Problem: dangerous complication

Massive pelvic floor hemorrhage is a potentially life-threatening condition associated with complicated obstetric and gynecologic procedures. Profuse bleeding from lateral pelvic floor veins or presacral veins can be encountered in various situations, such as during peripartum hysterectomy, placenta percreta, abdominal pregnancy, radical hysterectomy, and pelvic exenteration. When adequate control of bleeding is delayed, a secondary coagulopathy can complicate the situation by impairing hemostasis, consequently contributing to more blood loss. Hemorrhages exacerbated by coagulopathy commonly arise from venous plexuses or large raw surfaces at the pelvic floor that cannot be managed by clipping, ligating, suturing, or applying hemostatic agents.

Conventionally, this type of bleeding has been successfully controlled with abdominopelvic packing using Kerlex rolls (Kendall Co., Boston, MA) tied together in a bowel bag, the so-called pack and go back method, or with dry laparotomy pads, applied firmly to the bleeding sites. These methods temporarily contain bleeding until adequate volume replacement and correction of the coagulation defect can be achieved. However, a major disadvantage of both is the need to repeat laparotomy for pack removal.

First detailed by Logothetopoulos in 1926, a transvaginal pack, known variously as the Logothetopoulos, umbrella, parachute, or mushroom pack, relies on pressure to control copious posthysterectomy bleeding from pelvic floor venous plexuses and raw surfaces. A sterile plastic bag filled with continuously tied gauze rolls is used to compress the bleeding sites against the underlying fascia, muscle, and bone. The pack is inserted into the pelvis through the abdomen so that the opening of the bag and the tip of the assembled gauze line can be extended through the vaginal introitus. To ensure the pack furnishes the needed pressure, the open end of the bag, together with the gauze tail, is tied to a 1 L intravenous fluid bag that hangs at the foot of the bed for traction. A collection of case reports indicates this technique is successful in 85% of obstetric cases and 100% of gynecological cases. A major advantage is the ability to avoid a second laparotomy because the pack is removed vaginally.

Over the last decade, intrauterine placement of the Bakri postpartum balloon (Cook Medical Inc, Bloomington, IN) has been well accepted as a simple and effective measure for controlling intractable placental site bleeding after third-trimester delivery, second-trimester miscarriage, and cervical ectopic pregnancy. A newly devised application allows the Bakri balloon to be used in place of umbrella packing for pelvic floor hemorrhage after cesarean delivery. This technique provides several benefits.

Our solution

A 43 year old woman, gravida 2, para 1, was at 36 weeks’ gestation with a diagnosis of severe preeclampsia and fetal growth restriction. She underwent induction of labor using intravaginal misoprostol. In the early latent phase of labor, emergency cesarean section was performed because of a worrisome fetal heart rate pattern. The baby weighed 4 lb 2.6 oz (1890 g) and had a 1 minute Apgar score of 5. After that, the placenta was removed uneventfully, and the uterus was exteriorized for better exposure of the lower segment.

Following repair of the uterine lower segment and bilateral tubal ligation, profuse bleeding from raw surfaces between the posterior uterine wall and the posterior cul-de-sac peritoneum overlying the lower sacrum was evident. It was thought to occur during exteriorization of the uterus when a dense adhesion obliterating the cul-de-sac was accidentally broken. Bleeding was aggravated by the increased vascularity of the gravid uterus and adjacent pelvic tissues.

An initial attempt to curb the bleeding by direct suture ligation and electrocoagulation of the affected areas was unsuccessful. A gynecological oncologist was summoned to the operating room for a hemostasis consult. Bilateral low and high uterine artery ligation was performed with satisfactory reduction of bleeding at the posterior uterine wall. However, the pelvic floor hemorrhage continued, even after more suturing and coagulation of the area. The bleeding appeared venous in origin and came from the large raw surface and friable tissues covering the pelvic floor ( Figure 1 ).

Profuse bleeding came from the large raw surface and friable tissues overlying the pelvic floor.

Charoenkwan. Pelvic packing balloon. Am J Obstet Gynecol 2013 .

Manual direct compression with laparotomy pads was applied to the area for approximately 10 minutes to try to stem the bleeding. Yet as soon as the pressure was released, active bleeding resumed. From the surgeon’s experience, it was apparent that any further surgical tactics to control bleeding would only lead to more tissue trauma, more blood loss, and prolonged surgery. Pelvic packing, which affords a longer period of compression, was determined to be a better choice because it would supply time for adequate volume replacement and correction of any coagulation defect were one to arise. It would also reduce the chance of having to perform another laparotomy to pack still-bleeding areas if coagulopathy did develop perioperatively.

Although an umbrella pack was a valid option, it was deemed too cumbersome because space in the pelvis was limited because of the presence of the uterus. Therefore, it was determined that the use of the Bakri balloon would be equally effective and more convenient.

The patient was placed in the dorsal lithotomy position. A sponge was situated in the posterior fornix with holding forceps and pushed upward to facilitate posterior culdotomy. Once all arterial bleeding had been contained, a posterior culdotomy incision, approximately 2.5 cm in size, was made. This was just large enough to accommodate the deflated balloon. The stopcock at the distal part of the inflation port was detached from the unit to facilitate placement. Then the Bakri balloon was introduced abdominally via the laparotomy incision and passed, inflation port first, through the posterior cul-de-sac opening to the vagina ( Figure 2 ). Next, an assistant pulled the balloon’s shaft through the vaginal canal. SURGICEL absorbable hemostats (Ethicon, Inc, Somerville, NJ) were set on the surgical base for additional hemostasis.

The Bakri balloon was inserted abdominally via the laparotomy incision.

Charoenkwan. Pelvic packing balloon. Am J Obstet Gynecol 2013 .

After the best tamponade position was determined, the stopcock was reattached to the end of the inflation port, and the balloon was gradually inflated with sterile normal saline ( Figure 3 ). The goal was to fill the balloon to the minimal volume that could effectively compress the pelvic floor and completely control the bleeding. For this patient, given the anatomic location of the bleeding, that volume was judged to be 200 mL. The balloon was then firmly placed against the bleeding area in the pelvis ( Figure 4 ). Continuous traction was applied by attaching the balloon shaft to a 1 L intravenous fluid bag that was suspended at the end of the bed ( Figure 5 ). The balloon’s drainage port was connected to the fluid collection bag. A Video Clip demonstration of the procedure can be found as supplementary material to this article.

Sterile normal saline was used to gradually inflate the balloon.

Charoenkwan. Pelvic packing balloon. Am J Obstet Gynecol 2013 .

Inflated with 200 mL of sterile normal saline, the balloon was firmly placed against the bleeding area in the pelvis.

Charoenkwan. Pelvic packing balloon. Am J Obstet Gynecol 2013 .

A 1 L intravenous fluid bag suspended at the end of the bed provided traction.

Charoenkwan. Pelvic packing balloon. Am J Obstet Gynecol 2013 .

Following adequate hemostasis, all packs and instruments were removed, and the abdomen was closed. The patient was then transferred to the intensive care unit for close monitoring of her hemodynamic status and replacement of fluid and blood components. Her estimated blood loss was 1000 mL, and the entire operation lasted 200 minutes. She remained stable in the postoperative period. The fluid collection bag captured a total of 50 mL of drained fluid, and vaginal bleeding was minimal. Approximately 12 hours after placement, the traction on the balloon shaft was removed, and the Bakri balloon was deflated.

The patient’s vital signs, abdominal signs, and vaginal bleeding were closely observed for another 2 hours to guarantee hemostasis had been achieved before the balloon was removed. Had bleeding persisted, the balloon could be reinflated, but this proved unnecessary. Therefore, the balloon was removed by pulling it through the vaginal canal. This was accomplished at the bedside without sedation or anesthesia, and the patient remained reasonably comfortable throughout the procedure. She was discharged from the hospital on postoperative day 4 in stable condition. At 6 weeks’ postpartum, she had fully recovered without complication.

A PubMed search did not uncover any previous report demonstrating successful use of the Bakri balloon as a pressure pack for the control of relentless pelvic floor hemorrhage in a circumstance like this. As with umbrella packing, the Bakri balloon can be removed vaginally, a distinct advantage. Moreover, the Bakri balloon is assembled, applied, and removed faster and more easily. Of note, the balloon surface did not adhere to surrounding pelvic structures, so it could be gently removed. Repeat surface bleeding, which can be triggered by separation of adhesion, was avoided. Our patient required no extra surgical or anesthetic procedures at the time of administration or removal of the balloon.

In addition, any continuing intraabdominal blood loss coming through the inner lumen of the balloon drainage port could be conveniently monitored without extra drains. The inflation port allows ready adjustment of the balloon’s volume to accommodate the size of the bleeding area at the time of application. Before removal, the pressure can be gradually decreased by serially deflating the balloon to ensure satisfactory hemostasis is maintained. At the same time, the balloon can be easily reinflated if further compression is needed.

After hysterectomy, parts of the umbrella pack, the Bakri balloon, or other pelvic devices generally can be inserted through the vaginal canal by using the opening at the vaginal stump. When the uterus remains in situ, as it did in our patient, posterior culdotomy can be performed to permit passage of the balloon shaft into the vaginal canal. The transverse culdotomy incision is made on the posterior cul-de-sac under direct visualization and is facilitated by tenting the posterior vaginal fornix; upward force is exerted with a sponge holder in the vagina. This straightforward procedure caused only a small amount of incisional bleeding.

Bacterial contamination from the vagina to the pelvis should be minimized because the balloon shaft was passed downward from the pelvis to the vagina. Alternatively, the modified Seldinger technique, commonly used for vascular access and entry to the lumen of hollow viscera, can be considered. Direct visualization of the vascular area plus sequential application to the posterior vaginal fornix of an introducer needle, a guide wire, a suitably sized posterior cul-de-sac incision, and variously sized dilators gradually expands the opening, resulting in a safe and perhaps more bloodless entry.

During repair of the incision at cesarean delivery, some surgeons prefer to exteriorize the uterus for better exposure of the incised area. Recently a systematic review examined extraabdominal vs intraabdominal repair of the uterine incision at cesarean delivery and concluded that extraabdominal closure was associated with a lower rate of febrile morbidity (relative risk, 0.41; 95% confidence interval, 0.17–0.97) but a marginally longer hospital stay (weighted mean difference, 0.24 days; 95% confidence interval, 0.08–0.39). There were no differences between the repair methods with regard to other outcomes. Because the review was based on a limited number of small trials and only a few studies provided data for most of the outcomes identified, any differences between the extraabdominal and the intraabdominal repairs in rare but serious complications might not be detected.

In our patient’s case, exteriorization of the uterus probably contributed to the breakage of the dense adhesion between the posterior uterine wall and the cul-de-sac, leading to hemorrhage. Because no clear evidence favors either approach for uterine incision repair, surgeons must weigh the possible benefits and risks of each course. Although current evidence does not support routine exteriorization of the uterus, the approach would be useful when more inaccessible incision sites must be exposed or when the angle of the low transverse incision tears.

Note that the balloon was placed in the cul-de-sac between our patient’s uterus and the pelvic floor to compress the bleeding surfaces on the posterior uterine wall and pelvic floor peritoneum. Given the limited space, an inflated volume of only 200 mL was needed. One could assume that the uterus itself keeps the balloon in place and provides some tamponade by pressing against it. For pelvic floor hemorrhage after hysterectomy, a balloon with a larger filling volume (up to 500 mL) should be considered. The surgeon should pay strict attention to how the shape of the balloon conforms to the dimension of the pelvis so that even pressure is applied to the pelvic sidewalls and the pelvic floor.

Use of a tamponade balloon rather than umbrella packing to control pelvic floor hemorrhage after hysterectomy or cesarean delivery requires further study. Possibly the balloon might be applicable for control of bleeding from the pelvic floor or other areas in the pelvis that are difficult to access in gynecological, oncological, colorectal, or urological surgeries. Currently the application of a Bakri balloon under any of these circumstances would be an off-label use in the United States.