Tranexamic Acid Prophylaxis in Placenta Praevia Caesarean Delivery

Key Takeaways

• Postpartum haemorrhage was less frequent with tranexamic acid than with placebo in women with placenta praevia undergoing caesarean delivery.
• Findings were similar across age, placenta praevia type, antepartum haemorrhage, and placenta accreta spectrum analyses, and sensitivity checks pointed in the same direction.
• Operating-room adverse events and serious adverse events through six weeks were reported at similar rates in both groups.

Women with placenta praevia undergoing caesarean delivery had postpartum haemorrhage in 29.7% with prophylactic tranexamic acid and 35.1% with placebo in a multicentre BMJ phase 3 trial. Both groups also received prophylactic oxytocin, and study treatment began within five minutes of umbilical cord clamping. This compared prophylactic tranexamic acid plus oxytocin with placebo plus oxytocin in a high risk obstetric population. The reduction was statistically significant, although the absolute difference was modest.

A randomised, double blind, placebo controlled, phase 3 trial across 24 maternity units in China enrolled 1732 women with confirmed placenta praevia undergoing caesarean delivery from July 2023 to March 2025. Placenta accreta spectrum was diagnosed in 303 participants, or 17.9%. The intervention was tranexamic acid 1 g in 10 mL, diluted in 40 mL normal saline and infused intravenously over 10 minutes. The comparator was placebo 10 mL normal saline, and both groups received prophylactic oxytocin 10 IU immediately after delivery. The primary outcome was postpartum haemorrhage, defined as calculated estimated blood loss of at least 1000 mL or red cell transfusion within two days after delivery.

For the primary endpoint, postpartum haemorrhage occurred in 29.7% (251/845) of the tranexamic acid group and 35.1% (297/846) of the placebo group. The relative risk was 0.85, with a 95.2% confidence interval of 0.75 to 0.96 and P=0.01. Calculated estimated blood loss of at least 1000 mL occurred in 25.7% versus 31.6%, with a relative risk of 0.81 and 95% confidence interval 0.70 to 0.93. Red cell transfusion within two days occurred in 18.8% versus 21.6%, with a relative risk of 0.88 and 95% confidence interval 0.74 to 1.03. The number needed to treat for the composite outcome was 19, and the investigators described the benefit as statistically significant but modest.

Prespecified subgroup analyses showed no statistically significant variation by maternal age, placenta praevia type, or placenta accreta spectrum, and a post hoc analysis by antepartum haemorrhage also showed no statistically significant variation. Post hoc sensitivity analyses excluding delayed dosing or open-label tranexamic acid were consistent with the primary analysis. The protocol recommended delaying open-label tranexamic acid for at least 30 minutes if excessive bleeding occurred during surgery. Primary outcome data were available for 99.8% of the remaining women, or 1691 of 1694 participants. The direction of effect was similar across these analytic checks.

Operating-room adverse events occurred in 2.2% of women given tranexamic acid and 1.7% of women given placebo. Serious adverse events through six weeks occurred in 0.5% of each group, 4 of 837 versus 4 of 845, with RR 1.01 and 95% CI 0.25 to 4.00. Severe adverse effects, including thromboembolic events, seizures, acute kidney or liver injury, and maternal death, were monitored until six weeks postpartum. The single-country setting, partly recall-based follow-up, and exploratory secondary outcomes without multiplicity adjustment set the scope of the findings. In this setting, the regimen was associated with a modest reduction in postpartum haemorrhage without an observed increase in serious adverse events.