Postpartum hemorrhage remains one of the major causes of maternal morbidity and mortality throughout the world.1 The most common causes of postpartum hemorrhage are uterine atony, abnormal placentation, lower genital tract lacerations, retained placental tissue, coagulopathies, vessel malformation, and uterine rupture.2 Because the majority of postpartum hemorrhages remain unpredictable, it is imperative that practitioners identify and manage them early.3 Effective primary management requires the use of uterotonic agents, manual exploration of the uterus, suturing possible lacerations, and fundal massage.3,4 In cases of persistent bleeding, a surgical approach including bilateral uterine or hypogastric artery ligation, peripartum hysterectomy, or both traditionally has been required. Nevertheless, these procedures expose the patient to surgical complications such as infection, bleeding, and ureteral injury.4–6 Pelvic arterial embolization has become a reliable and safe alternative procedure for postpartum hemorrhage that is strongly recommended by various authorities worldwide, but only when the patient is hemodynamically stable and the embolization unit is located close to the delivery room.3,7–8 However, studies assessing the efficacy and safety of pelvic arterial embolization have included too small a number of patients, which limits our understanding as regards this procedure.9 In particular, to our knowledge, predictive factors of failed pelvic arterial embolization for severe postpartum hemorrhage remain unidentified. The primary purpose of this study was to attempt to identify specific risk factors associated with an increased likelihood of failed pelvic arterial embolization from a large, single-center cohort because identification of these factors may assist caregivers in optimizing management of postpartum hemorrhage. The secondary objective was to attempt to estimate efficacy of pelvic arterial embolization in rare conditions, as secondary postpartum hemorrhage, and after a failed conservative surgical procedure.
MATERIALS AND METHODS
This study was approved by our institutional review board. All consecutive patients who had pelvic arterial embolization for postpartum hemorrhage at our tertiary care obstetric center (Rouen University Hospital) from May 1994 to July 2007 were included in the study. The long-term maternal outcome of the first 28 patients treated between 1994 and July 1999 has been reported previously,10 and the remaining cases until July 2007 were added to the present study. Patients who were referred to our institution from other centers where pelvic arterial embolization was not available or who had undergone a surgical procedure before or after the pelvic arterial embolization were included in the study. The only exclusion criteria were subsequent pregnancies with postpartum hemorrhage requiring pelvic arterial embolization in patients with a previous history of pelvic arterial embolization for postpartum hemorrhage, so as not to violate the assumption of independent observations. Our policy management for postpartum hemorrhage was reported previously4,11 (Fig. 1). Blood transfusions were performed when there was clinical evidence of inadequate oxygen-carrying capacity or of a hemoglobin concentration of less than 70 g/L, and fresh frozen plasma was transfused in the presence of consumption coagulopathy and persistent bleeding. Digital subtraction angiography was performed using a right-sided unifemoral approach. After aortography to determine anatomy and locate possible extravasation of contrast agent, selective catheterization of the uterine artery or anterior trunk of the hypogastric artery was performed. Supraselective catheterization of the uterine artery was attempted in all cases. Other anastomotic vessels such as vaginal branches were assessed when necessary. The same procedure then usually was repeated for the contralateral artery. The embolization material was mainly pledgets of absorbable gelatin sponge (Gelfoam, Upjohn, Kalamazoo, MI) and inert microparticles (Embosphere, BioSphere, Rockland, MA), but also fibered steel (William Cook Europe, Bjaeverskov, Denmark) and fibered platinum microcoils (Target Therapeutics, Boston Scientific, Natick, MA), which were used at the discretion of the angiographer (E.C.) performing the procedure. A control aortography was performed routinely to exclude residual leakage. Femoral sheaths were left in place for 24 hours. During the study period, two different treatments were performed for placenta accreta/percreta—an extirpative approach, in accordance with most previous recommendations from Europe and North America, with routine manual removal of the placenta to leave the uterine cavity empty12 and a conservative management, in accordance with recent publications, of leaving the placenta in part or entirely in the uterus.13–15
For each patient, obstetrical chart review was done manually to gather data regarding patients’ and neonates’ characteristics, causative factors of postpartum hemorrhage, type of delivery, preembolization treatments, blood transfusion, pelvic arterial embolization characteristics, immediate complications, and outcome of the procedure. Success of the pelvic arterial embolization was defined as an arrest of the hemorrhage after pelvic arterial embolization, whatever the number of pelvic arterial embolization procedures, with no subsequent surgical procedure. Definition of primary and secondary postpartum hemorrhage was a postpartum hemorrhage occurring within the first 24 hours and 24 hours to 6 weeks after delivery, respectively.2 Vascular abnormality was defined by one of the following: aneurysm, pseudo-aneurysm, arteriovenous malformation, or uterine artery injury at surgery. Wound infection was defined as purulent discharge from a surgical site.
Outcome was compared between failed and successful pelvic arterial embolization for postpartum hemorrhage. Statistical analyses included the Student’s t test, Mann Whitney test, χ2 test and Fisher exact test when appropriate. Differences were considered significant when the P<.05. Statistical analysis was carried out using StatXact.4 (Cytel Software Corporation, Cambridge, MA).
During this 13-year period, there were 101 pelvic arterial embolizations performed in 33,510 deliveries (1 per 332 deliveries). One patient had undergone pelvic arterial embolization for two different pregnancies. After exclusions, the study population consisted of 100 pelvic arterial embolizations, performed by the same radiologist (E.C.) in 87% of cases. Pelvic arterial embolization failed to control hemorrhage in 11 of 100 cases (11%), requiring a peripartum hysterectomy in 7 of the 11 cases (64%). Causes of postpartum hemorrhage and maternal outcome in cases of embolization failure are reported in Table 1.
Uterine atony was the most common cause of bleeding (53 of 100 pelvic arterial embolizations, 53%). Rates of success and failure of pelvic arterial embolization according to cause of postpartum hemorrhage are shown in Table 2. Failed pelvic arterial embolization occurred in the three cases of uterine artery injury at surgery (one case was combined with a splenic artery aneurysm) (Table 1), whereas pelvic arterial embolization controlled postpartum hemorrhage in all the other causes of vascular abnormality including aneurysm, pseudo-aneurysm, and arteriovenous malformation (Table 2).
Fifty patients (50%) were transferred from nine other institutions (Table 3). Pelvic arterial embolization was performed in 13 cases (13%) after a surgical failure (Table 3) (peripartum hysterectomy [two cases], stepwise uterine devascularization isolated [10 cases] or associated with bilateral hypogastric artery ligation [one case]). In 9 of these 11 latter cases (82%), pelvic arterial embolization demonstrated a patency throughout at least one ligated pedicle. Bilateral embolization was performed routinely, except for 11 cases (11%) (Table 3) owing to the clear identification of a vaginal artery injury with extravasation of the contrast agent (five cases of lower genital tract laceration) and unilateral effective ligation (six cases). Two pelvic arterial embolizations were required to achieve hemostasis in six cases (6%) (Table 3); in these six cases, postpartum hemorrhage was secondary to placenta accreta in three cases, vascular abnormality in two cases, and lower genital tract laceration in one case. Pelvic arterial embolization was performed in eight cases (8%) for secondary postpartum hemorrhage (Table 3) that occurred with a median period after delivery of 16 days (range 8–40 days) related to vascular abnormality (three cases, 37.5%), uterine atony and placenta accreta (two cases each, 25%), and lower genital tract laceration (one case, 12.5%).
Comparison of the failed and successful pelvic arterial embolization groups showed no differences in maternal, neonatal, labor, or delivery characteristics (Table 3). On univariate analysis, failure of pelvic arterial embolization was associated only with a higher rate of estimated blood loss more than 1,500 mL and more than 5 red blood cell units transfused (Table 3). No learning curve has been identified because there was no significant difference in the rates of pelvic arterial embolization success between two periods of the study (for example 1994–1997 compared with 1998–2008) or between the main radiologist (E.C.) and the others.
The frequency of wound infections was significantly higher in the failed pelvic arterial embolization group (3 of 11 [27%] compared with 5 of 89 [6%], P=.04), but the major and minor complications did not differ between the failed and successful pelvic arterial embolization groups (Table 4).
The present study of 100 attempts at pelvic arterial embolization mainly by the same radiologist during a 13-year period at our tertiary care center yielded several key findings: 1) on univariate analysis, the only factors that were found to be associated significantly with failed pelvic arterial embolization were a higher rate of estimated blood loss more than 1,500 mL and more than 5 red blood cell units transfused; 2) attempted pelvic arterial embolization after a failed vessel ligation procedure or for a secondary postpartum hemorrhage is an interesting option with success rates of 91% (10 of 11) and 87.5% (seven of eight), respectively; 3) major complication rates after pelvic arterial embolization were low (3%).
To our knowledge, this is the second largest series of patients treated by pelvic arterial embolization for postpartum hemorrhage reported in the literature.9,16 One of the strengths of this study is that pelvic arterial embolizations were performed in a single center and, in 87% of the cases, by the same radiologist (E.C.), which resulted in a limitation of the potential confounders. Placenta accreta/percreta was not demonstrated to be associated significantly with failed pelvic arterial embolization. Nevertheless, previous series, as in our study, report a lower pelvic arterial embolization success rate in cases of placenta accreta/percreta, which ranged between 60% and 83%,17 whereas the overall pelvic arterial embolization success rate ranged between 85% and 100%.16,17 Interestingly, conservative treatment for placenta accreta/percreta was successful in the five cases in which it was attempted, contrary to the extirpative treatment that was successful in only 50% (four of eight) of cases. The possible higher success rate for the conservative rather than extirpative treatment for placenta accreta/percreta was reported previously by our team15 as well as by other authors.13,14,17
Surprisingly, no learning curve could be identified. Moreover, no true predictive factor was found to be significantly associated with failed pelvic arterial embolization, possibly owing to a low statistical power. Specifically, twin pregnancy, chorioamnionitis, operative vaginal delivery, hospital-to-hospital transfer, nature of embolizing agent and arteries embolized, failed surgical procedure, secondary postpartum hemorrhage, cause of postpartum hemorrhage, and more than one pelvic arterial embolization were not significantly associated with failed pelvic arterial embolization. In fact, estimated blood loss more than 1,500 mL and more than 5 red blood cell units transfused assessed the final and not initial clinical status of the patient. Therefore, these two items were, not surprisingly, significantly associated with failed pelvic arterial embolization and could not be considered true or useful predictive factors.
To our knowledge, only one previous study has assessed pelvic arterial embolization for secondary postpartum hemorrhage.18 Pelage et al report 14 cases of severe secondary postpartum hemorrhage in which the median time of bleeding onset was 16.3 days (similar to our series). Pelvic arterial embolization was successful in all cases. The suspected cause of postpartum hemorrhage was genital tract laceration and retention endometritis, but vascular abnormalities were found during angiography in three cases.18 These three cases in addition to our three cases document a total of 27% (6 of 22) of vascular abnormality in cases of secondary postpartum hemorrhage. Our findings are therefore in accordance with Pelage et al and confirm that pelvic arterial embolization is a reasonable option for secondary postpartum hemorrhage with a success rate higher than 88%. Moreover, angiography could rule out misdiagnosis of vascular abnormality, the incidence of which seems to increase in secondary postpartum hemorrhage. It is important to underline that 1) the relatively low success rate of pelvic arterial embolization in cases of vascular abnormality (70%) that was observed in our study (Table 2) was underestimated because the cause of one failure could not be treated by pelvic arterial embolization (splenic artery aneurysm requiring splenectomy) (Table 1), and 2) the other failures occurred in the two cases of uterine artery injury at surgery and not in cases of aneurysm, pseudo-aneurysm, or arteriovenous malformation.
No studies have specifically assessed pelvic arterial embolization after a failed surgical procedure and particularly after a failed vessel ligation procedure. Only a few cases have been reported. Ornan et al report three cases of successful pelvic arterial embolization after peripartum hysterectomy and hypogastric artery ligation,19 and Mathe et al report one case of successful pelvic arterial embolization after a failed stepwise uterine devascularization.20 Recently, Gaia et al reported five cases of successful pelvic arterial embolization after hypogastric (three cases) and uterine (two cases) artery ligation.16 As regards these cases, patency or not throughout the ligated pedicle was never mentioned.16,19,20 Interestingly, in our series, patency throughout at least one ligated pedicle was observed in 9 of the 11 cases (82%) in which pelvic arterial embolization was performed after a vessel ligation (ie, stepwise uterine devascularization isolated or associated with bilateral hypogastric artery ligation); our team had experience in performing this type of conservative surgical procedure.4,11,21,22 Therefore, our results and those of the literature suggest that pelvic arterial embolization could be a reasonable option in a hemodynamically stable patient when bleeding occurs after a vessel ligation, in particular because vessel ligation could have been performed inadequately.
In France, all the maternity wards have been organized to manage postpartum hemorrhage medically and surgically (conservative surgical procedure and peripartum hysterectomy).23 Maternal hospital-to-hospital transfer for pelvic arterial embolization is possible only when the patient is hemodynamically stable.23 To our knowledge, there is no reported study in the literature that has assessed the safety and efficacy of a regional policy of hospital-to-hospital transfer for pelvic arterial embolization in postpartum hemorrhage. In our series, 50 patients (50%) were transferred from nine other institutions separated from our center, on average, by a distance of 18.3 miles (range 1.9 mi to 47.2 mi). Hospital-to-hospital transfer was not associated with a higher rate of failed pelvic arterial embolization. Nevertheless, this finding is insufficient to support a regional policy of hospital-to-hospital transfer for pelvic arterial embolization in postpartum hemorrhage. In fact, during the 13-year study period, patients who required emergency laparotomy with no attempt of pelvic arterial embolization owing to unstable hemodynamics on arrival in our intensive care unit and the possible patients who died during transfer were not identified.
The first main limitation of our study was its retrospective design, and all flaws of retrospective analyses apply. In particular, the long study period and the inclusion of patients from other centers raise concern relative to consistency of preembolization medical management of postpartum hemorrhage and clinical status. The second main limitation was that, although the number of pelvic arterial embolizations reported in our study is important, this series lacked statistical power because the sample size of the outcome of interest was low (n=11), failed pelvic arterial embolization being a rare event, fewer than one per year in our experience. This lack of statistical power did not permit us to perform multivariable analysis and might be responsible for not identifying true predictive factors of failed pelvic arterial embolization, in particular for placenta accreta/percreta. Finally, because reliable data regarding initial clinical status before the pelvic arterial embolization procedure were not available, one cannot completely exclude that worrying initial status was a predictor of failed pelvic arterial embolization, estimated blood loss more than 1,500 mL and more than 5 red blood cell units transfused assessing only the final status of the patient.
In conclusion, our large, single-center series confirms that pelvic arterial embolization for postpartum hemorrhage is effective with a low major complication rate. No true predictive factor was found to be significantly associated with failed pelvic arterial embolization. Pelvic arterial embolization for secondary postpartum hemorrhage or after failed vessel ligation procedure seems to be a reasonable option.
1. Khan KS, Wojdyla D, Say L, Gülmezoglu AM, Van Look PF. WHO analysis of causes of maternal death: a systematic review. Lancet 2006;367:1066–74.
2. Cunningham FG, Gant NF, Leveno KJ, Gilstrap LC, Hauth JC, Wenstrom KD, editors. Williams Obstetrics, 21st ed. New York (NY): McGraw Hill Companies; 2001.
3. Postpartum hemorrhage. ACOG Practice Bulletin No. 76. American College of Obstetricians and Gynecologists. Obstet Gynecol 2006;108:1039–47.
4. Sentilhes L, Trichot C, Resch B, Sergent F, Roman H, Marpeau L, et al. Fertility and pregnancy outcomes following uterine devascularization for severe postpartum haemorrhage. Hum Reprod 2008;23:1087–92.
5. Joshi VM, Otiv SR, Majumder R, Nikam YA, Shrivastava M. Internal iliac artery ligation for arresting postpartum haemorrhage. BJOG 2007;114:356–61.
6. Knight M, UKOSS. Peripartum hysterectomy in the UK: management and outcomes of the associated haemorrhage. BJOG 2007;114:1380–7.
7. The role of emergency and elective interventional radiology in postpartum haemorrhage. Good Practice No. 6. Royal College of Obstetricians and Gynaecologists. Available at: www.RCOG.org.uk
. Retrieved March 18, 2009.
8. Goffinet F, Mercier F, Teyssier V, Pierre F, Dreyfus M, Mignon A, et al. Hémorragies du post-partum: recommandations du CNGOF pour la pratique clinique (Décembre 2004). Gynecol Obstet Fertil 2005;33:268–74.
9. Doumouchtsis SK, Papageorghiou AT, Arulkumaran S. Systematic review of conservative management of postpartum hemorrhage: what to do when medical treatment fails. Obstet Gynecol Surv 2007;62:540–7.
10. Descargues G, Mauger-Tinlot F, Douvrin F, Clavier E, Lemoine JP, Marpeau L. Menses, fertility and pregnancy after arterial embolization for the control of postpartum hemorrhage. Hum Reprod 2004;19:339–43.
11. Sentilhes L, Gromez A, Razzouk K, Resch B, Verspyck E, Marpeau L. B-Lynch suture for massive persistent postpartum hemorrhage following stepwise uterine devascularization. Acta Obstet Gynecol Scand 2008;87:1020–6.
12. American College of Obstetricians and Gynecologists. Postpartum hemorrhage. ACOG Educational Bulletin 243. Washington, DC: ACOG; 1998.
13. Kayem G, Davy C, Goffinet F, Thomas C, Clément D, Cabrol D. Conservative versus extirpative management in cases of placenta accreta. Obstet Gynecol 2004;104:531–6.
14. Bretelle F, Courbière B, Mazouni C, Agostini A, Cravello L, Boubli L, et al. Management of placenta accreta: morbidity and outcome. Eur J Obstet Gynecol Reprod Biol 2007;133:34–9.
15. Sentilhes L, Resch B, Clavier E, Marpeau L. Extirpative or conservative management for placenta percreta? Am J Obstet Gynecol 2006;195:1875–6.
16. Gaia G, Chabrot P, Cassagnes L, Calcagno A, Gallot D, Botchorishvili R, et al. Menses recovery and fertility after artery embolization for PPH: a single-center retrospective observational study. Eur Radiol 2009;19:481–7
17. Pelage JP, Limot O. Place actuelle de l’embolisation artérielle dans la prise en charge des hémorragies graves du post-partum immédiat. Gyn Obstet Fertil 2008;36:714–20.
18. Pelage JP, Soyer P, Repiquet D, Herbreteau D, Le Dref O, Houdart E, et al. Secondary postpartum hemorrhage: treatment with selective arterial embolization. Radiology 1999;212:385–9.
19. Ornan D, White R, Pollak J, Tal M. Pelvic embolization for intractable postpartum hemorrhage: long-term follow-up and implications for fertility. Obstet Gynecol 2003;102:904–10.
20. Mathe ML, Morau E, Vernhet-Kovacsik H, Picot MC, Boulot P. Impact of new French clinical practice recommendations in embolization in postpartum and post-abortion hemorrhage: study of 48 cases. J Perinat Med 2007;35:532–7.
21. Resch B, Sergent F, Blanc S, Baron M, Sentilhes L, Trichot C, et al. Comment je fais … une ligature des artères hypogastriques au décours d’une hémorragie grave de la délivrance. Gynecol Obstet Fertil 2008;36:85–7.
22. Resch B, Sergent F, Blanc S, Baron M, Sentilhes L, Trichot C, et al. Comment je fais … une ligature des artères utérines au décours d’une hémorragie grave de la délivrance. Gynecol Obstet Fertil 2008;36:88–9.
23. Bagou G, Collège National des Gynécologues et Obstétriciens Français, Agence Nationale d’Accréditation et d’Evaluation en Santé. Modalités d’un transfert maternel dans le cadre d’une hémorragie postpartum. J Gynecol Obstet Biol Reprod (Paris) 2004;33:4S89–92.