Objective
The purpose of this study was to investigate changes in invasive procedure rates after the addition of intrauterine balloon tamponade as an initial second-line therapy to our protocol for the management of severe postpartum hemorrhage.
Study Design
We compared the outcomes of all patients with postpartum hemorrhage that was unresponsive to prostaglandin during 2 equal periods, before (first period) and after (second period) the introduction of a balloon tamponade protocol.
Results
During the second period, 43 women had uterine tamponade with a Bakri balloon as their initial second-line therapy (after vaginal delivery, 31; after caesarean delivery, 12). The global success rate was 86% (37/43 women). Among patients who delivered vaginally, the rates of arterial embolization (8.2% vs 2.3% in the first and second period; P = .006; odds ratio, 0.26; 95% confidence interval, 0.09–0.72) and conservative surgical procedures (5.1% vs 1.4%, in the first and second period; P = .029; odds ratio, 0.26; 95% confidence interval, 0.07–0.95) were significantly lower after the uterine tamponade test was added to our protocol.
Conclusion
Intrauterine balloon tamponade is an attractive adjunctive strategy for the achievement of hemostasis in intractable hemorrhages and prevention of the need for invasive procedures.
Approximately 1-2% of deliveries in developed countries are associated with severe postpartum hemorrhage (PPH), which is among the most common causes of pregnancy-related death in the United States and Europe. In France, PPH remains the main cause of maternal death, responsible for 22% of maternal deaths from direct obstetric causes. Decreasing the prevalence of severe PPH remains a challenge and an important one, given the increase in the incidence of PPH recently reported in several developed countries. In the United States, these apparent increases in PPH rates are largely due to an increased prevalence of both uterine atony and placenta accreta, the latter associated with the rising cesarean delivery rate. The basic treatment of major PPH consists of medical management and uterotonic drugs such as oxytocin, ergometrine, and prostaglandin or their analogues. When these first-line treatments fail, second-line therapies (which include uterine compression sutures), pelvic vessel ligation (which includes ligation of the internal iliac or uterine arteries), or arterial embolization can be used individually or in combination. Hysterectomy is the ultimate measure for the control of bleeding and saving the mother’s life.
The use of an intrauterine balloon to facilitate tamponade is becoming a widely suggested technique and often is adopted as a second-line procedure in the management of PPH. The sparse literature about this noninvasive procedure is thus far limited to case reports and a few retrospective and prospective series that include small numbers of patients. They suggest, however, that it is an effective tool in the management of PPH. The potential benefits of intrauterine balloon tamponade for reducing the rate of invasive procedures after the failure of medical management have not yet been studied extensively. The objective of this study was therefore to investigate the trends in the invasive procedure rate in our department after the addition of the Bakri balloon (Cook Medical, Bloomington, IN) as the initial second-line therapy in our protocol for management of severe PPH, compared with an historic control.
Materials and Methods
We conducted a before-and-after study in the maternity unit of Poissy Saint-Germain Hospital, a tertiary referral university hospital, to evaluate the efficacy and tolerance of intrauterine balloon tamponade as an initial second-line treatment for severe PPH.
Before April 2008, our protocol for PPH included the following steps that are listed in the national guidelines : (1) active management of the third stage of labor that includes prophylactic injection of 5 IU oxytocin on delivery of the anterior shoulder in vaginal births or immediately after birth in cesarean deliveries, early clamping of the umbilical cord, and controlled cord traction; (2) manual removal of the placenta, which is performed without delay in case of hemorrhage or after 20 minutes in the absence of bleeding in vaginal births; (3) if bleeding appears excessive, administration of additional oxytocin (10 IU) after manual removal of any retained placental tissue, uterine massage until adequate contraction occurs, emptying of the bladder, and search for and treatment of any obvious bleeding in any episiotomy or tear; (4) if bleeding continues, circulatory support as necessary with crystalloids, colloids, and/or blood products; intravenous infusion of sulprostone (PGE2 analogue), starting at a dose of 500 μg/hour for 1 hour and then continuing as necessary at a dose of 100 μg/hour for 5 hours; bimanual compression is also performed; (5) if prostaglandin infusion does not control the bleeding, organ system support that includes oxygenation; if necessary, vasopressors are provided, and an invasive procedure is performed; at the discretion of the attending physician, this included a uterine artery embolization or surgical artery ligations, compressive uterine sutures, or hysterectomy.
In April 2008, the protocol was modified to require that an intrauterine balloon tamponade test be used before any invasive procedure in immediate hemorrhages because of uterine atony unresponsive to sulprostone after vaginal delivery or delayed PPH because of uterine atony unresponsive to sulprostone after cesarean delivery. Insertion of the balloon tamponade therapy followed the manufacturer’s recommended instructions for use. Depending on the situation, insertion of the Bakri balloon into the uterine cavity was transvaginal and digital or with a vaginal speculum, with the proximal end of the balloon held with sponge forceps. Once the balloon was placed in the cavity, it was inflated with 400-500 mL of sterile water. Its intrauterine position was monitored by abdominal ultrasound scanning. At this stage, if there was no bleeding through either the cervix or the balloon’s drainage channel, the tamponade test was considered successful. On the other hand, if bleeding continued, the tamponade test was considered a failure, and an emergency radiologic or surgical invasive procedure was performed immediately after the withdrawal of the balloon in the operating room. When the tamponade test was successful, the patient was transferred to the intensive care unit. A low-dose infusion of sulprostone (100 μg/hour) was continued up to a total dose of 500 μg. A low-dose infusion of oxytocin (40 IU in a liter of normal saline solution) then followed and continued to keep the uterus contracted until the balloon was withdrawn 24 hours after placement; it was deflated to one-half its volume, however, after 12 hours. All patients had a Foley catheter inserted into the bladder and received antibiotic prophylaxis (amoxicillin-clavulanate and gentamycin) for 48 hours.
In cases of persistent bleeding during cesarean delivery, surgical artery ligations and/or uterine compression sutures were performed initially. When these procedures failed, a hysterectomy was performed. Intrauterine balloon tamponade was never attempted in patients with placenta accreta, lacerations, or retained placenta.
For this before-and-after study, we compared the outcomes of all patients with PPH because of uterine atony that is unresponsive to sulprostone during both equal periods. From April 2008 to December 2010 (after balloon tamponade introduction), prospective data collection began in a group of patients that constituted the study group. From July 2005 to March 2008 (before balloon tamponade therapy was available; ie, the immediately previous period), the group of patients that constituted the control group were identified in our electronic database, and their files were reviewed to collect the data. The data that were collected included maternal age, body mass index, parity, singleton or multiple pregnancy, gestational age at delivery, mode of delivery (vaginal or cesarean), birthweight, and PPH management.
The primary outcome measures were the incidence of arterial embolization, conservative surgical procedures (artery ligations and/or uterine compression sutures), and hysterectomy among patients who delivered vaginally. Secondary outcomes were the incidence of maternal death, blood transfusion, and decreased peripartum hemoglobin of ≥2 g/dL (considered equivalent to the loss of >500 mL of blood) among patients who delivered vaginally. Other secondary outcomes were the incidence of arterial embolization, conservative surgical procedures, hysterectomy, maternal death, blood transfusion, and decreased peripartum hemoglobin of ≥2 g/dL among patients who delivered by cesarean section.
Prepartum hemoglobin was collected as part of routine prenatal care during the last weeks of pregnancy; postpartum hemoglobin level was the lowest level found in the 2 days after delivery.
Under French law, this study was exempt from institutional review board review because its methods (before-and-after study of a change in standard of care) is not considered experimental and is thus not covered by the Huriet Law on biomedical research.
The comparability of the study groups was determined by assessment of their baseline characteristics, which also made it possible to identify any potential confounding factors. Because continuous outcomes were not normally distributed, these data are presented as median values. Comparative data analysis used the Fisher exact test, a Mann-Whitney test or a χ 2 test, as appropriate. A 2-sided probability value of < .05 was considered statistically significant; the odds ratio (OR) with its 95% confidence interval (CI) was calculated for each variable. Statistical analysis was performed with Stata10 software (release 10; StataCorp, College Station, TX).
Results
During the control period, 11,781 women gave birth in our maternity ward, 25.6% (n = 3013) of them were delivered by cesarean section; during the study period, the corresponding figures were 12,082 and 27.2% for cesarean delivery (n = 3285; P = .005). The PPH rates were 6.9% for the control period (n = 820 births) and 5.4% for the study period (n = 663 births; P < .0001); the rates of PPH that required intravenous sulprostone were 2.4% (n = 290 births) and 3.2% (n = 395 births), respectively ( P = .0002). We note, however, that 5 patients in the first period and 6 patients in the second period had immediate hysterectomies without prostaglandin treatment because of placenta previa, placenta accreta, or uterine rupture during a vaginal birth after cesarean section ( Figure ).
Table 1 summarizes the general and obstetric characteristics of the women whose condition required intravenous sulprostone during these 2 periods. The groups did not differ significantly in maternal age, body mass index, percentage of nulliparous women or multiple pregnancies, gestational age at delivery, or birthweight. However, the proportion of cesarean deliveries was significantly higher among the patients who received a sulprostone infusion during the second period.
| Variable | Period 1: before balloon tamponade (n = 290) | Period 2: after balloon tamponade (n = 395) | P value |
|---|---|---|---|
| Maternal age, y a | 31 (26–34) | 30 (27–34) | .49 |
| Gestational age, wk a | 39 (38–40) | 39 (38–40) | .84 |
| Nulliparous, n (%) | 160 (55.2) | 212 (53.7) | .70 |
| Body mass index kg/m 2 a | 22.2 (20.3–25.4) | 21.9 (20.0–24.9) | .20 |
| Multiple pregnancy, n (%) | 35 (12.1%) | 46 (11.7%) | .87 |
| Birthweight, g a | 3360 (2930–3720) | 3396 (2900–3740) | .61 |
| Mode of delivery, n (%) | .005 | ||
| Vaginal delivery | 194 (66.9) | 218 (55.2) | |
| Elective cesarean delivery | 26 (9.0) | 58 (14.7) | |
| Cesarean delivery during labor | 70 (24.1) | 119 (30.1) |
The treatment of the women with PPH that was unresponsive to sulprostone during the 2 periods is reported in the Figure .
The general and obstetric characteristics of the women who delivered vaginally with PPH that was unresponsive to sulprostone during the 2 periods were similar ( Table 2 ).
| Variable | Period 1: before balloon tamponade (n = 38) | Period 2: after balloon tamponade (n = 72) | P value |
|---|---|---|---|
| Vaginal delivery a | |||
| Maternal age, y b | 30 (26–33) | 31 (27–36) | .93 |
| Gestational age, wk b | 40 (38–41) | 39 (38–41) | .73 |
| Nulliparous, n (%) | 12 (46.2) | 15 (42.9) | .80 |
| Body mass index, kg/m 2 b | 20.5 (19.2–23.4) | 21.9 (19.9–23.7) | .43 |
| Multiple pregnancy, n (%) | 4 (15.4) | 5 (14.3) | .91 |
| Birthweight, g b | 3335 (2950–3720) | 3490 (3080–3790) | .73 |
| Cesarean delivery c | |||
| Maternal age, y b | 32 (29–35) | 31 (27–35) | .53 |
| Gestational age, wk b | 39 (35–41) | 38 (36–40) | .82 |
| Nulliparous, n (%) | 4 (33.3) | 13 (35.1) | .91 |
| Body mass index, kg/m 2 b | 27.7 (26.2–28.6) | 23.0 (21.5–26.4) | .009 |
| Multiple pregnancy, n (%) | 1 (8.3) | 12 (32.4) | .10 |
| Birthweight, g b | 3850 (2315–4145) | 2990 (2330–3770) | .18 |
a Period 1, 26 women; period 2, 35 women;
b Data are given as median (quartile 1-quartile 3);
During the control period, 13.4% of patients (26/194) who delivered vaginally and whose condition required intravenous sulprostone had an invasive procedure. Two of the 10 conservative surgical procedures failed, which led to 2 hysterectomies. None of the 16 embolizations failed.
During the study period, 16.0% of patients (35/218) who delivered vaginally and whose condition required intravenous sulprostone had a second-line therapy, with a rate not significantly higher than during the control period ( P = .45): (1) Thirty-one women (14.2%) had a uterine tamponade test as the first second-line therapy. The tamponade test failed in 5 patients and led to 3 embolizations, 1 conservative surgical procedure, and 1 hysterectomy. The success rate was 83.8% (26/31 women). (2) Four other patients (1.8%) had invasive procedures as the first second-line therapy: 2 conservative surgical procedures and 2 embolizations. These 4 procedures were all performed during the first 8 months after we began using the balloon tamponade for severe PPH management.
The second-line therapy rates among the women with cesarean deliveries and PPH that was unresponsive to sulprostone were 12.5% (12/96 women) and 20.9% (37/177 women; P = .08), during the control and the study periods, respectively ( Figure ). The general and obstetric characteristics of these women during the 2 periods are reported in Table 2 .
During the study period, 12 women had a uterine tamponade test as the first second-line therapy after cesarean deliveries. The tamponade test failed in 1 patient who underwent a conservative surgical procedure. The success rate was thus 91.6% (11/12 women).
Finally, the global success rate of the uterine tamponade test that included the 31 patients who delivered vaginally and the 12 patients who delivered by cesarean section was 86% (37/43 women). Table 3 presents their general and obstetric characteristics.
| Variable | Measure |
|---|---|
| Maternal age, y a | 31 (27–34) |
| Gestational age, wk a | 39 (38–41) |
| Nulliparous, n (%) | 15 (34.9) |
| Body mass index, kg/m 2 a | 22.7 (20.7–25.7) |
| Multiple pregnancy, n (%) | 7 (16.3%) |
| Birthweight, g a | 3490 (2940–3810) |
| Mode of delivery, n (%) | |
| Vaginal birth | 31 (72.1) |
| Elective cesarean delivery | 7 (16.3) |
| Cesarean delivery during labor | 5 (11.6) |
Among women with vaginal deliveries, the rates of arterial embolizations (8.2% vs 2.3% during the control and study periods, respectively; P = .006; OR, 0.26; 95% CI, 0.09–0.72) and conservative surgical procedures (5.1% vs 1.4%, respectively; P = .029; OR, 0.26; 95% CI, 0.07–0.95) fell significantly after we added the uterine tamponade test to our protocol ( Table 4 ) . Accordingly, the global rate of patients whose condition requires an invasive procedure also dropped significantly (13.4% in the control period and 4.1% in the study period; P = .001; OR, 0.28; 95% CI, 0.13–0.61). Finally, significantly fewer women had a peripartum hemoglobin loss of ≥2 g/dL during the study period (53.7% vs 66.5% in the control period; P = .008; OR, 0.58; 95% CI, 0.39–0.87). On the other hand, neither transfusion (8.3% in the control period and 10.6% in the study period) nor hysterectomy rates (1% and 0.46%, respectively) changed significantly between the 2 periods.
| Variable | Period 1: before balloon tamponade (n = 194) | Period 2: after balloon tamponade (n = 218) | P value | Odds ratio (95% CI) |
|---|---|---|---|---|
| Vaginal deliveries, n (%) a | ||||
| Embolization | 16 (8.2) | 5 (2.3) | .006 | 0.26 (0.09–0.72) |
| Conservative surgical procedures b | 10 (5.1) | 3 (1.4) | .03 | 0.26 (0.07–0.95) |
| Hysterectomy | 2 (1.0) | 1 (0.46) | .50 | 0.44 (0.04–4.91) |
| Patients with invasive procedures | 26 (13.4) | 9 (4.1) | .001 | 0.28 (0.13–0.61) |
| Transfusion | 16 (8.3) | 23 (10.6) | .43 | 1.31 (0.67–2.56) |
| Peripartum hemoglobin loss ≥2 g/d | 129 (66.5) | 117 (53.7) | .008 | 0.58 (0.39–0.87) |
| Cesarean deliveries c | ||||
| Embolization | 0 | 2 (1.1) | .30 | — |
| Conservative surgical procedures b | 12 (12.5) | 23 (13.0) | .91 | 1.04 (0.50–2.20) |
| Hysterectomy | 1 (1.0) | 3 (1.7) | .67 | 1.64 (0.17–16.0) |
| Patients with invasive procedures | 12 (12.5) | 26 (14.7) | .62 | 1.20 (0.58–2.51) |
| Transfusion | 9 (9.4) | 20 (11.3) | .62 | 1.23 (0.54–2.82) |
| Peripartum hemoglobin loss ≥2 g/dL | 54 (56.3) | 77 (43.5) | .04 | 0.60 (0.36–0.99) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
