The effects of pelvic artery embolization (PAE) for postpartum hemorrhage (PPH) on subsequent pregnancies have been explored in small case series and one case-control study by mailed questionnaire with uncomplicated pregnancies as controls. We conducted a single-center retrospective cohort study using women with PPH without PAE for comparison.
From a cohort of 103 women undergoing PAE for primary PPH between January 1999 and December 2012 (exposed) and 189 pregnancies with PPH not requiring PAE between January 2008 and December 2012 (unexposed), we queried the electronic medical records for readmissions to labor and delivery in subsequent years. Outcomes of subsequent pregnancies continuing past 20 weeks were obtained by chart review.
Repeat pregnancies were documented in 17 of 103 exposed women (16.5%) and 18 of 189 unexposed women (9.5%). At delivery complicated by PPH, the groups did not differ in demographics, gestational age, units of blood transfused, or PPH cause. At the time of subsequent deliveries, there was a greater interdelivery interval in women exposed to PAE than those unexposed (1710 ± 938 days vs 904 ± 358 days; P = .002), and the 2 groups were similar in terms of gestational age and birthweight. However, there was a significantly higher rate of placenta accreta in exposed than unexposed women (23.5 % vs 0%; P = .04), with 3 of 17 sustaining total abdominal hysterectomy and 1 requiring repeat PAE for severe PPH.
Pregnancies following PAE for PPH were more likely than those not receiving PAE for treatment to be complicated by placenta accreta. Pregnancies following PAE should be followed up for imaging evidence of placenta accreta.
Pelvic artery embolization (PAE) has become part of the obstetrician’s armamentarium in the developed world as a fertility-sparing treatment of postpartum hemorrhage (PPH). There are multiple small case series of ongoing pregnancy following PPH. Some authors have reported complications including fetal growth restriction, recurrent PPH, and abnormal placentation, whereas others suggest no sequelae of the procedure in subsequent pregnancies.
The most recent reviews on the topic suggest that fertility can be preserved but that subsequent pregnancies may be at a small but enhanced risk for placental disorders and recurrence of PPH. The only prior case-control study used, as controls, women with normal pregnancies who were contacted by a mail-in questionnaire; this study suggested no difference in obstetric outcomes between the 2 groups.
Given the inconsistency in data regarding recurrent pregnancy after PAE for PPH and the relatively large experience with this procedure at Inova Alexandria Hospital, we have conducted a retrospective cohort study comparing women with PPH and receiving PAE with those with PPH not receiving PAE.
Materials and Methods
Following institutional review board approval, we identified a cohort of 103 women with primary PPH at longer than 20 weeks and receiving PAE between January 1999 and December 2012 (exposed) and a second cohort of 189 women with primary PPH >20 weeks not requiring PAE between January 2008 and December 2012 (unexposed). The reason for the different time frames in the 2 groups was that there was an extant, ongoing database specifically for PAE since its inception at our hospital, but a new recording method for general obstetric complications was introduced in 2008.
PPH was defined by the need for transfusion, dilation and curettage procedure, PAE, and/or hysterectomy following delivery. For this analysis, patients requiring hysterectomy were necessarily excluded. Electronic medical charts were queried for readmissions to labor and delivery in subsequent years with pregnancies longer than 20 weeks.
Our electronic medical record system is centralized for Inova Health Systems, which comprises 7 hospitals in northern Virginia, allowing access to data on patients whose subsequent deliveries occurred in the region but not at our hospital. All diagnoses of PPH were confirmed by chart review; in the case of accreta, pathology records specifically were reviewed. Exposed and unexposed cohorts were compared for clinical factors at the index pregnancy, including maternal age, gestational age at delivery, parity, route of delivery, number of units of packed red blood cells (pRBCs) transfused, and causes of PPH.
Exposed and unexposed cohorts were also compared regarding subsequent delivery factors including interdelivery interval, gestational age at delivery, birthweight, route of delivery, recurrence of PPH, and cause of PPH. Statistical analyses included a Student t test for continuous data, a Mann-Whitney U test for nonparametric data, and a Fisher exact test for categorical data, with P < .05 considered significant.
From 103 pregnancies treated with PAE for PPH, we identified 17 ongoing pregnancies (16.5%), and from 189 pregnancies treated without PAE for PPH, we identified 18 ongoing pregnancies (9.5%) ( Figure ). Exposed and unexposed cohorts were well matched in terms of all clinical factors queried ( Table 1 ), including etiology of PPH and units of blood transfused. As anticipated, there was a greater interval to a subsequent delivery in the exposed cohort than those unexposed, which likely reflects the earlier inception of the database specific to PAE ( Table 2 ).
|Characteristic||Exposed: received PAE (n = 17)||Unexposed: no PAE (n = 18)||P value|
|Maternal age, y||30.5 ± 5.5||29.0 ± 6.0||.5|
|Gestational age at delivery, wks||38.2 ± 2.3||36.4 ± 5.8||.3|
|Nulliparity, n (%)||12/17 (70.5)||8/18 (44.4)||.1|
|Cesarean delivery, n (%)||10/17 (58.8)||10/18 (55.6)||1.0|
|Median units of pRBC transfused||3 (0-30)||2 (0-8)||.1|
|Cause of PPH, n (%)||.3|
|Atony||10/17 (58.8)||5/18 (27.7)|
|Abruption||1/17 (5.9)||4/18 (22.2)|
|Previa||1/17 (5.9)||1/18 (5.6)|
|Cervical/vaginal laceration||3/17 (17.6)||2/18 (11.1)|
|Uterine artery laceration||0||1/18 (5.6)|
|Uterine rupture||0||1/18 (5.6)|
|Retained products of conception||0||2/18 (11.1)|
|DIC with preeclampsia||0||1/18 (5.6)|
|Variables||Exposed: received PAE (n = 17)||Unexposed: no PAE (n = 18)||P value|
|Interdelivery interval, d||1710 ± 938||904 ± 358||.002|
|Gestational age at delivery, wks||37.7 ± 2.6||38.7 ± 1.1||.2|
|Birthweight, g||3200 ± 612||3228 ± 467||.9|
|Cesarean delivery, n (%)||10/17 (58.8)||14/18 (77.7)||1.0|
|Cesarean hysterectomy, n (%)||3/17 (17.6)||0/18 (0)||.1|
|Recurrent PPH, n (%)||4/17 (23.5)||1/18 (5.5)||.2|
|Cause of PPH, n (%)|
|Previa/ruptured uterus||0/17 (0)||1/18 (6)||1.0|
|Accreta||4/17 (23.5)||0/18 (0)||.045|
Recurrent PPH was observed in 23.5% of the cohort exposed to PAE, and in all 4 cases, this was due to pathology-proven accreta. Three patients received hysterectomies and the fourth received a dilation and curettage, blood products, and a repeat PAE ( Table 3 ). In the unexposed group, there was 1 recurrent PPH requiring transfusion (5.5%), which was due to uterine rupture and placenta previa with a fetus in transverse lie and a repeat cesarean scar ( Table 3 ).
|Index pregnancy presenting with PPH||Subsequent pregnancy outcome|
|Clinical scenario||Cause of PPH||Treatment for PPH||Interdelivery interval (days)|
|36 year old P0, type 2 DM, large myoma |
Primary CD for abruption at 35 wks
Live born, 2305 g
30 U pRBC, FFP
Bilateral hypogastric artery embolization
|2551||Interpregnancy spontaneous abortion of myoma |
Cesarean hysterectomy at 29 wks for previa/accreta in setting of FDU
|29 year old G2P0, suspected macrosomia |
Primary CD at 39 wks
Live born, 4165 g
Bilateral uterine artery embolization
|910||Cesarean hysterectomy at 34 5/7 wks for bleeding previa/accreta |
Live born, 3050 g
|28 year old G1P0 |
NSVD at 38 wks
Live born, 3415 g
4 U pRBC
Bilateral uterine artery embolization
|1193||Cesarean hysterectomy at 36 6/7 wks for accreta, 5 units pRBC, FFP |
Live born, 3433 g
|32 year old G1 |
NSVD at 41 wks
Live born 2693 grams
3 U pRBC
Bilateral hypogastric artery embolization
D and C 6 wks PP
|1215||NSVD at 39 wks |
Live born, 3450 g
PPH with D and C for partial accreta
3 U pRBC and UAE
|28 year old G2P1, previa, prior CD |
Repeat CD at 37 wks
Live born, 3800 g
2 units pRBC
|367||Repeat CD at 36 wks previa, transverse lie |
Live born, 2336 g
PPH: previa/ruptured uterus
Uterotonics, surgical repair 3 U pRBC