Efficacy of early intrauterine balloon tamponade for immediate postpartum hemorrhage after vaginal delivery: a randomized clinical trial

Background Many questions remain about the appropriate use of intrauterine balloon devices in postpartum hemorrhage after vaginal delivery refractory to first-line uterotonics. Available data suggest that early use of intrauterine balloon tamponade might be beneficial. Objective This study aimed to compare the effect of intrauterine balloon tamponade used in combination with second-line uterotonics vs intrauterine balloon tamponade used after the failure of second-line uterotonic treatment on the rate of severe postpartum hemorrhage in women with postpartum hemorrhage after vaginal delivery refractory to first-line uterotonics. Study Design This multicenter, randomized, controlled, parallel-group, nonblinded trial was conducted at 18 hospitals and enrolled 403 women who had just given birth vaginally at 35 to 42 weeks of gestation. The inclusion criteria were a postpartum hemorrhage refractory to first-line uterotonics (oxytocin) and requiring a second-line uterotonic treatment with sulprostone (E1 prostaglandin). In the study group, the sulprostone infusion was combined with intrauterine tamponade by an ebb balloon performed within 15 minutes of randomization. In the control group, the sulprostone infusion was started alone within 15 minutes of randomization, and if bleeding persisted 30 minutes after the start of sulprostone infusion, intrauterine tamponade using the ebb balloon was performed. In both groups, if the bleeding persisted 30 minutes after the insertion of the balloon, an emergency radiological or surgical invasive procedure was performed. The primary outcome was the proportion of women who either received ≥3 units of packed red blood cells or had a calculated peripartum blood loss of >1000 mL. The prespecified secondary outcomes were the proportions of women who had a calculated blood loss of ≥1500 mL, any transfusion, an invasive procedure and women who were transferred to the intensive care unit. The analysis of the primary outcome with the triangular test was performed sequentially throughout the trial period. Results At the eighth interim analysis, the independent data monitoring committee concluded that the incidence of the primary outcome did not differ between the 2 groups and stopped inclusions. After 11 women were excluded because they met an exclusion criterion or withdrew their consent, 199 and 193 women remained in the study and control groups, respectively, for the intention-to-treat analysis. The women's baseline characteristics were similar in both groups. Peripartum hematocrit level change, which was needed for the calculation of the primary outcome, was missing for 4 women in the study group and 2 women in the control group. The primary outcome occurred in 131 of 195 women (67.2%) in the study group and 142 of 191 women (74.3%) in the control group (risk ratio, 0.90; 95% confidence interval, 0.79–1.03). The groups did not differ substantially for rates of calculated peripartum blood loss pf ≥1500 mL, any transfusion, invasive procedure, and admission to an intensive care unit. Endometritis occurred in 5 women (2.7%) in the study group and none in the control group (P=.06). Conclusion The early use of intrauterine balloon tamponade did not reduce the incidence of severe postpartum hemorrhage compared with its use after the failure of second-line uterotonic treatment and before recourse to invasive procedures. Many questions remain about the appropriate use of intrauterine balloon devices in postpartum hemorrhage after vaginal delivery refractory to first-line uterotonics. Available data suggest that early use of intrauterine balloon tamponade might be beneficial. This study aimed to compare the effect of intrauterine balloon tamponade used in combination with second-line uterotonics vs intrauterine balloon tamponade used after the failure of second-line uterotonic treatment on the rate of severe postpartum hemorrhage in women with postpartum hemorrhage after vaginal delivery refractory to first-line uterotonics. This multicenter, randomized, controlled, parallel-group, nonblinded trial was conducted at 18 hospitals and enrolled 403 women who had just given birth vaginally at 35 to 42 weeks of gestation. The inclusion criteria were a postpartum hemorrhage refractory to first-line uterotonics (oxytocin) and requiring a second-line uterotonic treatment with sulprostone (E1 prostaglandin). In the study group, the sulprostone infusion was combined with intrauterine tamponade by an ebb balloon performed within 15 minutes of randomization. In the control group, the sulprostone infusion was started alone within 15 minutes of randomization, and if bleeding persisted 30 minutes after the start of sulprostone infusion, intrauterine tamponade using the ebb balloon was performed. In both groups, if the bleeding persisted 30 minutes after the insertion of the balloon, an emergency radiological or surgical invasive procedure was performed. The primary outcome was the proportion of women who either received ≥3 units of packed red blood cells or had a calculated peripartum blood loss of >1000 mL. The prespecified secondary outcomes were the proportions of women who had a calculated blood loss of ≥1500 mL, any transfusion, an invasive procedure and women who were transferred to the intensive care unit. The analysis of the primary outcome with the triangular test was performed sequentially throughout the trial period. At the eighth interim analysis, the independent data monitoring committee concluded that the incidence of the primary outcome did not differ between the 2 groups and stopped inclusions. After 11 women were excluded because they met an exclusion criterion or withdrew their consent, 199 and 193 women remained in the study and control groups, respectively, for the intention-to-treat analysis. The women's baseline characteristics were similar in both groups. Peripartum hematocrit level change, which was needed for the calculation of the primary outcome, was missing for 4 women in the study group and 2 women in the control group. The primary outcome occurred in 131 of 195 women (67.2%) in the study group and 142 of 191 women (74.3%) in the control group (risk ratio, 0.90; 95% confidence interval, 0.79–1.03). The groups did not differ substantially for rates of calculated peripartum blood loss pf ≥1500 mL, any transfusion, invasive procedure, and admission to an intensive care unit. Endometritis occurred in 5 women (2.7%) in the study group and none in the control group (P=.06). The early use of intrauterine balloon tamponade did not reduce the incidence of severe postpartum hemorrhage compared with its use after the failure of second-line uterotonic treatment and before recourse to invasive procedures.