Complications and failure of uterine artery embolisation for intractable postpartum haemorrhage


Drs MS Maassen, Department of Obstetrics and Gynaecology, Albert Schweitzer Hospital, PO Box 444, 3300 AK Dordrecht, the Netherlands. Email:


Objective  Primary postpartum haemorrhage (PPH) is a major cause of maternal morbidity and mortality around the world. Most patients can be managed conservatively, but patients with intractable bleeding require more aggressive treatment. In these cases uterine artery embolisation (UAE) has proven to be a useful tool to control PPH. The reported success rate of UAE is over 90% with only minor complications. In this case series we studied the effectiveness and complications of UAE.

Design  Retrospective analysis of a case series.

Setting  Case series in a large peripheral hospital in the Netherlands.

Sample  Eleven patients who were treated with UAE for intractable PPH from November 2004 to February 2008.

Methods  In this paper we review the results of all patients treated with UAE for intractable PPH in our hospital and focus on the two cases with adverse outcomes.

Main outcome measures  Effectiveness, causes of failure of UAE, complications.

Results  Nine out of eleven patients were treated successfully with UAE. One patient needed an emergency hysterectomy for intractable bleeding. In the aftermath she developed a vesicovaginal fistula (VVF). Another patient suffered a major thrombo-embolic event of the right leg, for which she underwent embolectomies and despite fasciotomy a necrotectomy.

Conclusions  UAE is a valuable tool in managing major PPH and in most cases it can replace surgery and thus prevent sacrification of the uterus. However, due to blood supply of the uterus by one of the ovarian or aberrant arteries, UAE might fail to control the bleeding. In addition, serious complications such as a thrombo-embolic event or VVF may occur. We hereby present a case of migration of an embolus from the site of re-embolisation into the femoral artery requiring immediate intervention to prevent the loss of the lower leg. This complication demonstrates that gelatine sponge particles could migrate from the internal iliac artery into the external iliac artery.


Primary postpartum haemorrhage (PPH) is a major cause of maternal morbidity and mortality.1 In the Netherlands, PPH (defined as blood loss of 1000 ml or more from the genital tract) is the third most frequent cause (8%) of maternal mortality.2

Most patients can be managed by using uterotonic agents, surgical repair of lacerations, removal of retained placental tissue, uterine packing or correction of coagulation disorders. Women with intractable bleeding require more aggressive treatment; such methods include bilateral ligation of the uterine or internal iliac arteries and/or hysterectomy.1 However, surgical approaches may be associated with increased operative risk and morbidity, caused by compromised haemodynamic condition and coagulopathy. In addition, it can be very difficult to operate in a region where pregnancy has caused major anatomical changes.3

In 1979 Brown et al. reported on transcatheter embolisation of the internal iliac arteries or uterine arteries (UAE) to control PPH.4 The reported success rate of UAE is over 90%5–8 and up to 100% for intractable bleeding caused by atony.9 We adopted this procedure in our hospital and in this paper we review all cases of intractable PPH treated with UAE in our facility between November 2004 and February 2008. We draw attention to two cases with adverse outcomes, including one failure and two major complications.

Materials and methods

Between November 2004 and February 2008, 11 women with intractable PPH underwent UAE. Table 1 shows the relevant clinical characteristics of those studied.

Table 1.  Clinical characteristics of 11 patients who underwent arterial embolisation
Case no.AgeParityMaturityMode of delivery/indicationAetiologyTreatmentArteries embolisedSuccess*Transfusion (no. PC)
  • APH, antepartum haemorrhage; CS, caesarean section; IOL, induction of labour; IUFD, intrauterine fetal death; L, left; MROP, manual removal of placenta (l parts); PC, packed cells; PP, placenta praevia; R, right; (S)VD, (spontaneous) vaginal delivery,

  • *

    Success is bleeding controlled. No additional surgical intervention needed.

133P441 + 0IOL, VD, MROPAtonyConservative, uterine packingInternal iliac, twiceNo, hysterectomy17
226P136 + 6SVD, MROP, adherent placentaAtony, placenta incretaConservative, uterine packingInternal iliac R and L and ovarian artery RYes, after re-embolisation10
340P031 + 3IUFD, IOL, HELLP, VDAtonyConservative, Uterine packingInternal iliacYes8
431P041 + 0IOL, CS, prolonged 2nd stageAtonyConservativeInternal iliacYes8
527P040 + 1CS, prolonged 2nd stageHaematoma intra-abdominalRelaparotomyL uterine artery, R anterior internal iliacYes22
629P138 + 2CS, abruptio placentasAtonyConservative, B-lynch, uterine packing,Internal iliacYes11
735P041CS, prolonged 2nd stage, MROPAtonyConservative, B-lynchInternal iliacYes10
835P140 + 2SVD, MROP, adherent placentaAtony, placenta incretaConservative, uterine packingInternal iliacYes6
937P233 + 1CS, PP, APH, breechBleeding from CSRelaparotomy, bleeding from CSInternal iliacYes11
1039P041 + 3CS, prolonged 2nd stageAtonyConservativeInternal iliacYes4
1132P038 + 5IOL, HELLP, VD, MROPAtonyConservativeInternal iliacYes10

All women were initially managed by conservative means, including administration of oxytocin, manual uterine massage, fluid resuscitation, catheterisation of the bladder, prostagladine E2 analogues [sulprostone], manual removal of placental tissue, and, in three patients, uterine packing. In two patients a B-lynch suturing10 was performed. Two other patients needed UAE after relaparotomy for persistent bleeding following caesarean section.

The embolisation procedures were all performed by an interventional radiologist who injected small particles of absorbable gelatin sponge [Spongostan (also called Willospon), Will-Pharma, Wavre, Belgium] into the anterior division of the internal iliac artery via bilateral transfemoral catheters. A postembolisation selective angiogram was performed in all patients to ensure complete occlusion of the internal iliac arteries. Occlusion was considered to have been achieved when there was no flow for a duration of five heartbeats. The average time from the initial puncture of the common femoral artery until occlusion of the internal iliac artery was 28 minutes (range 22–43 minutes). The duration of the entire procedure was approximately 55 minutes (range 49–70). The sheaths were left in the femoral arteries to prevent leakage (no haematoma formation occurred) and to enable expeditious re-embolisation if necessary.


During the period reviewed there were 8779 deliveries in our hospital. There were 419 women with PPH (4.8%). All of them needed blood transfusion. There were no maternal deaths due to PPH.

Eleven women with PPH needed a UAE (2.6%) and the procedure was successful in nine. No patient underwent a hysterectomy before, or instead of, a UAE.

Case descriptions

Case 1

A 33-year-old woman, gravida 5 para 4, with an uneventful medical history, was induced at 41 weeks by means of artificial rupture of the membranes and intravenous administration of oxytocin. Her labour and delivery were uneventful and a healthy daughter, weighing 3460 g was born. The third stage of labour was managed actively with intravenous oxytocin; the placenta was delivered by controlled cord traction with an estimated blood loss of 400 ml. One hour later she had a PPH of 1200 ml due to an atonic uterus. Some retained placental tissue was removed manually in theatre. The uterus remained atonic and she continued to bleed after packing of the uterine cavity. The patient was transferred to the angiography suite for bilateral embolisation of the internal iliac arteries (start procedure 45 minutes after decision due to transportation to a different location of our hospital. Postembolisation arteriography showed occlusion of both internal iliac arteries. However, there was persistent bleeding and a re-embolisation was performed. Spongostan was again injected into both internal iliac arteries. Despite the fact that pre- and postembolisation angiography showed complete occlusion, the woman continued to bleed. Because of the massive blood loss (6500 ml) further angiographic exploration of the ovarian arteries was abandoned and an emergency hysterectomy was performed. At the time of laparotomy, a large atonic non-vital uterus of a greyish colour was removed. The only remaining blood supply appeared to come from the left ovarian artery. The parametria, cervix and apex of the vagina did not bleed excessively when transected, suggesting adequate embolisation at these levels.

In the puerperium, the patient complained of severe pain in her left leg and a neuropathy of the sciatic nerve was diagnosed. This spontaneously resolved within 3 months. The patient was discharged 2 weeks after delivery. Four weeks after the embolisation procedure the patient developed a vesicovaginal fistula (VVF) that required surgical repair.

Histopathological examination of the uterus showed massive oedema and loss of the mucosa. There was subserosal purple discoloration from ischaemia. A thrombus on the dorsal part of the uterus marked the implantation site of the placenta. There was widespread infiltration of the blood vessels of the myometrium by gelatin sponge particles.

Case 2

A 26-year-old woman, gravida 2 para 1, spontaneously delivered a healthy boy weighing 3310 g at 36 weeks and 6 days. A retained placenta and atony of the uterus caused a PPH of 2000 ml. During manual removal in theatre, the placenta appeared to be morbidly adherent to the uterus. Because of intractable haemorrhage, a bilateral embolisation of the internal iliac arteries was performed (the procedure was started 30 minutes after the decision was made). Postembolisation arteriography showed occlusion of both internal iliac arteries (Figure 1). The bleeding reduced, but did not stop satisfactorily. Therefore, a repeat embolisation was performed 6 hours later. At this time, persistent flow was noted through some branches of both uterine arteries and both internal iliac arteries were re-embolised. On juxtarenal arteriography, anastomic blood supply of the uterus via the right ovarian artery was seen (Figures 1 and 2). A decision was made to sacrifice the right ovary in an attempt to curtail the bleeding and this artery was occluded using a coil (Cook Europe, Bjaeverskov, Denmark). Postprocedural arteriography showed occlusion of the right ovarian artery and both internal iliac arteries. At the time of selective injection of the contrast agent into the internal iliac artery at the final angiogram (Figure 3), some gelatin sponge material was dislodged from the internal iliac artery and migrated retrogradely into the external iliac artery. This resulted in acute occlusion of the anterior and posterior tibial arteries, the peroneal artery, and small femoral arterial branches supplying the skin of the right lower leg (Figure 4). After the final angiogram the patient complained of pain in her right leg. There was marked pallor of the right leg with loss of sensation and motor function of the right foot. There were no arterial pulses detected in the anterior and posterior tibial arteries (Figure 5). An emergency embolectomy with fasciotomy was performed to prevent a compartment syndrome, but despite this blood flow was not restored. Re-embolectomy was performed, after which blood flow to the foot was seen through the posterior tibial artery. The anterior tibial artery remained occluded. Following the second embolectomy, there was a palpable pulse in the posterior tibial artery and the patient’s leg was warm with normal skin colour. It was decided to anticoagulate the patient with heparin to prevent reocclusion of the posterior tibial artery since her uterine blood loss had remained negligible. Although she suffered from a paralysis of the peroneal nerve, rehabilitation was initially satisfactory using a peroneal splint. Three weeks postpartum, the patient was discharged from our hospital. Unfortunately, despite the immediate fasciotomy, she developed necrosis of the peroneal, extensor digitorum and anterior tibial muscles and necrotectomy with closure of the wound using a split skin graft after secondary wound granulation was necessitated.

Figure 1.

Angiogram after embolisation of the right internal iliac artery. The large arrow indicates the internal iliac artery with its occluded side branches. Filling of the right ovarian artery (small arrow) via the internal iliac artery is seen.

Figure 2.

Blood supply of the uterus via the anastomic right ovarian artery is seen (arrow marks site of derivation from aorta).

Figure 3.

Retrograde flow from of Spongostan from the right internal iliac artery into the right external iliac artery (arrow) at the final angiogram.

Figure 4.

Occlusion of the right anterior tibial artery (small arrow) and posterior tibial artery (large arrow) and the peroneal artery (arrowhead), as well as occlusion of small femoral branches supplying the skin.

Figure 5.

Pale right leg with livores before embolectomy.


The blood supply of the uterine fundus is derived from the uterine (90%) and the ovarian (10%) arteries. Both the uterine artery and the vaginal and cervical branches originate from the internal iliac artery.11 Therefore, embolisation of the internal iliac artery should stem bleeding from any of these arteries. The ovarian arteries originate directly from the aorta at the level of the renal arteries or from the left renal artery.11 Embolisation of an ovarian artery usually results in the loss of that ovary.

In nine patients in our series the main indication for embolisation was uterine atony, and in two of these cases there was a morbidly adherent placenta. Two patients in our series suffered from persistent bleeding after caesarean section. Two patients had to be re-embolised for intractable haemorrhage (Table 1).

Based on this series, the PPH rate in our hospital was 4.8%, which is comparable with the rate in the literature.1

We performed a literature search on possible causes of intractable haemorrhage after embolisation. Table 2 shows a summary of the studies on uterine artery embolisation for PPH.5–8,12,13

Table 2.  Summary of studies on uterine artery embolisation for postpartum haemorrhage
Author (year)No. of ptsSuccess (%)Intervention (no. of pts)DeliveryCause HPPSuggested cause of failure
  1. ?, not mentioned; CS, caesarean section; PP, placenta praevia; PPROM, preterm prelabour rupture of membranes; SVD, spontaneous vaginal delivery. All patients who underwent hysterectomy before embolisation were excluded. Saraiya et al. treated one patient for leiomyomata.

Soncini et al.51493 (13/14)Hysterectomy (1)G1, 36 weeks, PIH, twinUterine atonyAbnormal placentation
Tsang et al.68100 (8/8)Collaterals
Pelage et al.72588 (22/25)Re-embolisation (2)??Abnormal placentation
Hysterectomy (1)?Placenta accreta
Yong and Cheung82893 (26/28)Hysterectomy (1), uterine packing (1)CSUterine atonyHaemodynamically unstable
Descargues et al.12771 (5/7)Hysterectomy (2)G3P1, PPROM 32 weeks, SVDPlacenta incretaAbnormal placentation
G3P1, PP, CSPP accreta
Saraiya et al.131100 (1/1)Collaterals by round ligament artery

These studies show that uterine atony is the most frequent cause of PPH, accounting for 67% of the UAE procedures.7 In cases of uterine atony, the reported success rate of UAE is 100%9; however, the numbers of patients in these case series are small (see Table 2).

It is suggested that abnormal placentation accounts for over half of the failures of UAE.12 Reported success rates of UAE for this indication are 71%.7,12 Failure is thought to be due to myometrial injury caused by difficult digital separation of the placenta. Delaying the UAE procedure could contribute to the failure rate.12 In our case of a failed UAE, the placenta was not pathologically adherent; only minimal residual tissue was discovered in the uterine cavity, and this was easily removed. Histopathological examination did not show any signs of abnormal placentation. The uterine atony did not respond to uterotonic agents and uterine packing. The most likely cause for the ongoing bleeding after UAE of both internal iliac arteries is the supply of the fundal part of the uterus by one of the ovarian arteries.

A re-embolisation can be done relatively quickly, since the sheaths in the femoral arteries are usually left in for 24 hours. During this procedure, it is important to check that occlusion of the internal iliac arteries has been achieved. In addition, an effort should be made to visualise any accessory arteries. Embolisation of these arteries could possibly prevent a hysterectomy for ongoing bleeding.

A juxtarenal angiogram is necessary to visualise uterine perfusion via the ovarian arteries. Such an angiogram was not performed in our first case, as the patient was haemodynamically too unstable to postpone surgery any longer. The findings during laparotomy did suggest persistent blood supply to the uterus by the left ovarian artery.

The second patient was haemodynamically stable enough to undergo a juxtarenal angiogram. Once the accessory blood supply to the uterus via the right ovarian artery had been shown, it was decided to embolise it, accepting the inevitable loss of the right ovary.

UAE has been reported to be a valuable and safe tool for managing intractable postpartum haemorrhage. Minor complications after UAE are frequently reported: pain, transient fever, mild transient numbness of the buttock, foot or thigh, haematoma formation at the site of the common femoral artery puncture, and pelvic infection.5,6 Complications from embolus migration to the general blood circulation are very rare.5 Ischaemic uterine necrosis after UAE with polyvinyl alcohol and Gelfoam (Pharmacia) particles was reported by Cottier in 200214 and in their case a hysterectomy was finally performed because of pelvic pain and persistent menorrhagia.14

It is known that urinary tract injury may complicate 15% of emergency postpartum hysterectomies.15 However, we feel that the VVF seen in our patient was more likely to have been caused by UAE-related necrosis. The involuntary loss of urine, which started 4 weeks after delivery, and the size of the VVF defect (2 × 2 cm) make a surgical cause less likely. In 2005 Porcu et al. reported uterus and bladder necrosis after UAE for PPH.16

It is unlikely that the neuropathy of the left sciatic nerve in our patient was caused by ischaemia since her complaints did not occur immediately after the procedure and the neuropathy resolved spontaneously within 3 months.

During the re-embolisation procedure of our second case, the radiologist was convinced that the position of the catheter in the internal iliac artery was correct. However, despite optimal localisation, retrograde flow of gelatin sponge particles occurred at the final angiogram. This case shows that it is very important to be careful when injecting contrast agent selectively in an artery filled with gelatin sponge as embolus migration to the systemic circulation can occur.


We present one failure and two major complications of embolisation for intractable PPH. The lesson learned from our first case is that despite successful UAE, ovarian arterial supply of the uterine fundus can still result in sufficient blood flow to cause persistent uterine bleeding. This accessory blood supply was confirmed by angiogram in our second case and the artery was occluded with a coil. We therefore suggest that if UAE fails, a juxtarenal angiogram should be performed to rule out blood supply by the ovarian arteries.

Injecting contrast agent into the internal iliac arteries after re-embolisation with gelatin sponge particles carries the risk of particle migration into the femoral artery. To the best of our knowledge we present the first case of a thromboembolic event during re-embolisation that needed immediate intervention to prevent amputation of a leg.

Disclosure of interests

I hereby declare on behalf of all authors that there is no conflict of interest as mentioned in the author guidelines.

Contribution of authorship

The contributions by each author is as follows: M.S. Maassen was main author, M.D.A. Lambers: first supervisor of main author, P.H.M. van der Valk: repeated and extensive revision of the article, R.P. Tutein Nolthenius: revision of case 2, and O.E. Elgersma: second supervisor of main author.

Details of ethical approval

Since this is a retrospective case-series, there was no protocol sent to any ethics committee for approval. IRB-approval: Both patients whose cases are described in this article approve of their case being described in this case series.


There is none.


Acknowledgements are due to L. Hofman for correcting our English grammar.