Balloon occlusion of the internal iliac arteries in the multidisciplinary management of placenta percreta

Authors


  • Conflict of interest
    The authors have stated explicitly that there are no conflicts of interest in connection with this article.

  • Clausen C, Stensballe J, Albrechtsen CK, Hansen MA, Lönn L, Langhoff-Roos J. Balloon occlusion of the internal iliac arteries in the multidisciplinary management of placenta percreta. Acta Obstet Gynecol Scand 2012;
    DOI 10.1111/j.1600-0412.2012.01451.x.

Caroline Clausen, Department of Cardiovascular Radiology, Rigshospitalet, Copenhagen University Hospital, Blegdamsvej 9, DK-2100 Copenhagen Ø, Denmark. E-mail: clausen.caroline@gmail.com

Abstract 

Objective. To evaluate our experience with prophylactic balloon occlusion of the internal iliac arteries as a part of a multidisciplinary algorithm for the management of placenta percreta. Design. Consecutive case series. Setting. Rigshospitalet, Copenhagen University Hospital, Denmark. Sample. Seventeen women with placenta percreta. Methods. Demographic characteristics, intraoperative data and outcomes are summarized and discussed. Main outcome measures. Feasibility of local resection, intraoperative blood loss and transfusion requirements. Results. The multidisciplinary management allowed for local resections in nine of the 11 women who requested preservation of fertility. The mean intraoperative blood loss was 4050 mL (range 450–16 000 mL, median 2500 mL). Adhesions to the bladder or the parietal peritoneum were associated with an intraoperative blood loss >6000 mL. Conclusions. Prophylactic balloon occlusion of the internal iliac arteries as part of a multidisciplinary algorithm allowed for a safe management of all cases in our consecutive series of 17 women with placenta percreta. However, intraoperative blood loss and transfusion requirements were significant. We have therefore decided to modify our multidisciplinary algorithm to include balloon occlusion of the common iliac arteries rather than the internal iliac arteries.

Key Message

Prophylactic balloon occlusion of the internal iliac arteries as part of a multidisciplinary treatment algorithm, together with hemostatic control resuscitation and appropriate surgical measures, allows for safe management of women with placenta percreta and for preservation of fertility in selected cases.

Introduction

Placenta percreta is a rare life-threatening obstetrical condition in which the placenta invades the entire depth of the uterine wall. Placenta accreta and increta represent less severe variants of the same condition (1,2). The most important risk factors are a placenta previa and previous cesarean section (1,3–6). Women with placenta percreta are at risk of severe third-trimester bleeding, uterine rupture and severe postpartum hemorrhage (1,3–6). Severe hemorrhage may be associated with massive transfusion, surgical complications, hysterectomy, coagulopathy and disseminated intravascular coagulation, multi-organ failure and death (1,3–6).

The incidence of placenta percreta has increased over recent decades, probably as a consequence of the rise in cesarean section rates (1,3,4). Several centers have argued for centralization of management in highly specialized maternity units. Maternal morbidity is reduced in these high-risk pregnancies by a carefully planned multidisciplinary approach (7,8). Early diagnosis by ultrasound in high-risk patients is central to identifying cases for follow-up during pregnancy and for multidisciplinary planning of the complicated cesarean section procedure that is required. The purpose of this study was to evaluate our experience with prophylactic balloon occlusion of the internal iliac arteries as a part of a multidisciplinary algorithm for the management of placenta percreta.

Material and methods

In 2008, we developed a multidisciplinary treatment algorithm for the management of placenta percreta patients treated at Rigshospitalet, Copenhagen University Hospital, Denmark. In this study, we report the results of 17 consecutive placenta percreta cases from February 2008 to June 2011. After the initial diagnosis in pregnancy, by ultrasound and color Doppler imaging, an individualized plan was made for each woman, covering multidisciplinary intervention in the event of excessive bleeding during pregnancy, and for a multidisciplinary activity at elective cesarean section at 34 weeks gestation (9).

The multidisciplinary treatment algorithm on the day of delivery included the following steps.

  • (i) Cesarean section performed under epidural analgesia, and conversion to general anesthesia after the infant was born unless bleeding was sparse.
  • (ii) Prophylactic temporary occlusion balloon catheters were placed in both internal iliac arteries shortly after the insertion of the epidural catheter.
  • (iii) A local extirpative surgical technique was chosen to maintain fertility when requested and feasible.
  • (iv) Close hemostatic monitoring and hemostatic resuscitation with an early proactive balanced transfusion therapy guided by thrombelastography. This approach was chosen on the basis of current understanding of the cell-based model of hemostasis, and because it is associated with improved survival in massive bleeders (10,11).
  • (v) Tranexamic acid was given immediately after delivery and repeated after every 10th to 15th blood product administered.
  • (vi) Intraoperative blood loss was estimated by adding the volume in suction containers, the weight of swabs, and a visual estimation of vaginal blood loss.

The multidisciplinary team involved the departments of obstetrics, anesthesiology, blood bank, interventional radiology, neonatology and fetal medicine.

Endovascular technique

Arterial access to both common femoral arteries was gained using the percutaneous Seldinger technique under local regional anesthesia with lidocaine. Occlusion balloon catheters were placed in the most proximal division of the internal iliac arteries through 6 French cross-over sheaths in the contralateral common femoral artery. Sufficient diluted contrast medium to occlude the arteries was registered; usually about 0.5–1.0 mL (Figure 1). The syringes with the appropriate volume were fixed to both catheters to avoid the use of fluoroscopy during cesarean delivery. Sheaths and balloon catheters were fixed with skin sutures and a tape cover. Following the cutting of the umbilical cord, the occlusion balloons were inflated, and deflated again after approximately 30 min when hemostasis was secured. Endovascular instruments were removed immediately after wound closure, and hemostasis at the puncture sites was obtained by manual compression.

Figure 1.

Injection of contrast media in the right common iliac artery, proximal to the inflated occlusion balloon in the right internal iliac artery, passes through collaterals from the right external iliac artery to the right internal iliac artery and the placenta. The right ovarian artery is seen.

Approval for this study was granted by the Danish Data Protection Agency.

Results

All 17 consecutive patients were diagnosed with placenta percreta by ultrasound and color Doppler imaging in pregnancy, and the diagnosis was confirmed clinically at the operation or by histological examination. The placental position was anterior in nine cases and anterior/posterior in eight cases. Two cesarean sections had to be performed earlier than planned; one because of vaginal bleeding and the other because of preterm premature rupture of membranes and contractions. Prophylactic placement of internal iliac artery occlusion balloon catheters was not possible in these two patients for logistic reasons. One surgeon performed 16 of 17 cesarean sections and one interventional radiologist performed 13 of 15 occlusion balloon placement procedures. Demographic characteristics, intraoperative data and outcomes are summarized in Table 1. All 17 cesarean sections could be done under epidural analgesia. In 14 cases, the patient underwent general anesthesia after the infant was born.

Table 1. Summary of demographic characteristics, intraoperative data and outcomes.
Case no.Age (years)Gravida/paraPrevious cesarean section/or previous uterus surgeryGestational age at delivery (weeks+days)Balloon occlusion of the internal iliac arteriesSurgical procedureIntraoperative blood loss (mL)Packed red blood cell transfusion 0 –24 hour (mL)Fresh frozen plasma transfusion 0 –24 hour (mL)Platelet transfusion 0 –24 hour (mL)Length of surgery (min)
  1. a Myomectomy.

  2. b Tubotomy.

  3. c Conization of the cervix.

  4. d Cesarean section earlier than planned.

  5. e Abnormal placentation.

  6. f Re-operated without balloon occlusion. Intraoperative bloodloss 3500 mL.

1335/33/032+6YesHysterectomy2000000155
2320/00/1a33+6YesLocal resection19009000045
3343/32/032+2YesHysterectomy7000390033001500120
4343/32/034+1YesHysterectomy20006000075
5363/42/032+0YesHysterectomy8450420036001500188
6343/2e1/032+4YesLocal resection3500180090060080
7351/21/034+3YesLocal resection140000038
8373/32/035+3YesLocal resection1300600600090
9383/21/035+2YesLocal resection2500600600090
10302/32/033+5dNoHysterectomy600030001500600145
11393/11/033+6YesHysterectomy200060060030065
12391/21/034+5YesHysterectomy16 000750045001800205
13323/1e0/1b34+1YesLocal resection25006000060
14322/1e1/033+1YesLocal resection3500f2250100035075
15392/11/1c30+1dNoHysterectomy45025000110
16354/21/033+6YesLocal resection600020001890200100
17342/11/2c33+4YesLocal resection23401500600097
Mean34.93/21/033+4405017801120400100
Median343/21/033+625009006007090

Five women had an estimated intraoperative blood loss >6000 mL or more. These women had significant adhesions to the bladder or the parietal peritoneum, which was not observed in the remaining women. Hysterectomy was also associated with increased intraoperative blood loss. The mean intraoperative blood loss for the eight women who had a hysterectomy was 5490 mL (range 450–16000 mL) and for the nine women who had a local resection it was 2770 mL (range 1300–6000 mL). The average duration of the surgical procedure for hysterectomy was 133 min(range 65–205 min), compared with 75 min(range 38–100 min) for local resection. Six of eight hysterectomies were requested prior to operation.

One woman's intraoperative blood loss was significantly higher than in the rest of the series; it was 16 000 mL. She had significant adhesions to the bladder, a hysterectomy was done and, furthermore, it was not possible to inflate the right occlusion balloon immediately when requested.

Hemostatic control resuscitation with thrombelastography-guided transfusion therapy revealed no severe coagulopathy or disseminated intravascular coagulation. Goal-directed use of coagulation factor concentrates, such as recombinant factor VIIa, fibrinogen or prothombin complex, was not indicated according to the thrombelastography. No thromboembolic events, or other complications related to the endovascular intervention, were detected. The mean fluoroscopic time was five minutes (range 3–7.5 min) and a mean of 40 mL (range 20–80 mL) contrast media was used (iodixanol; 270 mg iodine/mL).

One woman needed postoperative intensive care for 8.5 h. Another woman had to be re-operated 1 h after the cesarean section due to excessive postoperative bleeding. In another case, the remaining cervical tissue was considered insufficient for cervical competence in future pregnancies, and an abdominal cerclage was applied. Two of five patients with significant adhesions to the bladder or the anterior abdominal wall experienced complications. In one case, it led to rupture of the bladder and in the other to hematoma on both sides of the peritoneum. In one case, retained placental tissue was removed hysteroscopically because of bleeding 7 weeks after cesarean section.

The mean birthweight of the babies was 2360 g (range 1550–3660 g). The Apgar scores at five minutes were normal. The patients stayed in hospital for a mean of 27 days before cesarean section (range 0–59 days), and for a mean of 8 days after cesarean section (range 5–21 days).

Discussion

The preoperatively planned multidisciplinary treatment algorithm, including temporary occlusion balloon insertion into the internal iliac arteries, well-planned surgical techniques and hemostatic control during resuscitation, allowed for a safe management of all cases in our consecutive series of 17 women with placenta percreta. Furthermore, the overall set-up allowed for local resection in nine of the 11 who requested preservation of fertility if possible. We found that prophylactic placement of temporary occlusion balloon catheters within the most proximal division of the internal iliac arteries was technically feasible, safe and without complications. However, blood loss and transfusion volumes were significant, and it is uncertain to what extent balloon occlusion of the internal iliac arteries contributed to controlling the bleeding intraoperatively.

Different techniques in surgery, anesthesiology and interventional radiology have been used to reduce the intraoperative blood loss in these high-risk pregnancies. Hysterectomy is often necessary in the management of placenta percreta, but in recent years there are an increasing number of women who request maintained fertility with local extirpative or conservative treatment. There are several different surgical hemostatic techniques, such as ligation, uterine compression sutures, Cho's compression sutures, B-Lynch and so on, in the management of these high-risk pregnancies (12,13). Usually, the surgeon's personal preferences, individual differences in patients’ anatomy and the course of events during surgery determine the choice, priority and sequence of the techniques applied. In our setting, one radiologist and one surgeon carried out the large majority of procedures, which might be one of the reasons for the good results. Another reason is probably that we have established a good multidisciplinary team to take care of the rather restricted number of percreta cases. Both speak in favor of regional “centers of excellence.”

Hemostatic control during resuscitation seeks early control of coagulopathy as a supplement to the surgical control of bleeding, and provides transfusions in an immediate and sustained manner as part of the transfusion protocol for massively bleeding patients. Transfusion of red blood cells, fresh frozen plasma and platelets are initiated in a similar proportion to that seen in whole blood (ratio 1:1:1). This prevents both hypovolemia and coagulopathy, and is guided by thrombelastography (11). Epidural analgesia was chosen rather than spinal analgesia for the cesarean section, because it causes less vasodilatation, can be controlled more precisely, and because the catheter can be used for postoperative pain management. Furthermore, it allows the women to participate in the delivery. Conversion to general anesthesia after the delivery was chosen for the sake of optimal management of massive hemorrhage. However, all women were also given the opportunity to have general anesthesia for the entire procedure.

The contributions of interventional radiologists have been the placement of occlusion balloons, pelvic arterial embolization or a combination of these two techniques (14–23). Placement in the internal iliac arteries is most frequently employed for occlusion balloons, but positions as proximal as the abdominal aorta and as distal as within the anterior division of the internal iliac arteries have been employed (15,16,18,19,22,23). Regarding interventional radiological hemostasis techniques, 33 reports presenting one or more placenta percreta cases have been published to date. Heterogeneity characterizes these case series. They often include not only patients with placenta percreta, but also placenta accreta, increta and previa. Furthermore, most case series include only one or two placenta percreta cases, the condition associated with the highest risk of maternal and fetal morbidity and mortality, including severe hemorrhage and cesarean hysterectomy. Soyer et al. found that less than 50% of women in a case series including placenta accreta, increta or percreta needed pelvic arterial embolization (21). Within the same report, cases are often not managed in a uniform fashion regarding surgical techniques, as well as interventional radiological techniques. Outcome measures also vary among studies, and transfusion protocols and ways of estimating the intraoperative blood loss differ substantially between centers. Therefore, comparison of results from different studies is obviously complicated and entails reservations.

Tan et al. and Carnevale et al. found that balloon occlusion of the main trunk of the internal iliac arteries reduced intraoperative blood loss and transfusion requirements when comparing the patients retrospectively with a control group (16,22). Shrivastava et al. compared occlusion of the anterior divisions of the internal iliac arteries with a control group and found no significant differences (19). A review of the literature by Alanis et al. in 2006 demonstrated arterial embolization to be effective in treating placenta increta in women who wished to preserve fertility. In 72 women, they found a 76.9% success rate and an 11% complication rate (24). Additionally, recent studies have been in agreement with the suggestions from this review and have concluded that arterial embolization is effective in the majority of cases with abnormal placentation (14,17,20,21). However, in one of these studies blood transfusions were reported to be 12.5 red blood cell units (range 4–41 units) and 5.2 fresh frozen plasma units (range 2–20 units). These large transfusion volumes indicate that the intraoperative blood loss was considerable (21). Bodner et al., in 2006, did not find that the combination of balloon occlusion of the anterior division of the internal iliac arteries and arterial embolization reduced intraoperative blood loss in abnormal placentation (15). In comparison to embolization, the use of occlusion balloons has the advantage of being completely reversible immediately after the procedure, and the exposure to ionizing radiation is lower.

It can be supposed that the extensive collateral circulation to the placenta percreta uterus limits the hemostatic effect of balloon occlusion, embolization and arterial ligation (21,25,26). Anastomoses are possible among the uterine, ovarian, vesical and vaginal arterial systems, besides the lumbar, median sacral, inferior mesenteric and iliolumbar arteries (21,25,26). Furthermore, the vaginal arteries can get a supply from the external iliac artery via anastomoses among the medial femoral circumflex, internal pudendal and obturator arteries in cases of of placenta percreta (21,25,26). This is clearly shown in Figure 1, where contrast media injected into the right common iliac artery, proximal to the inflated occlusion balloon in the right internal iliac artery, passes through collaterals from the right external iliac artery to the right internal iliac artery and the placenta. Furthermore, the right ovarian artery is seen.

Balloon occlusion of the infrarenal aorta is often described in the management of hemorrhagic shock associated with pelvic fractures, and reported to be without serious complications related to the procedure (27). The ischemia caused by temporary occlusion of the aorta followed by reperfusion increases the oxidative stress and local inflammatory response, and thus can affect the glutathione redox status (28). It is described that the duration of aortic clamping can be 1 h without altering the glutathione redox status in patients without iliac occlusions, undergoing elective abdominal aneurysm surgery (28). However, occlusion of the infrarenal aorta does cause extensive hemodynamic changes. Reports on a few cases have described the use of endovascular balloon occlusion of the infrarenal aorta or common iliac arteries in the management of placenta percreta (14,23,29,30). Balloon occlusion of the infrarenal aorta or common iliac arteries, compared with balloon occlusion of the internal iliac arteries, will additionally block the supply to the placenta percreta from the external iliac arteries. Position of an occlusion balloon in the infrarenal aorta also arrests blood flow from the median sacral, low lumbar and, in some cases, the inferior mesenteric artery, depending on the origin of the artery and the position of the balloon. However, blood flow from the ovarian and higher lumbar arteries will not be controlled.

Based on the evaluation of our consecutive case series of 17 women with placenta percreta, the clinical experience of a small change in bleeding when balloons were inflated in the internal iliac arteries, and considering the rich collateral circulation from the external iliac arteries, we have decided to modify our multidisciplinary treatment algorithm. The algorithm now includes balloon occlusion of the common iliac arteries rather than the internal iliac arteries in the elective setting, and the use of an infrarenal aortic occlusion is reserved explicitly for the acute setting.

In conclusion, prophylactic balloon occlusion of the internal iliac arteries as part of a multidisciplinary treatment algorithm, together with hemostatic control resuscitation and appropriate surgical measures, allows for safe management of women with placenta percreta. However, further observational studies are needed on the optimal balloon location, but we suggest balloon occlusion of the common iliac arteries rather than the internal iliac arteries.

Funding

No specific funding.

Acknowledgements

The authors wish to thank all colleagues participating in the multidisciplinary treatment algorithm at Rigshospitalet, Copenhagen University Hospital.

Ancillary