Maternal death due to bleeding

Overview

Definition

Death of the woman for which the clinician recorded bleeding as the main immediate cause of death, counted from randomisation until death, discharge or 42 days, whichever came first.

Unit

women

Measured in

Studies in this atlas that report this outcome, by the claim each finding supports. Each finding keeps its own population, denominator and time window on the study page.

• Secondary endpoint WOMAN trial (World Maternal Antifibrinolytic Trial) · See the findings

  Applies to Maternal

  • Women aged 16 years and older with a clinical diagnosis of postpartum haemorrhage after a vaginal birth or caesarean section. WOMAN specified that the diagnosis could rest on clinically estimated blood loss of more than 500 mL after vaginal birth or 1000 mL after caesarean section, or any blood loss sufficient to compromise haemodynamic stability. This is the trial eligibility wording, not a WHO definition. The fundamental eligibility criterion was the clinician uncertainty about whether to use tranexamic acid in that woman. The overall result covers every treatment time.
  • Hospitals (193 in the paper).
  • Hospitals in 21 countries. The paper site list names Nigeria, Pakistan, Uganda, Kenya, Cameroon, Sudan, United Kingdom, Tanzania, Nepal, Zambia, Albania, Democratic Republic of Congo, Bangladesh, Ethiopia, Burkina Faso, Jamaica, Ghana, Papua New Guinea, Egypt, Colombia and Cote d'Ivoire.
  • Postpartum, treated within hours of birth or later
  • Gestational age Not reported
  • Postnatal age Not reported

  Follow-up Outcome data were collected at death, discharge or 42 days after randomisation, whichever came first.
  In the WOMAN trial, women with heavy bleeding after birth who were given tranexamic acid were less likely to die of bleeding than women given a placebo. 155 of 10 036 women (1.5%) died of bleeding with tranexamic acid, against 191 of 9985 (1.9%) with placebo. The risk ratio was 0.81 (95% confidence interval 0.65 to 1.00, p=0.045). The authors named death due to bleeding as the key secondary outcome. The upper end of the interval touches 1.00, and the primary outcome, death from any cause or hysterectomy, did not differ clearly between the groups (separate claim).

  Source problem. A cited source is corrected. Check the sources before relying on this claim.

  Limitations. Death due to bleeding is a cause-specific secondary outcome. The primary composite showed no clear difference. The upper confidence bound is 1.00. The planned sample of 15 000 was raised to 20 000 during the trial to estimate this effect, after hysterectomy decisions were found to be often made at the time of randomisation. The authors state that they pre-specified this outcome and the time-to-treatment subgroups before un-blinding. Clinicians recorded the immediate cause of death, the main cause if there were several. The trial was double-blind. The paper also gives an adjusted risk ratio of 0.78 (95% CI 0.62 to 0.98). Adverse events, including thromboembolic events, did not differ significantly between groups in WOMAN. Pfizer is among the funders and made the trial drug. Analysed numbers are those printed for this analysis (10 036 and 9985); some randomised women were excluded from analysis and the paper prints no total. Source correction: the Lancet printed a correction to this paper (a Department of Error notice; the online article was corrected on 5 May 2017). It changes one number in the procedures section, the tranexamic acid concentration, from 10 mg/mL to 100 mg/mL. The correction notice lists this one change and no other, and the text read for this claim is the corrected text. The source state shows Corrected for that reason.

• Secondary endpoint WOMAN trial (World Maternal Antifibrinolytic Trial) · See the findings

  Applies to Maternal

  • The subgroup of WOMAN women treated within 3 hours of giving birth. The trial pre-specified subgroup analyses of death due to bleeding by the time interval between giving birth and treatment, as a protocol change made before un-blinding. The same eligibility applies as for the whole trial. The comparison with later treatment is given in the same claim.
  • Hospitals (193 in the paper).
  • Hospitals in 21 countries, as for the whole trial.
  • Postpartum, within 3 hours of birth
  • Gestational age Not reported
  • Postnatal age Not reported

  Follow-up Outcome data were collected at death, discharge or 42 days after randomisation, whichever came first.
  Among women in WOMAN who were given tranexamic acid within 3 hours of giving birth, 89 died of bleeding against 127 given placebo, a risk ratio of 0.69 (95% CI 0.52 to 0.91, p=0.008). When treatment came after 3 hours the counts were 66 and 63, a risk ratio of 1.07 (95% CI 0.76 to 1.51, p=0.70). This is a pre-specified subgroup analysis of the secondary outcome, not the primary result. The trial stated its own rule. Unless the test for different effects across subgroups gave p below 0.001, the overall relative risk (0.81 for death due to bleeding) was regarded as the most reliable guide to subgroup effects. Figure 3 of the paper gives p=0.085 for the test of different effects across the three time-from-delivery groups (up to 1 hour, over 1 to 3 hours, over 3 hours). That is above the trial bar of 0.001, so by the trial rule the overall relative risk of 0.81 stays the better guide. Because earlier trauma data gave strong reason to expect a time effect, the authors also planned to read the timing result alongside all available data on that effect.

  Source problem. A cited source is corrected. Check the sources before relying on this claim.

  Limitations. The p=0.085 in Figure 3 tests three time groups together, not the within-3-hours split alone, and the paper text gives no heterogeneity result for time to treatment on this outcome. Figure 5 of the paper sets WOMAN beside CRASH-2 results and has its own heterogeneity p value; this record did not read it. Figure 3 prints separate denominators for the up to 1 hour and over 1 to 3 hours groups and no combined within-3-hours denominator. WHO 2025 Recommendation 27 remarks rest the 3-hour window on WOMAN timing analysis and on a pooled individual-participant analysis of 40 138 bleeding patients, which includes WOMAN participants and trauma patients. Death due to bleeding is a cause-specific secondary outcome and the primary composite was not clearly reduced. Source correction: the Lancet printed a correction to this paper (a Department of Error notice; the online article was corrected on 5 May 2017). It changes one number in the procedures section, the tranexamic acid concentration, from 10 mg/mL to 100 mg/mL. The correction notice lists this one change and no other, and the text read for this claim is the corrected text. The source state shows Corrected for that reason.

Record details

Record id

maternal:outcome-maternal-death-due-to-bleeding

Content hash

6b1586b2facd1599cb912859424eca9d32a27401b3c622df8b57f6e90777b7fb