Placenta praevia is a major cause of obstetric haemorrhage1 and an important source of maternal and fetal morbidity throughout the world, resulting in high demands on health care resources. The risk of morbidity is higher when placenta praevia is complicated by accreta and percreta.2
Placenta praevia occurs when the placenta is located completely or partly in the lower uterine segment. The terms placenta accreta, placenta increta and placenta percreta describe successively more serious conditions in which the placenta morbidly adheres to the myometrium, penetrates the decidua basalis and then penetrates the entire myometrium to the uterine serose.2 The increasing number of caesarean sections, combined with increasing maternal age, means that it is likely that the number of cases of placenta praevia and its complications, including placenta accreta, will continue to increase.1
In the past, managing massive obstetric haemorrhage associated with placenta praevia was difficult and hysterectomy was often required, with the inevitable loss of fertility. However, when placenta praevia is complicated by percreta, hysterectomy may not guarantee the cessation of bleeding.1
Following the introduction of recent interventional radiological procedures, it has become possible to control postpartum haemorrhage and, therefore, conserve the uterus.2
Materials and methods
A 31-year-old woman, gravida 3, parity 2, presented for the first time at 38 weeks' gestation with mild vaginal bleeding. She had had one previous caesarean section and one vaginal delivery. She had been diagnosed with complete placenta praevia, and accreta was suspected after her admission to hospital. She provided a history of three previous episodes of vaginal bleeding at 27 weeks' and two episodes at 28 weeks' gestation. These episodes had been managed conservatively at another institution, where she had received steroids (intravenously or orally) for respiratory distress syndrome prophylaxis.
On admission to hospital, her vital signs were normal and bleeding was minimal. On examination, her uterus size was consistent with normal limits for 38 weeks' gestation, the fetus was in the cephalic presentation and cardiotocography of the fetus was normal. Colour Doppler ultrasonography revealed central placenta praevia covering the internal os. Blood tests showed maternal platelet counts of 85–89 × 109/l on repeated samples, with a normal coagulation profile and haemoglobin of 11.5%.
Pre-eclampsia was ruled out by a series of normal blood pressure readings, the absence of protinuria and normal liver function tests. Blood urea electrolytes and normal blood film ruled out the possibility of HELPP syndrome (haemolysis, elevated liver enzymes and low platelet count). In view of the patient's low platelet count, a haematologist reviewed the patient and gave a provisional diagnosis of gestational thrombocytopenia or immune thrombocytopenic purpura, with further investigations to be performed after delivery. He advised that 6 units of platelets and 4 units of packed red blood cells be prepared before surgery.
Prior to the delivery, a multidisciplinary team was arranged, including obstetricians, an interventional radiologist, an anaesthetist and a haematologist. The intensive care unit and neonatologist were also involved in the subsequent care of the mother and the newborn.
The patient and her husband were both counselled regarding the findings and the risks and benefits of caesarean section. Particular emphasis was placed on the increased risk of bleeding in the presence of major placenta praevia and accreta. They couple was also informed of the interventional radiological procedures, for which written consent was obtained, and they were counselled regarding the possible need for hysterectomy. They were extremely reluctant to agree to a hysterectomy, which added to the challenge of management.
Preoperatively, bilateral uterine artery catheterization and balloon insertion were performed through a puncture in the bilateral common femoral arteries. This was achieved by a selective catheterization of the uterine arteries via the anterior division of the internal iliac arteries. A contrast injection was performed to obtain angiographic images in order to measure the diameter of these arteries and select the correct size of balloons. After assessing the angiographs, 6 mm × 2 cm long balloons were inserted bilaterally into the uterine arteries. They were fixed securely in place to avoid any displacement during patient transportation and positioning. The two balloons were tested in the angiography suite to confirm their effectiveness and position.
During the caesarean section, it was noted that the lower uterine segment was extremely vascular. The placenta was centrally placed, covering the whole lower segment anterolaterally. A high transverse uterine incision, above the site of the placental attachment, was made and the baby was delivered successfully and had good Apgar scores.
A 100-μg dose of carbetocin (Pabal®, Ferring, Saint-Prex, Switzerland) was given intravenously and spontaneous separation of the placenta occurred within 2–3 minutes. A small, 2–3 cm, anterior segment of placenta remained unseparated but was later found to be accreta and, therefore, separated manually. The lower segment and the cervical area were explored carefully but no remnants of placental tissue were found; however, in view of active bleeding from the placental bed, further uterotonics were given alongside uterine massage and heat therapy. The uterus began to contract but the placental bed continued to bleed; therefore, the radiologist inflated the balloons to the calculated volume to achieve vessel occlusion. This quickly reduced the vaginal bleeding and, as a result, B-Lynch sutures did not need to be considered as the uterus was contracting and the bleeding was already controlled by uterine artery balloon inflation. The total intraoperative blood loss was approximately 600–700 ml. The decision to perform uterine artery embolization was made and the patient was transferred to the angiography suite under continued sedation and mechanical ventilation.
Uterine artery embolization
Bilateral iliac angiography was performed through the in situ groin sheaths. The angiograms confirmed the correct positioning of the balloons and the effective closure of both uterine arteries (Figure 1). The two balloons were removed over a wire and selective catheterization of both uterine arteries was performed. There was active extravasation from the left uterine artery; therefore, both uterine arteries were embolized using a gelfoam mixture (Figures 2 and 3). The angiograms showed that the post-embolization appearance was good and there was no flow into the uterine arteries or active extravasation (Figure 4).
The patient suffered from recurrent vaginal bleeding post embolization, while still undergoing angiography. Fluid resuscitation was given with the replacement of red blood cells, platelets, fresh-frozen plasma and recombinant factor VIIa. Further uterotonic medicines including oxytocin (Syntocinon®, Novartis, Basel, Switzerland) infusion, methylergometrin (methylergnovie malerate, Novartis, Basel, Switzerland) injection and two further doses of intramuscular carboprost (Hemabate®, Pfizer, New York, NY, USA) and misoprostol (Cytotec®, Pfizer, New York, NY, USA) per rectum were given and bimanual uterine compression was continued.
Uterine artery embolization was repeated for the whole anterior division of internal iliac artery (bilateral and unilateral). The angiograms showed a good post-embolization appearance and the immediate cessation of bleeding. There was no recurrent vaginal bleeding and blood pressure was maintained throughout the procedure. The patient had good urine output and an estimated total blood loss was approximately 3000 ml. She received 6 units of red blood cells, 6 units of platelets, 4 units of fresh-frozen plasma and two doses of recombinant factor VIIa during this procedure.
The patient was transferred to the intensive care unit, where she was taken off sedation and was extubated the following day. Her haematological variables returned to acceptable limits and a chest radiograph showed no signs of acute lung injury. After 2 days, she was moved to the postnatal ward, where she had a smooth postnatal recovery, and she was discharged on the sixth day after the operation with her baby, who was fit and healthy.
Obstetric haemorrhage due to placenta praevia with accreta is a difficult condition for obstetricians to manage and presents new challenges in modern times.1 When haemorrhage is severe, successful management depends on timely recognition and intervention with the appropriate measures. Prior to delivery, all women with placenta praevia, and their partners, should be advised of the risk of haemorrhage, blood transfusion and the possibility of hysterectomy. Any concerns, queries or refusals of treatment should be dealt with effectively and documented clearly.2 Management of massive haemorrhage should follow a standard locally agreed protocol.
Temporary measures can be divided into immediate and advanced measures.
Immediate measures include:
Medications. The uterotonic agents can be very helpful in reducing the blood loss associated with atony, as in the patient described above, in whom bleeding from the relatively atonic lower uterine segment and the placental bed created additional problems.
Bimanual compression, which was used effectively in the case above.
Bilateral uterine artery balloon inflation.
Intrauterine balloon or packing.
Intraoperative compression of lower abdominal aorta. This is required only if the above measures are not effective.
Advanced measures include:
Uterine compression sutures. B-Lynch and lower-segment caesarean section mass sutures were not required in the case above as uterine atony was not evident at the time of initial surgery. Atonic postpartum haemorrhage occurred postoperatively and post embolization while the patient was in the radiology suite and equipment for re-embolization was quickly available.
Uterine artery embolization, as was performed in the case above.
Ligation of hypogastric arteries or uterine arteries could be performed if embolization was not available, had failed or was inadequate.
Advanced techniques may also be required if local measures, used to control bleeding, are unsuccessful. Ligation of the uterine arteries or hypogastric arteries may be performed; however, both operations require laparotomies. Bilateral hypogastric artery ligation has a success rate ranging from 40% to 100%, whereas uterine artery ligation has been shown to have a success rate of 92% and a complication rate of 1%.2
Transcatheter arterial embolization has been a recognized method of haemorrhage control since the 1960s and has been used in the control of pelvic haemorrhage due to malignancy, trauma and radiation.3 More recently, this technique has been used successfully in the control of postpartum haemorrhage, with the first case reported in 1979.2 Uterine artery embolization may be performed as an alternative management and interventional radiology can be life saving and uterus sparing in cases of massive obstetric haemorrhage2 and to have this facility available locally is highly desirable.4
Placement of intravascular balloon catheters has been performed at various sites, from as proximal as the aorta to more distally within the anterior division of the internal iliac arteries. More often than not, this technique has been combined with concomitant arterial embolization. The rational of inserting intravascular balloon catheters is to quickly decrease uterine blood flow as required during, or after, surgery. This has two advantages: first, it leads to reduced total blood loss and a lower morbidity rate and, second, surgery can be performed under easier and more controlled circumstances with less profuse haemorrhage.4
Uterine artery embolization has several advantages, including easy identification of the bleeding site, preservation of the uterus and potential fertility and decreased rebreeding from collaterals with more distal occlusion of the bleeding vessels.4
Garmi et al. 5 suggested that failure of intravascular balloon catheters to reduce blood loss may be explained by the extensive degree of uterine blood flow with pregnancy and the extensive vascular anastomoses present in the gravid pelvis. In addition, although reduction of blood flow to the uterine arteries probably occurs following balloon inflation in the hypogastric arteries, collateral circulation from cervical, ovarian, rectal, femoral, lumbar and sacral arteries probably contributes to the overall blood loss.5
In our case, the uterine artery balloons reduced blood loss successfully and allowed easier transfer to embolization. Following deflation and removal of the balloons and the gelfoam embolization, there was a significant rebleed. The factors likely to be responsible for this recurrent haemorrhage were considered and included (a) cervical implantation of the placenta leading to recruitment of extrauterine blood supply that had not been embolized, (b) poor uterine contraction capability of the lower segment, or (c) a low platelet count. This case illustrates that a prompt re-embolization of a wider arterial territory may be required in complex cases.
Complications of pelvic embolization for postpartum haemorrhage occur at a rate of 8.7%.4 The most common complication is low-grade fever and rarer complications include pelvic infection, groin haematoma, iliac artery perforation, transient buttock ischaemia, transient foot ischaemia and bladder gangrene.4 Regarding fertility, Stancato-Pasiket al. 6 reported the follow-up of 12 embolized patients: in 11 (92%) of the 12 women, normal menses resumed within 2–5 months of the procedure. There were no complications related to embolotherapy and the follow-up period was 1–6 years. All three patients who wished to conceive had full-term, healthy newborns. The results of another study suggested that women who undergo arterial embolization for obstetric haemorrhage should expect to return to normal menses with preservation of future fertility and successful uneventful pregnancies.7 The exact effect on infertility from uterine artery embolization is unknown for postpartum haemorrhage but it does provide a higher fertility rate than the alternative, which is hysterectomy.
If both uterine artery embolization and uterine artery ligation are available, embolization is the first choice as it obviates laparotomy and ligation can be attempted subsequently if embolization is unsuccessful. In contrast, after an unsuccessful uterine artery ligation, embolization may be extremely difficult or even impossible, leaving hysterectomy as the only remaining option.4
From a review of the British literature, it would appear that arterial embolization for the control of postpartum haemorrhage is a technique that is not widely used or available. A previous survey of maternity units in the UK showed that 86% of these units had never used this technique.4 It is believed that this therapy is underused and its increasingly important role in the control of postpartum haemorrhage has been emphasized.4
A recommendation of the latest triennial report on maternal mortality is to consider uterine artery embolization in postpartum haemorrhage. This has led to attempts to increase 24-hour availability of emergency uterine artery embolization to all maternity units by providing interventional on-call radiologists.
Obstetric haemorrhage is uncommon but can lead to life-threatening complications during childbirth and a high morbidity rate. It is vital not to lose treatment time by late diagnosis and/or inadequate care. With so many treatment options, ultimately the choice depends on the nature of clinical scenario, individual preference and availability of resources. Immediate effectiveness, a lower morbidity rate and conservation of fertility make embolization the method of choice in the framework of a multidisciplinary care system. In this case, as in many other cases, selective uterine arterial embolization was effective in the control of intraoperative and postpartum haemorrhage and avoided the need for hysterectomy.