches [51, 52].
  • ◦ In the absence of severe bleeding, cord traction will be the preferred method for placental delivery during the cesarean delivery, given its association with less blood loss than manual removal [53].
  • ◦ Because prevention of an incision of the anterior placenta previa has been shown to reduce the frequency of intraoperative hemorrhage (> 1000 mL) [54], clinicians will be asked to perform a partial manual separation of the placenta and then an amniotomy, as in accordance with guidelines of the French College of Gynecologists and Obstetricians [4].

• Health care providers will apply the French guidelines [4] and their center’s protocol in deciding whether to administer additional uterotonics and the other aspects of PPH management. In particular, practitioners are authorized to use tranexamic acid for the treatment of PPH consistent with their center’s protocol.

• The following standardized methods for quantifying blood loss will be standardized and include:

  • ◦ Recording the volume of all blood collected by suction.
  • ◦ Weighing all swabs or material used to measure additional blood loss in the operating room and the postanaesthetic care unit (PACU).

• Finally, the obstetrician's level of experience (categorized as junior or senior based on years of specialist training) will be recorded because it may affect the estimated blood loss during or after cesarean delivery [44].

Outcome measures

Primary outcome measure

The trial's primary outcome will be the incidence of RBC transfusion between the child's delivery and the mother's discharge from her postpartum hospital stay.

We chose this primary outcome for the following reasons: assessment of postpartum blood loss is challenging after all births [55] but particularly in the context of cesarean delivery, and its accuracy has been questioned, regardless of the method used [41, 56–58]. Moreover, RBC transfusion is considered as significant maternal morbidity, equivalent to blood loss of 1000 mL or more [3, 4, 59, 60]; it is rare that patients receive RBC transfusion for a blood loss less than 1000 mL [25, 26, 61, 62]. It is included in the core set of outcomes recommended for studies evaluating interventions for PPH prevention [46] and was the main component of the primary outcome of a recent large US RCT assessing tranexamic acid for the prevention of blood loss among women with cesarean delivery [27].

Secondary outcomes

Secondary outcome measures assessing postpartum blood loss

Clinical

• Proportion of women requiring a transfusion of blood products, defined as any of the following: packed red blood cells, fresh frozen plasma, and platelets (until discharge).

• Mean quantitative (i.e. gravimetrically estimated) blood loss, by measuring the suction volume and swab weight according to Gai et al. (estimated blood loss = (weight of materials used + materials not used—weight of all materials before surgery)/1.05 + volume included in the suction container) [26, 42], recorded from placental delivery until the woman's discharge from the PACU), i.e. about two hours without complications after the end of the cesarean delivery.

• Quantitative blood loss > 1000 mL

• Quantitative blood loss > 1500 mL

• Quantitative blood loss > 2000 mL

• Incidence of provider-assessed clinically significant PPH, defined by the providers’ response to the question: “Was there a PPH?” [25, 26, 40, 41]

• Mean shock index, defined by the ratio of heart rate to systolic blood pressure, measured at 15, 30, 45, 60, and 120 min after birth, or if PPH occurs [45]

• Proportion of women requiring supplementary uterotonic treatment, including sulprostone [25, 26, 40, 41]

• Proportion of women receiving an RBC transfusion from the child's birth to 24 h postpartum

• Mean number of RBC units transfused until discharge

• Incidence of intrauterine hemorrhage-control devices including intrauterine balloon tamponade or vacuum-induced hemorrhage control

• Incidence of arterial embolization and emergency surgery for PPH including vessel ligation, uterine compression suture or peripartum hysterectomy [25, 26, 40, 41]

• Incidence of peripartum hysterectomy

• Incidence of transfer to intensive care unit (ICU)

• Incidence of maternal death from any cause [25, 26, 40, 41]

• Incidence of iron sucrose perfusion until discharge

• Incidence of breastfeeding at hospital discharge (as PPH may impact the woman’s ability to establish or maintain breastfeeding [46])

Laboratory

• Incidence of calculated blood loss > 1000 mL, 1500 mL and > 2000 mL

• Mean total calculated blood loss

• Mean peripartum change in hemoglobin (difference between the most recent hemoglobin within one week before delivery and at day 2 postpartum) [25, 26, 40, 41]

• Mean peripartum change in hematocrit (difference between the most recent hematocrit within one week before delivery and at day 2 postpartum) [25, 26, 40, 41].

For all laboratory indicators, the predelivery reference examination will be the most recent blood count obtained in the week before delivery. As previously described [25, 26, 40, 41], all patients included in the trial will provide a blood sample on the second day postpartum (D2) to measure their peripartum hemoglobin and hematocrit and calculate the change in these two indicators. If no blood sample is available from D2, a D3 blood sample will be used, if available. If no blood sample is available from D2 or D3, the blood sample closest to D2 in the absence of transfusion will be used.

The occurrence of potential adverse effects of tranexamic acid, most as previously reported [25, 26, 40, 41], with some modifications.

Clinical

• Hemodynamic parameters (heart rate, blood pressure) 15, 30, 45, 60, and 120 min after delivery [25, 26, 40, 41]

• The occurrence of potential mild adverse effects of tranexamic acid in the operating room:

  • nausea
  • vomiting
  • phosphenes
  • dizziness
• The occurrence of potential severe adverse effects of tranexamic acid during the hospital stay and up to 12 weeks postpartum [25, 26, 40, 41]:

  • deep vein thrombosis, if the diagnosis is confirmed by Doppler ultrasound [25, 26, 40, 41]
  • pulmonary embolism, if the diagnosis is confirmed by radiological examination [25, 26, 40, 41]
  • myocardial infarction [25, 26, 40, 41]
  • seizure [25, 26, 40, 41]
  • renal failure requiring dialysis [25, 26, 40, 41]
  • any other unexpected adverse events [25, 26, 40, 41].

As previously [25, 26, 40, 41], the medical team will check these events during hospitalization and then again at 12 weeks postpartum, by a telephone interview of each woman. To minimize loss to follow-up, at least 10 calls at different hours over the course of the week will be attempted. Should women report severe adverse effects after discharge, objective data will be collected from medical files, transmitted either by the woman or her general practitioner.

• Women’s satisfaction and psychological status

  • This will be assessed by a self-administered questionnaire on D2 postpartum and by mail or email at 8 weeks postpartum [25, 26, 40, 41, 47–50].

Statistical analysis

In line with our previous TRAAP and TRAAP2 trials [25, 26, 40, 41], data analysis and reporting will follow the CONSORT guidelines for randomized controlled trials and will be conducted with the trial statistician and researchers blinded to randomization group. The two groups will be compared for the women's demographic characteristics and standard PPH risk factors. The modified intention-to-treat (ITT) population will include the women who undergo randomization and cesarean delivery, excluding those who withdraw consent or are deemed ineligible post-randomization. The main analysis of the primary and secondary outcomes will be performed in this population. All women receiving either tranexamic acid or placebo will be included in the safety population.

Two separate per-protocol populations will also be analyzed. Per-protocol group 1 will comprise the women from the modified ITT population receiving a uterotonic and then tranexamic acid or placebo in the three minutes after delivery, as prespecified in the protocol. For a situation closer to routine clinical practice and thus more generalizable, per-protocol group 2 will include women from the modified ITT population receiving a uterotonic and then tranexamic acid or placebo in the 10 min after delivery.

Descriptive statistics will be used to compare the participants'baseline characteristics, management of the third stage of labor, and protocol adherence. Means with standard deviations and medians with interquartile ranges will be used to express quantitative (continuous) variables, which will be compared respectively by Student’s t-test and the Wilcoxon rank-sum test. Chi-square or Fisher’s exact tests will be used, as appropriate, to compare categorical (qualitative) variables. The effects of tranexamic acid will be reported for quantitative variables as mean differences with 95% confidence intervals and for categorical outcomes as relative risks with 95% confidence intervals. Binary outcomes will also be expressed as absolute risk differences with 95% confidence intervals.

Two prespecified subgroup analyses will be used to test the effect of tranexamic acid on the primary outcome in subgroups characterizing various contexts of care: the risk of bleeding by the type of placental location (low-lying or placenta previa), and the occurrence of prenatal bleeding less than 500 mL in the 12 h before cesarean (yes or no).

Feasibility

The large majority of the participating centers have previously collaborated in the TRAAP and TRAAP2 trials [25, 26]. Moreover, all participating centers belong to a national network called Groupe de Recherche en Obstétrique et Gynécologie (GROG) and have recently participated successfully in several RCTs. All have demonstrated their ability to manage large RCTs assessing perinatal outcomes [63–68].

Sample size

Based on the results from the PACCRETA study, a population-based study of women with placenta previa conducted in 9 French regions in 2014–2016 [69, 70], we assumed a 20% incidence of our primary outcome (RBC transfusion before maternal discharge) in the absence of tranexamic acid. Previous trials showing a preventive effect of tranexamic acid in various surgical contexts at risk of bleeding outside obstetrics have always reported an effect magnitude greater than or around one third for the reduction in the proportion of women requiring transfusion [19]. The study thus requires 670 women in each group, and a total of 1340 participants to show a relative reduction of at least one third (33%) in the incidence of the primary outcome in the tranexamic acid arm — that is, an incidence of 13.4% or less in this arm, with α = 0.05, 1 − β = 0.90, and a bilateral test. We hypothesize that 3% of the included women will withdraw consent or will lack a primary outcome information. The study will therefore include 690 women in each group for a total of 1380 participants.

Timetable

The participating maternity units totalize about 117,600 deliveries per year. Based on the French national data from the French Perinatal Survey 2016 [71], the proportion of eligible women was estimated to be 0.5% of deliveries; this is probably a low estimate since most of the participating units are reference university hospitals where women with placenta previa are referred for delivery from less specialized hospitals. Assuming a participation rate of at least 40%; we should be able to recruit 1380 patients during an inclusion period of 72 months.

Data management

Each investigator will be responsible for ensuring the accurate recording of the data, which will be completed by clinical research technicians (CRTs) throughout the trial, by using Clinsight software. The electronic case report file for each woman will contain 6 components, as previously described [25, 26, 40, 41]:

• One completed by the CRT from the obstetric file: woman’s characteristics, course of the pregnancy, labor, and delivery.

• One completed by the CRT about the postpartum events after leaving the delivery room and the results of the postpartum blood count.

• One completed by the surgeon to collect details related to the cesarean techniques that may not be recorded in the obstetric file as well as the surgeon’s characteristics.

• One questionnaire about women’s satisfaction on D2 postpartum, completed by the women, with responses entered secondarily in the electronic file by the CRT.

• One questionnaire about the women’s satisfaction and psychological status at 8 weeks postpartum, sent by the CRT to the women and completed by them, with responses entered secondarily in the electronic file by the CRT.

• One questionnaire about the occurrence of thromboembolic and any other unexpected events at 12 weeks postpartum, completed by the CRT during a telephone interview with the woman, with responses secondarily entered in the electronic file by the CRT.

The Methodology and Data Management Centre of the Bordeaux University Hospital will handle the data management and statistics centrally under the supervision of the study’s Scientific Director (CDT).

Quality control will adhere to the standard operating procedures of the sponsor, Bordeaux University Hospital. The research in the investigational centers and management of subjects will comply with the Declaration of Helsinki and Good Clinical Practices. CRTs will conduct regular visits to each investigative center and report to the Data and Safety Monitoring Committee (DSMC). During these onsite inspections and in accordance with Good Clinical Practices and as in the previous TRAAP studies [40, 41], the following items will be reviewed:

• Adherence with the research protocol and defined procedures

• Verification of patients’ informed consents for all included women

• Comparison of source documents with the data reported in the eCRFs for accuracy and consistency of the data and the missing data

• End-of-trial visit for archiving of research documents.

Trial steering committee

A trial steering committee (TSC) will be set up and will be responsible for overall supervision of the trial. It will meet before the trial starts and then at least every 6 months until the trial is completed. The TSC will meet within a month of every Safety Monitoring Committee meeting to consider their recommendations (as modified from [25, 26, 40, 41]).

Safety consideration

As in the TRAAP and TRAAP2 trials [25, 26, 40, 41], an independent Data and Safety Monitoring Committee will also be formed. Its members will meet yearly to examine recruitment figures, baseline data, and will review the result of the interim safety data analysis when half of the planned sample size has been recruited.

Suspected unexpected serious adverse reactions (SUSARs) will be recorded and reported with the ANSM (National Agency for Drug Safety) approved SUSAR form. As in the first TRAAP trials [25, 26, 40, 41], SUSARs include maternal death, surgery (other than cesarean delivery) in the 12 weeks after randomization, transfusion of more than 4 units of blood, ICU admission, deep vein thrombosis, pulmonary embolism, myocardial infarction, seizure, renal failure defined by the need for dialysis, or suspected drug reactions. In case of any SUSAR, the local coordinator will complete the form and transmit it to the trial coordinating center at Bordeaux University Hospital within 72 h. Copies of the form will then be sent to the trial statistician and the chair of the Safety Monitoring Committee. The Safety Monitoring Committee may ask for unblinding, while the investigators will remain blinded to the treatment allocation. The ANSM, the trial sponsor, and the Chair of the Ethics Committee will also be informed by the Safety Monitoring Committee if considered appropriate, in particular when SUSARS related to the Investigational Medicinal Products are suspected. Finally, if the Safety Monitoring Committee judges it necessary, it can recommend that the Scientific Committee stop the trial.

Discussion

Potential and implementation of the findings

Placenta previa is a well-known condition that can significantly increase severe PPH-related maternal morbidity and the risk of life-threatening postpartum bleeding, the incidence of which is increasing along with the worldwide rise in cesarean rates. Very few interventions have been specifically tested among women with placenta previa to reduce blood loss and maternal morbidity. Tranexamic acid is an inexpensive agent that is simple to administer and acts by inhibiting the fibrinolytic pathway and protecting blood clots from being degraded [34]. It can easily be added to the routine management of cesarean deliveries worldwide. It has been considered a promising candidate for prevention, although four large multicenter well-conducted and adequately powered randomized controlled trials report either no or a weak prophylactic effect on clinical secondary blood loss-related outcomes after childbirth [25–27, 30].

Nevertheless, some authorities recommend the use of the tranexamic acid for PPH prevention for women with cesarean delivery at high risk of PPH [72], such as those with placenta previa. It is possible that the effect magnitude of tranexamic acid for the prevention of blood loss is much higher among women at high rather than low or moderate risk for PPH. On the safety side, the possibility that the antifibrinolytic properties of tranexamic acid may increase the thromboembolic risk associated with it cannot be ruled out given the limited power to detect differences in the incidence of thromboembolic events in the trials, although the data available are reassuring [23–31, 73]. That is, it remains possible that tranexamic acid may be associated with a significant increase in the rate of thromboembolic events among women with cesarean delivery at high risk of transfusion, such as those with placenta previa, for both cesarean and massive hemorrhage requiring transfusion are risk factors for thrombosis. Therefore, although some authorities consider that prophylactic tranexamic acid in these high-risk conditions may be valuable because it may have a stronger effect, we cannot rule out the possibility that this medication may have an effect opposite to that intended, i.e., no prophylactic effect on blood loss, but an increase in the rate of postpartum thromboembolic events.

This large, multicenter, randomized placebo-controlled trial aims to determine with adequate power if the prophylactic use of tranexamic acid among women with cesarean delivery and placenta previa or low-lying placenta would decrease the incidence of transfusion. It also aims to provide safety data to attempt to determine if the benefits of the routine prophylactic use of tranexamic acid among those women significantly outweigh its risks for the prevention of transfusion.

Abbreviations

ANSM

Agence Nationale de Sécurité du Médicament et des produits de santé (National agency for drug safety)

BPP

Bonnes Pratiques de Pharmacie (Good Pharmacy Practices)

CI

Confidence interval

CHU

Centre Hospitalier Universitaire (University Hospital)

CONSORT

Consolidated standards of reporting trials

CPP

Comité de Protection des Personnes (Committee for protection of persons)

CRT

Clinical research technician

D

Day

DSMC

Data and safety monitoring committee

eCRF

Electronic Case Report Form

g

Gram

GMP

Good manufacturing practices

GROG

Groupe de Recherche en Obstétrique et Gynécologie (Obstetrics and gynecology research group)

HELLP

Hemolysis elevated liver low platelet

ICU

Intensive care unit

IU

International unit

ITT

Intention-to-treat

Kg

Kilogram

L

Liter

mL

Milliliter

PPH

Postpartum hemorrhage

PACU

Postanaesthetic care unit PHRC, Programme Hospitalier de Recherche Clinique (hospital clinical research program)

PPRIGO

production pharmaceutique pour la recherche institutionnelle du grand ouest (Pharmaceutical production for Western France institutional research)

RBC

Red blood cells

RCT

Randomized controlled trial

RCP

Résumé des Caractéristiques du Produit (Summary of Product Characteristics)

RR

Risk ratio

TRACOR

Traction of the cord

TRAAP

Tranexamic acid for preventing postpartum hemorrhage following a vaginal delivery

TRAAP2

Tranexamic acid for preventing postpartum hemorrhage after cesarean delivery

TSC

Trial steering committee

SUSARs

Suspected unexpected serious adverse reactions

US(A)

United States (of America)

WOMAN

World maternal antifibrinolytic

Authors' contributions

LS wrote the first draft of the manuscript, HM, AI, JMC and CDT reviewed the manuscript, with input from other members of the TRAAPREVIA Study Group.

Funding

The TRAAPREVIA trial is supported by a grant from the French Ministry of Health (National PHRC, 2018, 18–0461). The funding body has no role of the in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This research study was designed and is implemented in accordance with the Declaration of Helsinki’s principles on studies involving human subjects.

TRAAPREVIA trial was approved.

- by the Agence Nationale de Sécurité du Médicament et des produits de santé (ANSM) (national agency for drug safety), March 12, 2020.

- by the following Approval Committee: Comité de Protection des Personnes (CPP South East III) (Committee for protection of persons involved in biomedical research), March 5, 2020.

Informed written consent will be obtained from all participants.

Consent for publication

Not applicable.

Competing interests

Dr. Loïc Sentilhes reports receiving lecture fees from Norgine, and lecture and consulting fees from Ferring Pharmaceuticals, GlaxoSmithKline, Pfizer, Organon and Bayer. The other authors did not report any potential conflicts of interest.