Balloon tamponade during cesarean section is useful for severe post-partum hemorrhage due to placenta previa


Dr Kenjiro Sawada, Department of Obstetrics and Gynecology, Osaka University Graduate School of Medicine, 2-2 Yamadaoka, Suita, Osaka, Japan. Email:


Aim:  Severe post-partum hemorrhage during cesarean section due to placenta previa is still one of the leading causes of maternal mortality. The aim of this study was to evaluate the efficiency of intrauterine tamponade with a Sengstaken-Blakemore tube (SB-tube) for the treatment of severe post-partum hemorrhage in cases of placenta previa.

Material and Methods:  Data were collected from our departmental clinical records on all patients who underwent caesarian section due to placenta previa between 2007 and 2009.

Results:  During the period analyzed, 37 patients underwent caesarian section due to placenta previa/low-lying placenta. Four (11%) underwent hysterectomy due to placenta accreta and 33 (89%) were treated conservatively. Of the 33 patients with conserved uterus, 10 (28%) patients required a SB-tube during the cesarean section because of continuous post-partum hemorrhage despite appropriate medical treatment. The median bleeding during the operation was 2030 ± 860 mL in the patients who used SB-tube. None of them presented severe complications related to these procedures or required any further invasive surgery.

Conclusion:  Intrauterine balloon-tamponade could successfully control severe hemorrhage from a lower uterine segment of a patient with placenta previa. This technique is simple to use, scarcely invasive, and available at a low cost to all maternity wards, and should be considered as one of the first management options to reduce the risk of undesirable hysterectomy.


The incidence of placenta previa at the time of birth varies widely in published series, but on average it occurs once in every 150–250 live births. Obstetrical bleeding (intrapartum/post-partum) secondary to placenta previa with variable degrees of accretion is not uncommon. Post-partum bleeding is usually from the placental bed at the lower uterine segment and occurs immediately after the placenta is delivered. Although recent developments in transvaginal ultrasonography allow clinicians to diagnose prenatally, it is still one of the leading causes of maternal mortality.1

Hysterectomy can be an undesirable action to take, especially in the case of a low parity patient. Usually, this step is taken when other traditional measures to stop hemorrhage fail.2 Various management options are utilized for control of bleeding caused by this clinical abnormality and conservative approaches are becoming increasingly used instead of hysterectomy. Arterial embolization under fluoroscopic guidance requires expertise in interventional radiology and specialized equipment, although the success rate is high and the procedure has the potential to preserve fertility. This procedure is limited to centers with a high degree of expertise.3

Nowadays, the use of intrauterine balloons has been well described in the literature for the control of massive post-partum hemorrhage (PPH) due to atonic uterus not responding to oxytocics such as prostaglandins.4 Placement of a uterine balloon tamponade Foley,5 Bakri balloon,6 or Sengstaken-Blakemore tube (SB-tube),7 which may be inserted either after cesarean section or vaginal delivery, is an option with interesting advantages, and is thus often preferred to gauze packing. Placement of a uterine balloon can act as a diagnostic test to screen those women who need hysterectomy. In addition, it minimizes the risk of occult bleeding and removal of the balloon is not a painful procedure. However, the efficacy of the use of intrauterine balloon in PPH after cesarean section complicated by placenta previa remains unclear because only a small series of a few cases have been reported.8,9

With these points in mind, we analyzed the clinical outcomes of uterine tamponade with SB-tube for the treatment of severe PPH due to placenta previa in a larger series. In all 10 cases analyzed, hemostasis was adequately achieved after the insertion of the tube and no patient required any further invasive surgical procedures. Uterine balloon-tamponade was highly effective in controlling PPH originating from the placental site and should be considered as the first step in order to reduce undesirable hysterectomy.


Between January 2007 and December 2009, there were 1731 deliveries in the Perinatal Medical Center of Osaka University Hospital (Osaka, Japan). Of these, 37 (2.1%) underwent caesarian section due to placenta previa/low-lying placenta. All were Asian, aged 27–41 years and from a middle-class community in Osaka or Hyogo prefecture in Japan. Since our unit accepts high-risk pregnant women who can not be treated in usual neighborhood hospitals, the percentage of placenta previa/low-lying placenta was relatively high. Placenta previa was diagnosed by transvaginal ultrasonography in late pregnancy and it was confirmed that the placental edge overlapped the internal cervical os just before the operation. Low-lying placenta was diagnosed when the distance of the placenta edge to the internal os was less than 2 cm. Of 37 patients, four patients (11%) required hysterectomy because the placenta did not separate from the uterus due to placenta accretion. Six patients (16%) had no major troubles because the placenta separated promptly and bleeding was controlled easily. In the remaining 27 cases, after delivery of the placenta, PPH originating from the placental site continued to be resistant to medical therapies, which included intravenous infusions of oxytocin (10–20 U), intravenous ergometrine (0.5 mg), intra muscular prostaglandin F2α or rectal insertion of misoprostol (400 µg). In 10 of those with no response to these medical treatments, the SB-tube was inserted into the uterine cavity during the operation. The insertion of the tube was performed by a single qualified operator (K. S.). In the remaining 17 cases, rolled gauze was used to control severe PPH.

The insertion of SB-tube was based on the method reported by Condous et al.10 Briefly, after cutting the distal end of the tube beyond the balloon, the insertion was facilitated by grasping the anterior and lateral margins of the cervix with sponge forceps and placing the esophageal or stomach balloon into the uterine cavity via the cervix. The balloon was positioned to fit the cervix of the uterus and inflated with 200–300 mL of saline as appropriate for the uterine size. Applying gentle traction confirmed that the SB-tube was firmly fixed in situ in the uterine cavity and good control of the hemorrhage was provided. A representative picture of the procedure is shown in Figure 1. The hysterotomy incision was carefully sutured with VicrylR 1–0 without entrapping the balloon. Broad-spectrum antibiotic therapy was systematically used just before the operation in all cases. In several cases, prophylactic antibiotics were used after the procedure based on the operators' decision. The blood drainage was collected through the distal end of the shaft by attaching a collection bag, and closely observed. SB-tube was removed 24 h after the operation and complete hemostasis was confirmed.

Figure 1.

A representative picture of Sengstaken-Blakemore tube inserted from hysterotomy incision.

Clinical data were collected from medical records. The clinical, biochemical and hematological data were recovered together with data on age, number of pregnancies, parity, gestational age, duration of time between delivery of baby and placenta, estimated blood loss and volume of blood transfused. Postoperative fever was defined as a temperature rise above 38°C maintained over 24 h or recurring during the period from the 1st to the 10th day after childbirth.

Statistical analysis was performed with Stat View (Abacus Conceptus Inc, Berkeley, CA, USA). Statistical differences between groups were analyzed by Kruskal-Wallis test along with post hoc test (Scheffe's test).


Balloon tamponade with SB-tube was used in a total of 10 patients during cesarean section due to uterine hemorrhage resistant to medical therapy. The clinical outcomes of those 10 patients are summarized in Table 1. Complete hemostasis was achieved in all cases. The drainage blood was continuously monitored until the next morning after the operation. The median drainage blood until the next morning was 125.7 ± 81.3 (40–320) mL. The median age of the women was 34.3 ± 2.8 years. The median gestational age was 36.5 weeks of gestation. Six women (60%) were nulliparas and three were para 1. The median birth weight was 2.74 ± 0.21 kg and none of the neonates had apparent congenital abnormalities. The median bleeding during the operation was 2030 ± 860 mL. Two cases (cases 1 and 3) required transfusion due to massive bleeding during the operations. One (case 5) experienced postoperative fever. There were no severe adverse events in any cases and no further invasive surgical procedures were required.

Table 1.  Summary of clinical outcomes of the patients who used Sengstaken-Blakemore tube
Case No.Age (y)Gravidity and parityGestationPosition of placentaDuration between baby delivery and placenta deliveryEstimated blood loss (mL)Drainage blood loss (mL)TransfusionAntibiotics after operationPostoperative fever
  1. RCC, red cell concentrate.

137G4P037w0dPrevia totalis5 min3160120RCC 8UYesNo
230G2P136w1dPrevia totalis1 min100056NoneYesNo
334G5P436w3dPrevia totalis2 min3220121RCC 6UYesNo
434G2P034w6dPrevia totalis4 min130040NoneNoNo
535G2P036w3dPrevia totalis4 min1500170NoneNoYes
629G1P037w3dLow lying placenta8 min170040NoneNoNo
737G2P036w5dPrevia totalis5 min1360320NoneYesNo
837G2P136w6dPrevia totalis4 min1450140NoneNoNo
936G1P037w0dLow lying placenta6 min2830125NoneYesNo
1034G2P136w0dPrevia totalis2 min2780125NoneYesNo

Next, we compared clinical outcomes among the treatment options employed in Table 2. The median blood loss of six patients who required no further treatments was 935 ± 271 mL and significantly lesser than that of rolled gauze (P = 0.043). The median bleeding during the operation was 2030 ± 860 mL in the patients who used SB-tube and 2241 ± 1378 mL in those that used the gauze. No significant differences were noted between these two groups. One case of gauze packing group required uterine artery embolization due to continuous bleeding after the procedure. In the remaining 26 cases in which SB-tube or rolled gauze was used, adequate hemostasis (less than 100 mL/2 h) was achieved soon after the insertion and no further treatments were required. Two cases (20%) of SB-tube group and three cases (18%) of gauze packing required transfusion due to massive bleeding during the operations. Two (20%) of the SB-tube group and eight (47%) of gauze group experienced postoperative fever. Both uterine gauze packing and balloon-tamponade were similarly effective in controlling PPH, and there were no severe adverse events (i.e. severe infectious symptoms) in any cases.

Table 2.  The summary of clinical outcomes of the cases of cesarean section due to placenta previa
 Sengstaken-Blakemore tubeRolled gauzeNo treatmentHysterectomy required
  1. N.A., not applicable.

Age (y)34.3 ± 2.833.7 ± 5.235.0 ± 2.435.3 ± 4.0
Previous cesarean section0/102/171/64/4
Previous myomectomy1/101/170/60/4
Estimated blood loss (mL)2030 ± 8602241 ± 1378935 ± 2713300 ± 1764
Uterine artery embolization required0/101/170/6N.A.
Postoperative fever2/108/171/62/4

In Table 3, the clinical outcomes of the cases required hysterectomy are summarized. In case 1, 3 and 4, placenta accrete was strongly suspected prenatally by MRI and Doppler ultrasonography. Since a 25% to 50% incidence of placenta accreta in patients with placenta previa with prior cesarean delivery has been well recognized, the patient and her family agreed with hysterectomy when placenta was not separated spontaneously. In case 4, to avoid the risk of hemorrhage, we scheduled and performed stepwise treatment suggested by Sumigama et al. in Nagoya University.11 The brief procedure was as follows: a cesarean section was performed without separation of the placenta; on the operation day, transcatheter angiographic uterine arterial embolization was conducted with gelatin sponge particles and platinum coils; one week later, total hysterectomy was carried out. In Table 3, the blood loss during the initial cesarean section was shown. In case 2, although the placenta was partially separated spontaneously, the other part of placenta was tightly attached to the uterine wall and massive bleeding occurred. Since the operator considered it was not possible to control bleeding, the hysterectomy was emergently performed.

Table 3.  Summary of clinical outcomes of the patients who required hysterectomy
CaseAge (y)Gravity and parityPrevious CSGestationPrenatal diagnosisEstimated blood loss (mL)Transfusion requiredPathological diagnosis
  • Estimated blood loss during the initial cesarean section. CS, cesarean section; FFP, fresh frozen plasma; RCC, red cell concentrate.

131G2P1135w2dYes2900RCC 4UPlacenta percreta
237G4P3235w4dNo5000RCC 8U FFP 8UPlacenta accreta
340G4P2236w0dYes4300RCC 8UPlacenta accreta
433G3P1135w3dYes1000NonePlacenta increta


Post-partum hemorrhage in cases of placenta previa remains a serious obstetric complication. Successful control of bleeding can often be achieved medically using uterotonics including oxytocin, ergometrine, 15-methyl prostaglandin F2α and misoprostol.9 However, once these medical treatments fail, it is often necessary to intervene surgically with uterine or internal iliac ligation, uterine compression sutures or hysterectomy. When dealing with young women who may wish to have more children, the cesarean-hysterectomy without delay, resulting in devastating emotional and/or cultural consequences, should be avoided wherever possible.12 Actually, in our reports, six women were nulliparas and all 10 cases who used SB-tube strongly wished to preserve fertility.

While a variety of surgical techniques have been proposed to avoid hysterectomy, a suitable conservative technique is still lacking in the literature and all the advantages shown by the proposed options are counterbalanced by some risks.2 Herein, we described that not only uterine gauze packing but uterine balloon tamponade with a SB-tube showed similarly high efficacy in controlling PPH originating from the placental site of the lower uterine segment during caesarian sections. Historically, the use of uterine gauze packing in the management of PPH fell into disfavor after the 1960s, following concerns of concealing ongoing hemorrhage, development of infection and its ‘non-physiological approach’.13 Condous et al. commented, based on their experiences, that uterine packing with gauze packs is outdated and should only be reserved in cases when a balloon catheter is unavailable.10 Control of PPH by uterine balloon packing itself is not a new idea. Although some of the balloons, such as the Rusch balloon and the condom catheter, are reported to be effective,14 they do not allow blood drainage from the uterine cavity. On the contrary, in SB-tube system, the blood drainage is collected through the distal end of the shaft by attaching a collection bag. This drainage system helps prevent blood collection inside the uterine cavity and provide an accurate estimation of bleeding. Besides, if the procedure fails to stop the bleeding, the failure is immediately visible, as opposed to the use of the conventional balloons as tamponade. With these reasons, we consider SB-tube system to be superior to the other treatment options such as rolled gauze or the conventional balloon.

Reports are accumulating suggesting that methods of uterine tamponade are effective to avoid hysterectomy in frequently unstable patients and can preserve fertility, especially in the case of severe PPH secondary to uterine atony.15 A recent American College of Obstetricians and Gynecologists practice bulletin suggests that tamponade of the uterus can be an effective way to decrease hemorrhage secondary to uterine atony, and procedures such as uterine artery ligation or B-Lynch suture may be used to obviate the need for hysterectomy. Furthermore, it is suggested that if hysterectomy is performed for uterine atony, there should be documentation of these therapies' attempts.16 On the other hand, the use of a balloon tamponade in severe PPH due to placenta previa has been reported only in a small series of a few cases. Bowen et al. reported the first attempt to archive hemostasis in case of PPH complicated by placenta previa by compression using a Foley catheter.17 Bakri et al. used a self-made original tamponade balloon in two cases of placenta previa together with additional surgical procedures such as bilateral hypogastric ligation.6 Recently, Condous et al. reported the use of SB-tube in the management of PPH.10 In this series of 17 cases, two patients experienced severe PPH complicated by placenta previa but were successfully treated without any further invasive procedures. Those reports offer evidence that further clinical attention should be given to this procedure, and herein we reported a larger series of 10 cases and demonstrated that intrauterine balloon tamponade should be considered as a management option before performing surgical procedures in PPH resulting from placental site bleeding.

In this report, uterine tamponade with SB-tube controlled PPH during cesarean section for placenta previa in all 10 patients, although the success rate of uterine balloon tamponade as a single measure for the management of major PPH has been reported as 77.5–88.8% in a recent systematic review.13 One possible reason for this discrepancy is that the use of a balloon tamponade has been reported in the management of severe PPH mainly resulting from uterine atony. In our experience, a balloon tamponade is more effective in cases of placenta previa than those of uterine atony. The intrauterine balloon is considered to act by exerting in ‘inward-to-outward pressure’ that is greater than the systemic arterial pressure to prevent continual bleeding.18 Since the uterine cavity itself is well contracted in cases of placenta previa, adequate ‘inward-to-outward pressure’ produced by the tube is likely to be achieved easily. In addition, by halting ongoing hemorrhage from the placental bed promptly during the operation, consumptive coagulopathy can be reversed in most cases. Further data from a larger prospective study are needed to verify our hypothesis.

One possible problem with using the SB-tube for placenta previa is that it takes some time to insert the SB-tube because it must be inserted into the uterine cavity transvaginally during the operation. The SB-tube has two separated drainage catheters with attached parts, which might cause further tearing of the lower uterine segment if inserted through the hysterotomy incision because the uterine cervical canal of the patients is usually almost closed in cases of placenta previa. Since PPH occurs immediately after the placenta is delivered, even a few minutes of lost time might cause additional bleeding. In that sense, the more simple and flexible tamponade balloon, which is specially designed for a uterus and can be inserted through the hysterotomy incision, might decrease PPH during the operation. Indeed, the SOS Bakri Tamponade Balloon Catheter is designed just for the uterus and is reported easily administered, although this type of catheter is not commercially available in Japan. Such an improved device should be examined to achieve better outcomes in cases of severe PPH.

In conclusion, the uterine balloon-tamponade was effective in controlling PPH originating from the placental site, although our results are from a retrospective study of a small series. This method is very effective, simple to use, scarcely invasive, and available at a low cost for all maternity wards. It should be considered as the first step in order to reduce undesirable hysterectomy.


This work was supported in part by a Grant-in-Aid for scientific research from the Ministry of Education, Science, Sports and Culture of Japan. The authors are grateful to Remina Emoto and Ayako Okamura for their secretarial assistance.