- Email: [email protected]
The association between intrauterine balloon tamponade duration and postpartum hemorrhage outcomes
Accepted Manuscript The association between intrauterine balloon tamponade duration and postpartum hemorrhage outcomes Brett D. Einerson, MD MPH, Moeun Son, MD, Patrick Schneider, MD, Ian Fields, MD, Emily S. Miller, MD MPH PII:
S0002-9378(16)30976-0
DOI:
10.1016/j.ajog.2016.10.040
Reference:
YMOB 11370
To appear in:
American Journal of Obstetrics and Gynecology
Received Date: 15 September 2016 Revised Date:
24 October 2016
Accepted Date: 28 October 2016
Please cite this article as: Einerson BD, Son M, Schneider P, Fields I, Miller ES, The association between intrauterine balloon tamponade duration and postpartum hemorrhage outcomes, American Journal of Obstetrics and Gynecology (2016), doi: 10.1016/j.ajog.2016.10.040. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Einerson
1
Title: The association between intrauterine balloon tamponade duration and postpartum hemorrhage outcomes.
RI PT
Authors: Brett D EINERSON MD MPH1; Moeun SON MD2; Patrick SCHNEIDER, MD3; Ian FIELDS, MD4; Emily S. MILLER, MD MPH2
Affiliations 1Department
of Obstetrics & Gynecology, Division of Maternal-Fetal Medicine, University of
M AN U
Utah Health Sciences Center, Salt Lake City, Utah 2Department
SC
City: Chicago, Illinois
of Obstetrics & Gynecology, Division of Maternal-Fetal Medicine,
Northwestern Feinberg School of Medicine, Chicago, Illinois 3Department
of Obstetrics & Gynecology, Division of Maternal-Fetal Medicine, University of
4Department
TE D
Chicago, Chicago, Illinois
of Obstetrics & Gynecology, Northwestern Feinberg School of Medicine,
Chicago, Illinois
EP
Corresponding author: Brett D. Einerson, 50 N Medical Drive, Salt Lake City, UT 84132;
AC C
Phone 801-585-5156; Email: [email protected] Disclosures: The authors report no conflict of interest. Presented: The 63rd Annual Scientific Meeting of the Society for Reproductive Investigation (SRI), March 16-19, 2016 in Montreal, Quebec, Canada.
Word Count: Abstract 386, Main text 2,006
ACCEPTED MANUSCRIPT Einerson
2
Condensation: Postpartum hemorrhage-related outcomes were similar in those with intrauterine balloon tamponade duration beyond 12 hours compared to those with
AC C
EP
TE D
M AN U
SC
Short title: Optimal intrauterine balloon tamponade duration
RI PT
removal at or before 12 hours.
ACCEPTED MANUSCRIPT Einerson
3
ABSTRACT Background: Intrauterine balloon tamponade (IUBT) is an effective treatment for
management of postpartum hemorrhage (PPH) is unclear.
RI PT
postpartum hemorrhage when first line treatments fail. The optimal duration of IUBT for
Objective: To determine whether IUBT removal >12 hours duration is associated with
SC
PPH-related clinical outcomes.
M AN U
Study Design: A retrospective cohort study of women with PPH from 2007-2014 who underwent IUBT. We excluded failures of IUBT (IUBT expulsion with duration <2 hours or if hysterectomy was required prior to planned IUBT). Patients who underwent IUBT for 212 hours were compared to those who underwent IUBT for >12 hours. Examined PPHrelated clinical outcomes included estimated blood loss (EBL) after IUBT placement, blood
TE D
product transfusion, use of adjuvant measures to control PPH after IUBT (either uterine artery embolization (UAE) or hysterectomy), and maternal intensive care unit (ICU) admission. Secondary outcomes examined included postpartum fever and hospital length
EP
of stay (LOS). Multivariable logistic regression models were used to control for
AC C
confounding variables.
Results: Of 274 eligible women, 206 (75%) underwent IUBT for >12 hours and 68 (25%) underwent IUBT for 2-12 hours. The median EBL after IUBT placement (190 versus 143 mL, p=0.116), as well as the frequencies of blood product transfusion (62.1% versus 51.5%, p=0.120), transfusion of 4 or more units of packed red blood cells (17.0% versus 14.7%, p=0.659), UAE (15.1% versus 16.2%, p=0.823), hysterectomy (0.0% versus 1.5%, p=0.248), and ICU admission (8.7% versus 7.4%, p=0.721) were not statistically different between
ACCEPTED MANUSCRIPT Einerson
4
the groups and this lack of association persisted in multivariable regressions. IUBT duration >12 hours was associated with a higher frequency of postpartum fever (27% versus 15%, p=0.047) and a longer mean hospital LOS (3.7 versus 3.1 days, p=0.002). After
RI PT
adjusting for variables that differed statistically between groups, the difference in length of stay associated with IUBT duration was no longer present, but the association between IUBT duration >12 hours and postpartum fever persisted (OR 2.33, 95% CI 1.07-5.11).
SC
Including chorioamnionitis as an independent variable in a post hoc multivariable analysis diminished the association between IUBT >12 hours and postpartum fever (adjusted OR
M AN U
2.04, 95% CI 0.92-4.53).
Conclusion: There are no significant differences in PPH-related outcomes associated with IUBT duration >12 hours compared to removal 2-12 hours. If ongoing hemorrhage has
initial placement.
TE D
abated, it is reasonable to consider removal of an intrauterine balloon by 12 hours after its
AC C
EP
Key words: intrauterine balloon tamponade, postpartum hemorrhage
ACCEPTED MANUSCRIPT Einerson
5
TEXT Introduction
RI PT
Aggressive treatment is essential in the management of postpartum hemorrhage (PPH), the world’s leading cause of maternal mortality.1 When first line treatments such as uterotonic agents and bimanual uterine massage fail, intrauterine balloon tamponade (IUBT) has been
SC
shown to be effective, with reported bleeding cessation rates of 55-94%.2–4 However,
previous studies have been limited to case series and small observational studies, and thus
M AN U
many questions remain about the appropriate management of the intrauterine balloon devices once they have been successfully placed and hemorrhage has abated. Specifically, it remains unknown what the optimal duration of IUBT is with regard to PPH-related outcomes. IUBT is proposed to stop obstetric hemorrhage by direct tamponade of the
TE D
vascular placental bed, encouragement of uterine contractility, and/or direct hydrostatic pressure on the uterine arteries.5,6 Whether these actions are successful and sustained after a set duration of IUBT has not been previously studied.
EP
We sought to determine if IUBT duration is associated with PPH-related outcomes. We hypothesized that a longer duration of IUBT would be associated with increased
AC C
effectiveness in the management of PPH.
Materials and Methods
This is a retrospective cohort study of all women who underwent IUBT at Northwestern Memorial Hospital in Chicago, Illinois from January 2007 to October 2014. The estimated
ACCEPTED MANUSCRIPT Einerson
6
blood loss (EBL) is routinely recorded in the delivery form of our electronic medical record. A query of cases with an EBL of 500 mL or more after a vaginal delivery or 1,000 mL or more after a cesarean yielded all cases of PPH. Medical records for each case of PPH were
RI PT
then reviewed to identify women undergoing IUBT. At the study institution, IUBT was
recommended by protocol as a second-line option for the management of PPH refractory to uterotonic agents and bimanual uterine massage. The protocol indications for use of IUBT
SC
included placental site bleeding and uterine atony. The IUBT device used was the Bakri Postpartum Balloon (Cook Medical Inc, Bloomington, Indiana). We excluded women whose
M AN U
balloon was removed within the first two hours of placement due to expulsion (n=33) as well as women who underwent hysterectomy prior to planned IUBT removal for refractory hemorrhage (n=7) as these circumstances indicated a failure of IUBT rather than an outcome related to the duration of IUBT placement. We identified no cases in which IUBT
TE D
was discontinued intentionally before 2 hours due to perceived success in controlling PPH. After placement of IUBT, patients underwent a period of observation on the labor and
EP
delivery unit with close monitoring of vital signs and ongoing estimation of blood loss. The duration of IUBT was at the discretion of the clinical provider and was not dictated by
AC C
protocol; thus, this practice pattern varied at the study institution. A designated nurse was tasked with recording the events of the PPH episode, including ongoing estimation of blood loss as well as timing and type of intervention. When the decision was made for removal, the intrauterine balloon was deflated over the course of five to ten minutes and the patient monitored on the labor and delivery unit for the subsequent 2-3 hours to ensure clinical stability of the PPH.
ACCEPTED MANUSCRIPT Einerson
7
Demographic data, labor and delivery data, and postpartum data were gathered from the electronic medical record. Demographic and baseline clinical data collected included maternal age, self-reported race and ethnicity, and delivery admission body mass index.
RI PT
Obstetric data abstraction included gestational age at delivery, gravidity and parity, multiple gestation, the presence of placenta previa, polyhydramnios, preeclampsia
requiring use of magnesium sulfate infusion, chorioamnionitis, mode of delivery, estimated
SC
blood loss (EBL) before balloon placement, the use of uterotonic agents, and antibiotic use. Chorioamnionitis was defined as a minimum temperature of 100.4 degrees Fahrenheit with
M AN U
clinical signs of intrauterine infection (fetal or maternal tachycardia, purulent discharge, uterine tenderness, or leukocytosis). All cases of chorioamnionitis received broadspectrum antibiotics with initiation during the intrapartum period. Patients undergoing cesarean delivery also received preoperative antibiotics for surgical prophylaxis; and in
TE D
those who had chorioamnionitis and underwent cesarean broad spectrum antibiotics were continued for 24 hours postpartum. Additionally, night time placement (between 6pm and 7am) and day time removal of IUBT were included to measure the influence of provider
EP
preference for removal of IUBT during the day.
AC C
The PPH outcomes evaluated included nadir hemoglobin ≤6 mg/dL, transfusion of any blood products, transfusion of 4 or more units of packed red blood cells (PRBC), EBL after balloon placement, uterine artery embolization (UAE), hysterectomy performed after IUBT removal, and maternal intensive care unit (ICU) admission. Transfusion was defined as transfusion of any blood product, including red blood cells, platelets, fresh frozen plasma, or cryoprecipitate. EBL after IUBT placement was estimated clinically by a combination of blood-soaked sponge count, weighed disposable pads and drapes, and output from the
ACCEPTED MANUSCRIPT Einerson
8
IUBT device itself. Clinical staff members who participate in the care of patients with PPH are trained in the estimation of obstetric blood loss through mandatory ongoing EBL
RI PT
training at the study institution. Secondary outcomes included postpartum fever (defined as a temperature greater than or equal to 100.4 degrees Fahrenheit) and postpartum length of stay. Postpartum length of
SC
stay was calculated forward from the time of delivery until discharge.
National guidelines for IUBT duration are lacking, so based on an observed bimodal
M AN U
distribution of timing of IUBT duration in our institution (one peak leading up to 12 hours, the other peak leading up to 24 hours), we compared women who had IUBT duration of 212 hours to those who had an IUBT duration >12 hours (see Figure). The groups were compared using the Wilcoxon Rank-Sum test, Chi-square test, and Fischer’s exact test, as
TE D
appropriate. Tests of hypothesis were two-tailed with a predefined alpha of 0.05. Logistic regression models that included statistically significant differences (p<0.05) in bivariable analyses were performed to determine the effect of confounding on the results.
EP
We used Stata Statistical Software, release 14 (StataCorp. 2015. College Station, TX). The
AC C
Northwestern University Research Institutional Review Board (IRB) approved the study.
Results
During the study period, 93,825 deliveries occurred, of which 5,731 (6.1%) resulted in PPH. An intrauterine balloon was placed in 314 (5.5%) of women with PPH. Of these, 40 (13%) were excluded due to expulsion in the first two hours or excessive ongoing hemorrhage
ACCEPTED MANUSCRIPT Einerson
9
requiring immediate alternative interventions prior to planned balloon removal, leaving 274 patients for analysis. There were 206 (75.2%) women who had an IUBT duration of > 12 hours and 68 (24.8%) women with an IUBT duration of 2-12 hours. The IUBT duration
RI PT
range was 2 hours to 51 hours, with most durations between 10 and 24 hours, as seen in the histogram (Figure).
SC
Women with IUBT duration > 12 hours were more likely to have undergone cesarean
section (56.3% versus 35.3%, p=0.003) and less likely to have had the intrauterine balloon
M AN U
placed during the night time (between 6pm and 7am) (36.9% versus 75.0%, p<0.001). Other baseline characteristics and markers of management leading up to the hemorrhage were similar between the groups (Table 1).
The PPH outcomes in women with IUBT are shown in Table 2. PPH outcomes including
TE D
transfusion, transfusion of 4 or more units of PRBC, nadir hemoglobin ≤ 6.0 mg/dL, and the EBL after balloon placement were similar between groups. The median EBL after IUBT placement did not differ between groups (190 versus 143 mL in the >12 hour and 2-12
EP
hour group respectively, p=0.116). Similarly, the need for adjuvant UAE or hysterectomy did not differ between the groups.
AC C
Postpartum length of stay was longer in the IUBT >12 hour group compared to the 2-12 hour group (3.7 days versus 3.1 days, P<0.001). Furthermore, women undergoing IUBT for >12 hours were almost twice as likely to have postpartum fever ≥100.4 Fahrenheit (26.7% versus 14.7%, P=0.044). After adjusting for variables that differed statistically between groups, including cesarean delivery and night time placement, there remained no association between IUBT duration
ACCEPTED MANUSCRIPT Einerson
10
and PPH-related outcomes. In addition, the association between IUBT >12 hours and postpartum length of stay was no longer significant (adjusted OR 1.29, 95% CI 0.81-2.06). However, the association between IUBT >12 hours and postpartum fever remained
RI PT
(adjusted OR 2.33, 95% CI 1.07-5.11).
To further test the association between IUBT>12 hours and postpartum fever, we
SC
performed a post hoc multivariable analysis that included chorioamnionitis as an
independent variable. Although the incidence of chorioamnionitis was not significantly
M AN U
different between the cohorts, the incidence was higher in the IUBT>12 hour cohort and this clinically meaningful difference could confound the observed result for postpartum fever. Its inclusion in the multivariable model diminished the association between IUBT >12 hours and postpartum fever to the point of being not statistically significant (adjusted
Comment
TE D
OR 2.04, 95% CI 0.92-4.53).
EP
This study did not demonstrate a benefit in PPH control associated with leaving an IUBT in
AC C
place for more than 12 hours. Furthermore, IUBT duration of more than 12 hours may be associated with an increase in febrile morbidity, an association that persisted in analyses controlling for baseline differences in cesarean delivery and night time placement, but that did not persist in post hoc analyses controlling for intrapartum chorioamnionitis. Our results suggest that IUBT duration of 12 hours or less may be sufficient for treatment of PPH.
ACCEPTED MANUSCRIPT Einerson
11
We were unable to find other studies comparing PPH outcomes by IUBT duration. Experience in treating PPH with the Bakri Postpartum Balloon was first reported in 2001, when Bakri and colleagues successfully treated four patients with hemorrhage due to low
RI PT
lying placenta or previa and one patient with hemorrhage due to a cervical ectopic
pregnancy.7 Since then, the indications and availability for use have expanded, due in part to international guidelines encouraging the use of intrauterine tamponade.1,8,9 In a
SC
systematic review of twelve observational studies published between 2001 and 2013, with a total of 273 cases of IUBT used in a wide variety of clinical circumstances, Wright and
M AN U
colleagues found the rate of transfusion was 43%.10 By way of comparison, our observed rate of transfusion was 59%. Duration of IUBT was not reported in this previous systematic review and has received limited attention in the obstetric literature. One case series of 5 patients noted a median duration of 11 hours (range 10-24 hours) but this series did not
TE D
describe outcomes stratified by duration of IUBT.11 Another case series of 109 patients from Hong Kong reported a mean ± standard deviation duration of 22 hours ± 3 hours, but
EP
again they did not examine the relationship between IUBT duration and clinical outcomes.3 Timing of balloon placement in relation to work day hours appears to have played a role in
AC C
the clinicians’ decisions about timing of balloon removal. This is evidenced by the finding that night time placement was much more common in the 2-12 hour group than in the >12 hour group, which likely contributed to the finding that a majority of patients had the balloon removed during the day. We included night time placement in our multivariable regression model to account for this important difference in practice.
ACCEPTED MANUSCRIPT Einerson
12
We acknowledge that our study has limitations. First, the study was performed at a single academic institution with a high obstetric volume, a PPH protocol, and experience with IUBT. Thus external generalizability may be limited. Second, the study is predisposed to
RI PT
confounding and selection bias inherent to its retrospective observational study design. By performing multivariable analyses, we attempted to account for measured differences in the groups, but it is possible that unmeasured differences (such as length of labor, duration
SC
of rupture of membranes, indications for cesarean, or timing of transfusion in relation to hemorrhage onset) in the study groups influenced our final results. Finally, while this is the
M AN U
largest cohort of IUBT to date, it lacks the power to detect meaningful differences in rare but serious outcomes like hysterectomy and ICU admission.
In summary, we found no association between IUBT duration greater than 12 hours and PPH outcomes after adjusting for baseline differences in the groups. Until more definitive
TE D
data from prospective trials are available, it is reasonable to consider removal of an intrauterine balloon by 12 hours after its initial placement if ongoing hemorrhage has
AC C
EP
abated.
ACCEPTED MANUSCRIPT Einerson
13
REFERENCES WHO Recommendations for the Prevention and Treatment of Postpartum Haemorrhage. Geneva: World Health Organization; 2012. 2.
RI PT
1.
Dildy GA, Belfort MA, Adair CD, et al. Initial experience with a dual-balloon catheter
SC
for the management of postpartum hemorrhage. Am J Obstet Gynecol
3.
M AN U
2013;210:136.e1-e6.
Yu G, Wang Z, Zhang J, et al. Efficacy and safety of intrauterine Bakri balloon tamponade in the treatment of postpartum hemorrhage: a multicenter analysis of 109 cases. Chinese J Obstet Gynecol 2014;49:670-675.
Kaya B, Tuten A, Daglar K, et al. Balloon tamponade for the management of
TE D
4.
postpartum uterine hemorrhage. J Perinat Med 2014;42:745-753. 5.
Cho Y, Rizvi C, Uppal T, Condous G. Ultrasonographic visualization of balloon
EP
placement for uterine tamponade in massive primary postpartum hemorrhage.
6.
AC C
Ultrasound Obstet Gynecol 2008;32:711-713. Georgiou C. Intraluminal pressure readings during the establishment of a positive “tamponade test” in the management of postpartum haemorrhage. BJOG 2010;117:295-303.
7.
Bakri YN, Amri A, Abdul Jabbar F. Tamponade-balloon for obstetrical bleeding. Int J Gynecol Obstet 2001;74:139-142.
ACCEPTED MANUSCRIPT Einerson 8.
14
Postpartum Hemorrhage. ACOG Practice Bulletin No. 76. Obstet Gynecol 2006;108:1039-1047. Leduc D, Senikas V, Lalonde AB. Active management of the third stage of labour:
RI PT
9.
prevention and treatment of postpartum hemorrhage. J Obs Gynaecol Can 2009;31:980-993.
Wright CE, Chauhan SP, Abuhamad AZ. Bakri Balloon in the Management of
SC
10.
Postpartum Hemorrhage: A Review. Am J Perinatol 2014;31:957-964.
M AN U
Nelson WL, O’Brien JM. The uterine sandwich for persistent uterine atony: combining the B-Lynch compression suture and an intrauterine Bakri balloon. Am J Obstet
EP
TE D
Gynecol 2007;196:2006-2007.
AC C
11.
ACCEPTED MANUSCRIPT Einerson
15
IUBT >12 hours (n=206) 21 (17-24) 33 (29-37) 101 (49.0) 25 (12.1) 17 (8.3) 10 (4.8) 53 (25.7) 30.3 (27.6-35.2) 38.8 (37.0-39.6) 110 (53.4) 46 (22.3) 10 (4.9) 1 (0.5) 24 (11.7) 30 (14.6) 21 (10.2) 116 (56.3)
p value <0.001 0.720 0.711
AC C
Zero One Two Three
EP
TE D
M AN U
SC
Table 1. Demographic and obstetric characteristics IUBT 2-12 hours Characteristic (n=68) Total IUBT duration, median hours 11 (8-12) Age 33 (29-36) Race / ethnicity Non-Hispanic white 40 (58.8) Non-Hispanic black 7 (10.3) Asian 5 (7.4) Hispanic 2 (2.9) Other 14 (20.6) 2 Body mass index (kg/m ) 30.1 (26.8-35.8) Gestational age (weeks) 39.1 (37.3-40.1) Nulliparous 30 (44.1) Multiple gestation 13 (19.1) Placenta previa 2 (2.9) Polyhydramnios 1 (1.5) Preeclampsia 12 (17.7) Magnesium sulfate infusion* 13 (19.2) Chorioamnionitis 3 (4.4) Cesarean delivery 24 (35.3) Uterotonic agents used (total number)
RI PT
TABLES
Four
0 (0) 3 (3.4) 4 (5.9) 20 (29.4)
6 (2.9) 8 (3.9) 17 (8.3) 69 (33.5)
41 (60.3)
106 (51.5)
EBL before IUBT placement (mL) 2000 (1317-2373) 2000 (1555-2520) Night time placement (6pm – 7am) 51 (75.0) 76 (36.9) Prophylactic antibiotic use 61 (89.7) 196 (95.2) Day time removal (7am – 6pm) 50 (73.5) 165 (80.1) IUBT, intrauterine balloon tamponade; EBL, estimated blood loss Data are listed as n (%) or median (interquartile range) *For either fetal neuroprotection or seizure prophylaxis
0.440 0.374 0.184 0.576 0.504 0.408 0.204 0.371 0.144 0.003 0.494
0.173 <0.001 0.107 0.253
ACCEPTED MANUSCRIPT Einerson
16
AC C
EP
TE D
M AN U
SC
RI PT
Table 2. Postpartum hemorrhage outcomes in women with intrauterine balloon tamponade IUBT 2-12 IUBT >12 Odds Ratio Adjusted hours hours Odds Outcome (n=68) (n=206) P Ratio* Nadir hemoglobin ≤ 6.0 12 (17.7) 39 (18.9) 0.935 1.20 1.06 mg/dL [0.43-3.38] [0.49-2.29] Transfusion of any blood 35 (51.5) 128 (62.1) 0.120 1.55 1.55 product(s) [0.89-2.69] [0.85-2.80] Transfusion of 4 or more 10 (14.7) 35 (17.0) 0.659 1.19 1.28 units PRBC [0.55-2.55] [0.56-2.91] Estimated blood loss 39 (57.4) 136 (66.0) 0.197 1.44 1.29 after balloon placement [0.82-2.53] [0.71-2.35] >200 mL Uterine artery 11 (16.2) 31 (15.1) 0.823 0.91 0.76 embolization [0.43-1.94] [0.34-1.72] Hysterectomy 1 (1.5) 0 (0) 0.248 --Maternal ICU admission 5 (7.4) 18 (8.7) 0.721 1.21 1.30 [0.43-3.38] [0.44-3.88] IUBT, intrauterine balloon tamponade; PRBC, packed red blood cells; mg/dL, milligrams per deciliter; mL, milliliter; ICU, intensive care unit. Data are listed as n (%) or median (interquartile range). *Model adjusted for baseline differences in cesarean section and night time IUBT placement.
ACCEPTED MANUSCRIPT Einerson
17
FIGURE
AC C
EP
TE D
M AN U
SC
RI PT
Figure. Histogram of intrauterine balloon tamponade duration in two-hour time periods.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0002-9378(16)30976-0
DOI:
10.1016/j.ajog.2016.10.040
Reference:
YMOB 11370
To appear in:
American Journal of Obstetrics and Gynecology
Received Date: 15 September 2016 Revised Date:
24 October 2016
Accepted Date: 28 October 2016
Please cite this article as: Einerson BD, Son M, Schneider P, Fields I, Miller ES, The association between intrauterine balloon tamponade duration and postpartum hemorrhage outcomes, American Journal of Obstetrics and Gynecology (2016), doi: 10.1016/j.ajog.2016.10.040. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Einerson
1
Title: The association between intrauterine balloon tamponade duration and postpartum hemorrhage outcomes.
RI PT
Authors: Brett D EINERSON MD MPH1; Moeun SON MD2; Patrick SCHNEIDER, MD3; Ian FIELDS, MD4; Emily S. MILLER, MD MPH2
Affiliations 1Department
of Obstetrics & Gynecology, Division of Maternal-Fetal Medicine, University of
M AN U
Utah Health Sciences Center, Salt Lake City, Utah 2Department
SC
City: Chicago, Illinois
of Obstetrics & Gynecology, Division of Maternal-Fetal Medicine,
Northwestern Feinberg School of Medicine, Chicago, Illinois 3Department
of Obstetrics & Gynecology, Division of Maternal-Fetal Medicine, University of
4Department
TE D
Chicago, Chicago, Illinois
of Obstetrics & Gynecology, Northwestern Feinberg School of Medicine,
Chicago, Illinois
EP
Corresponding author: Brett D. Einerson, 50 N Medical Drive, Salt Lake City, UT 84132;
AC C
Phone 801-585-5156; Email: [email protected] Disclosures: The authors report no conflict of interest. Presented: The 63rd Annual Scientific Meeting of the Society for Reproductive Investigation (SRI), March 16-19, 2016 in Montreal, Quebec, Canada.
Word Count: Abstract 386, Main text 2,006
ACCEPTED MANUSCRIPT Einerson
2
Condensation: Postpartum hemorrhage-related outcomes were similar in those with intrauterine balloon tamponade duration beyond 12 hours compared to those with
AC C
EP
TE D
M AN U
SC
Short title: Optimal intrauterine balloon tamponade duration
RI PT
removal at or before 12 hours.
ACCEPTED MANUSCRIPT Einerson
3
ABSTRACT Background: Intrauterine balloon tamponade (IUBT) is an effective treatment for
management of postpartum hemorrhage (PPH) is unclear.
RI PT
postpartum hemorrhage when first line treatments fail. The optimal duration of IUBT for
Objective: To determine whether IUBT removal >12 hours duration is associated with
SC
PPH-related clinical outcomes.
M AN U
Study Design: A retrospective cohort study of women with PPH from 2007-2014 who underwent IUBT. We excluded failures of IUBT (IUBT expulsion with duration <2 hours or if hysterectomy was required prior to planned IUBT). Patients who underwent IUBT for 212 hours were compared to those who underwent IUBT for >12 hours. Examined PPHrelated clinical outcomes included estimated blood loss (EBL) after IUBT placement, blood
TE D
product transfusion, use of adjuvant measures to control PPH after IUBT (either uterine artery embolization (UAE) or hysterectomy), and maternal intensive care unit (ICU) admission. Secondary outcomes examined included postpartum fever and hospital length
EP
of stay (LOS). Multivariable logistic regression models were used to control for
AC C
confounding variables.
Results: Of 274 eligible women, 206 (75%) underwent IUBT for >12 hours and 68 (25%) underwent IUBT for 2-12 hours. The median EBL after IUBT placement (190 versus 143 mL, p=0.116), as well as the frequencies of blood product transfusion (62.1% versus 51.5%, p=0.120), transfusion of 4 or more units of packed red blood cells (17.0% versus 14.7%, p=0.659), UAE (15.1% versus 16.2%, p=0.823), hysterectomy (0.0% versus 1.5%, p=0.248), and ICU admission (8.7% versus 7.4%, p=0.721) were not statistically different between
ACCEPTED MANUSCRIPT Einerson
4
the groups and this lack of association persisted in multivariable regressions. IUBT duration >12 hours was associated with a higher frequency of postpartum fever (27% versus 15%, p=0.047) and a longer mean hospital LOS (3.7 versus 3.1 days, p=0.002). After
RI PT
adjusting for variables that differed statistically between groups, the difference in length of stay associated with IUBT duration was no longer present, but the association between IUBT duration >12 hours and postpartum fever persisted (OR 2.33, 95% CI 1.07-5.11).
SC
Including chorioamnionitis as an independent variable in a post hoc multivariable analysis diminished the association between IUBT >12 hours and postpartum fever (adjusted OR
M AN U
2.04, 95% CI 0.92-4.53).
Conclusion: There are no significant differences in PPH-related outcomes associated with IUBT duration >12 hours compared to removal 2-12 hours. If ongoing hemorrhage has
initial placement.
TE D
abated, it is reasonable to consider removal of an intrauterine balloon by 12 hours after its
AC C
EP
Key words: intrauterine balloon tamponade, postpartum hemorrhage
ACCEPTED MANUSCRIPT Einerson
5
TEXT Introduction
RI PT
Aggressive treatment is essential in the management of postpartum hemorrhage (PPH), the world’s leading cause of maternal mortality.1 When first line treatments such as uterotonic agents and bimanual uterine massage fail, intrauterine balloon tamponade (IUBT) has been
SC
shown to be effective, with reported bleeding cessation rates of 55-94%.2–4 However,
previous studies have been limited to case series and small observational studies, and thus
M AN U
many questions remain about the appropriate management of the intrauterine balloon devices once they have been successfully placed and hemorrhage has abated. Specifically, it remains unknown what the optimal duration of IUBT is with regard to PPH-related outcomes. IUBT is proposed to stop obstetric hemorrhage by direct tamponade of the
TE D
vascular placental bed, encouragement of uterine contractility, and/or direct hydrostatic pressure on the uterine arteries.5,6 Whether these actions are successful and sustained after a set duration of IUBT has not been previously studied.
EP
We sought to determine if IUBT duration is associated with PPH-related outcomes. We hypothesized that a longer duration of IUBT would be associated with increased
AC C
effectiveness in the management of PPH.
Materials and Methods
This is a retrospective cohort study of all women who underwent IUBT at Northwestern Memorial Hospital in Chicago, Illinois from January 2007 to October 2014. The estimated
ACCEPTED MANUSCRIPT Einerson
6
blood loss (EBL) is routinely recorded in the delivery form of our electronic medical record. A query of cases with an EBL of 500 mL or more after a vaginal delivery or 1,000 mL or more after a cesarean yielded all cases of PPH. Medical records for each case of PPH were
RI PT
then reviewed to identify women undergoing IUBT. At the study institution, IUBT was
recommended by protocol as a second-line option for the management of PPH refractory to uterotonic agents and bimanual uterine massage. The protocol indications for use of IUBT
SC
included placental site bleeding and uterine atony. The IUBT device used was the Bakri Postpartum Balloon (Cook Medical Inc, Bloomington, Indiana). We excluded women whose
M AN U
balloon was removed within the first two hours of placement due to expulsion (n=33) as well as women who underwent hysterectomy prior to planned IUBT removal for refractory hemorrhage (n=7) as these circumstances indicated a failure of IUBT rather than an outcome related to the duration of IUBT placement. We identified no cases in which IUBT
TE D
was discontinued intentionally before 2 hours due to perceived success in controlling PPH. After placement of IUBT, patients underwent a period of observation on the labor and
EP
delivery unit with close monitoring of vital signs and ongoing estimation of blood loss. The duration of IUBT was at the discretion of the clinical provider and was not dictated by
AC C
protocol; thus, this practice pattern varied at the study institution. A designated nurse was tasked with recording the events of the PPH episode, including ongoing estimation of blood loss as well as timing and type of intervention. When the decision was made for removal, the intrauterine balloon was deflated over the course of five to ten minutes and the patient monitored on the labor and delivery unit for the subsequent 2-3 hours to ensure clinical stability of the PPH.
ACCEPTED MANUSCRIPT Einerson
7
Demographic data, labor and delivery data, and postpartum data were gathered from the electronic medical record. Demographic and baseline clinical data collected included maternal age, self-reported race and ethnicity, and delivery admission body mass index.
RI PT
Obstetric data abstraction included gestational age at delivery, gravidity and parity, multiple gestation, the presence of placenta previa, polyhydramnios, preeclampsia
requiring use of magnesium sulfate infusion, chorioamnionitis, mode of delivery, estimated
SC
blood loss (EBL) before balloon placement, the use of uterotonic agents, and antibiotic use. Chorioamnionitis was defined as a minimum temperature of 100.4 degrees Fahrenheit with
M AN U
clinical signs of intrauterine infection (fetal or maternal tachycardia, purulent discharge, uterine tenderness, or leukocytosis). All cases of chorioamnionitis received broadspectrum antibiotics with initiation during the intrapartum period. Patients undergoing cesarean delivery also received preoperative antibiotics for surgical prophylaxis; and in
TE D
those who had chorioamnionitis and underwent cesarean broad spectrum antibiotics were continued for 24 hours postpartum. Additionally, night time placement (between 6pm and 7am) and day time removal of IUBT were included to measure the influence of provider
EP
preference for removal of IUBT during the day.
AC C
The PPH outcomes evaluated included nadir hemoglobin ≤6 mg/dL, transfusion of any blood products, transfusion of 4 or more units of packed red blood cells (PRBC), EBL after balloon placement, uterine artery embolization (UAE), hysterectomy performed after IUBT removal, and maternal intensive care unit (ICU) admission. Transfusion was defined as transfusion of any blood product, including red blood cells, platelets, fresh frozen plasma, or cryoprecipitate. EBL after IUBT placement was estimated clinically by a combination of blood-soaked sponge count, weighed disposable pads and drapes, and output from the
ACCEPTED MANUSCRIPT Einerson
8
IUBT device itself. Clinical staff members who participate in the care of patients with PPH are trained in the estimation of obstetric blood loss through mandatory ongoing EBL
RI PT
training at the study institution. Secondary outcomes included postpartum fever (defined as a temperature greater than or equal to 100.4 degrees Fahrenheit) and postpartum length of stay. Postpartum length of
SC
stay was calculated forward from the time of delivery until discharge.
National guidelines for IUBT duration are lacking, so based on an observed bimodal
M AN U
distribution of timing of IUBT duration in our institution (one peak leading up to 12 hours, the other peak leading up to 24 hours), we compared women who had IUBT duration of 212 hours to those who had an IUBT duration >12 hours (see Figure). The groups were compared using the Wilcoxon Rank-Sum test, Chi-square test, and Fischer’s exact test, as
TE D
appropriate. Tests of hypothesis were two-tailed with a predefined alpha of 0.05. Logistic regression models that included statistically significant differences (p<0.05) in bivariable analyses were performed to determine the effect of confounding on the results.
EP
We used Stata Statistical Software, release 14 (StataCorp. 2015. College Station, TX). The
AC C
Northwestern University Research Institutional Review Board (IRB) approved the study.
Results
During the study period, 93,825 deliveries occurred, of which 5,731 (6.1%) resulted in PPH. An intrauterine balloon was placed in 314 (5.5%) of women with PPH. Of these, 40 (13%) were excluded due to expulsion in the first two hours or excessive ongoing hemorrhage
ACCEPTED MANUSCRIPT Einerson
9
requiring immediate alternative interventions prior to planned balloon removal, leaving 274 patients for analysis. There were 206 (75.2%) women who had an IUBT duration of > 12 hours and 68 (24.8%) women with an IUBT duration of 2-12 hours. The IUBT duration
RI PT
range was 2 hours to 51 hours, with most durations between 10 and 24 hours, as seen in the histogram (Figure).
SC
Women with IUBT duration > 12 hours were more likely to have undergone cesarean
section (56.3% versus 35.3%, p=0.003) and less likely to have had the intrauterine balloon
M AN U
placed during the night time (between 6pm and 7am) (36.9% versus 75.0%, p<0.001). Other baseline characteristics and markers of management leading up to the hemorrhage were similar between the groups (Table 1).
The PPH outcomes in women with IUBT are shown in Table 2. PPH outcomes including
TE D
transfusion, transfusion of 4 or more units of PRBC, nadir hemoglobin ≤ 6.0 mg/dL, and the EBL after balloon placement were similar between groups. The median EBL after IUBT placement did not differ between groups (190 versus 143 mL in the >12 hour and 2-12
EP
hour group respectively, p=0.116). Similarly, the need for adjuvant UAE or hysterectomy did not differ between the groups.
AC C
Postpartum length of stay was longer in the IUBT >12 hour group compared to the 2-12 hour group (3.7 days versus 3.1 days, P<0.001). Furthermore, women undergoing IUBT for >12 hours were almost twice as likely to have postpartum fever ≥100.4 Fahrenheit (26.7% versus 14.7%, P=0.044). After adjusting for variables that differed statistically between groups, including cesarean delivery and night time placement, there remained no association between IUBT duration
ACCEPTED MANUSCRIPT Einerson
10
and PPH-related outcomes. In addition, the association between IUBT >12 hours and postpartum length of stay was no longer significant (adjusted OR 1.29, 95% CI 0.81-2.06). However, the association between IUBT >12 hours and postpartum fever remained
RI PT
(adjusted OR 2.33, 95% CI 1.07-5.11).
To further test the association between IUBT>12 hours and postpartum fever, we
SC
performed a post hoc multivariable analysis that included chorioamnionitis as an
independent variable. Although the incidence of chorioamnionitis was not significantly
M AN U
different between the cohorts, the incidence was higher in the IUBT>12 hour cohort and this clinically meaningful difference could confound the observed result for postpartum fever. Its inclusion in the multivariable model diminished the association between IUBT >12 hours and postpartum fever to the point of being not statistically significant (adjusted
Comment
TE D
OR 2.04, 95% CI 0.92-4.53).
EP
This study did not demonstrate a benefit in PPH control associated with leaving an IUBT in
AC C
place for more than 12 hours. Furthermore, IUBT duration of more than 12 hours may be associated with an increase in febrile morbidity, an association that persisted in analyses controlling for baseline differences in cesarean delivery and night time placement, but that did not persist in post hoc analyses controlling for intrapartum chorioamnionitis. Our results suggest that IUBT duration of 12 hours or less may be sufficient for treatment of PPH.
ACCEPTED MANUSCRIPT Einerson
11
We were unable to find other studies comparing PPH outcomes by IUBT duration. Experience in treating PPH with the Bakri Postpartum Balloon was first reported in 2001, when Bakri and colleagues successfully treated four patients with hemorrhage due to low
RI PT
lying placenta or previa and one patient with hemorrhage due to a cervical ectopic
pregnancy.7 Since then, the indications and availability for use have expanded, due in part to international guidelines encouraging the use of intrauterine tamponade.1,8,9 In a
SC
systematic review of twelve observational studies published between 2001 and 2013, with a total of 273 cases of IUBT used in a wide variety of clinical circumstances, Wright and
M AN U
colleagues found the rate of transfusion was 43%.10 By way of comparison, our observed rate of transfusion was 59%. Duration of IUBT was not reported in this previous systematic review and has received limited attention in the obstetric literature. One case series of 5 patients noted a median duration of 11 hours (range 10-24 hours) but this series did not
TE D
describe outcomes stratified by duration of IUBT.11 Another case series of 109 patients from Hong Kong reported a mean ± standard deviation duration of 22 hours ± 3 hours, but
EP
again they did not examine the relationship between IUBT duration and clinical outcomes.3 Timing of balloon placement in relation to work day hours appears to have played a role in
AC C
the clinicians’ decisions about timing of balloon removal. This is evidenced by the finding that night time placement was much more common in the 2-12 hour group than in the >12 hour group, which likely contributed to the finding that a majority of patients had the balloon removed during the day. We included night time placement in our multivariable regression model to account for this important difference in practice.
ACCEPTED MANUSCRIPT Einerson
12
We acknowledge that our study has limitations. First, the study was performed at a single academic institution with a high obstetric volume, a PPH protocol, and experience with IUBT. Thus external generalizability may be limited. Second, the study is predisposed to
RI PT
confounding and selection bias inherent to its retrospective observational study design. By performing multivariable analyses, we attempted to account for measured differences in the groups, but it is possible that unmeasured differences (such as length of labor, duration
SC
of rupture of membranes, indications for cesarean, or timing of transfusion in relation to hemorrhage onset) in the study groups influenced our final results. Finally, while this is the
M AN U
largest cohort of IUBT to date, it lacks the power to detect meaningful differences in rare but serious outcomes like hysterectomy and ICU admission.
In summary, we found no association between IUBT duration greater than 12 hours and PPH outcomes after adjusting for baseline differences in the groups. Until more definitive
TE D
data from prospective trials are available, it is reasonable to consider removal of an intrauterine balloon by 12 hours after its initial placement if ongoing hemorrhage has
AC C
EP
abated.
ACCEPTED MANUSCRIPT Einerson
13
REFERENCES WHO Recommendations for the Prevention and Treatment of Postpartum Haemorrhage. Geneva: World Health Organization; 2012. 2.
RI PT
1.
Dildy GA, Belfort MA, Adair CD, et al. Initial experience with a dual-balloon catheter
SC
for the management of postpartum hemorrhage. Am J Obstet Gynecol
3.
M AN U
2013;210:136.e1-e6.
Yu G, Wang Z, Zhang J, et al. Efficacy and safety of intrauterine Bakri balloon tamponade in the treatment of postpartum hemorrhage: a multicenter analysis of 109 cases. Chinese J Obstet Gynecol 2014;49:670-675.
Kaya B, Tuten A, Daglar K, et al. Balloon tamponade for the management of
TE D
4.
postpartum uterine hemorrhage. J Perinat Med 2014;42:745-753. 5.
Cho Y, Rizvi C, Uppal T, Condous G. Ultrasonographic visualization of balloon
EP
placement for uterine tamponade in massive primary postpartum hemorrhage.
6.
AC C
Ultrasound Obstet Gynecol 2008;32:711-713. Georgiou C. Intraluminal pressure readings during the establishment of a positive “tamponade test” in the management of postpartum haemorrhage. BJOG 2010;117:295-303.
7.
Bakri YN, Amri A, Abdul Jabbar F. Tamponade-balloon for obstetrical bleeding. Int J Gynecol Obstet 2001;74:139-142.
ACCEPTED MANUSCRIPT Einerson 8.
14
Postpartum Hemorrhage. ACOG Practice Bulletin No. 76. Obstet Gynecol 2006;108:1039-1047. Leduc D, Senikas V, Lalonde AB. Active management of the third stage of labour:
RI PT
9.
prevention and treatment of postpartum hemorrhage. J Obs Gynaecol Can 2009;31:980-993.
Wright CE, Chauhan SP, Abuhamad AZ. Bakri Balloon in the Management of
SC
10.
Postpartum Hemorrhage: A Review. Am J Perinatol 2014;31:957-964.
M AN U
Nelson WL, O’Brien JM. The uterine sandwich for persistent uterine atony: combining the B-Lynch compression suture and an intrauterine Bakri balloon. Am J Obstet
EP
TE D
Gynecol 2007;196:2006-2007.
AC C
11.
ACCEPTED MANUSCRIPT Einerson
15
IUBT >12 hours (n=206) 21 (17-24) 33 (29-37) 101 (49.0) 25 (12.1) 17 (8.3) 10 (4.8) 53 (25.7) 30.3 (27.6-35.2) 38.8 (37.0-39.6) 110 (53.4) 46 (22.3) 10 (4.9) 1 (0.5) 24 (11.7) 30 (14.6) 21 (10.2) 116 (56.3)
p value <0.001 0.720 0.711
AC C
Zero One Two Three
EP
TE D
M AN U
SC
Table 1. Demographic and obstetric characteristics IUBT 2-12 hours Characteristic (n=68) Total IUBT duration, median hours 11 (8-12) Age 33 (29-36) Race / ethnicity Non-Hispanic white 40 (58.8) Non-Hispanic black 7 (10.3) Asian 5 (7.4) Hispanic 2 (2.9) Other 14 (20.6) 2 Body mass index (kg/m ) 30.1 (26.8-35.8) Gestational age (weeks) 39.1 (37.3-40.1) Nulliparous 30 (44.1) Multiple gestation 13 (19.1) Placenta previa 2 (2.9) Polyhydramnios 1 (1.5) Preeclampsia 12 (17.7) Magnesium sulfate infusion* 13 (19.2) Chorioamnionitis 3 (4.4) Cesarean delivery 24 (35.3) Uterotonic agents used (total number)
RI PT
TABLES
Four
0 (0) 3 (3.4) 4 (5.9) 20 (29.4)
6 (2.9) 8 (3.9) 17 (8.3) 69 (33.5)
41 (60.3)
106 (51.5)
EBL before IUBT placement (mL) 2000 (1317-2373) 2000 (1555-2520) Night time placement (6pm – 7am) 51 (75.0) 76 (36.9) Prophylactic antibiotic use 61 (89.7) 196 (95.2) Day time removal (7am – 6pm) 50 (73.5) 165 (80.1) IUBT, intrauterine balloon tamponade; EBL, estimated blood loss Data are listed as n (%) or median (interquartile range) *For either fetal neuroprotection or seizure prophylaxis
0.440 0.374 0.184 0.576 0.504 0.408 0.204 0.371 0.144 0.003 0.494
0.173 <0.001 0.107 0.253
ACCEPTED MANUSCRIPT Einerson
16
AC C
EP
TE D
M AN U
SC
RI PT
Table 2. Postpartum hemorrhage outcomes in women with intrauterine balloon tamponade IUBT 2-12 IUBT >12 Odds Ratio Adjusted hours hours Odds Outcome (n=68) (n=206) P Ratio* Nadir hemoglobin ≤ 6.0 12 (17.7) 39 (18.9) 0.935 1.20 1.06 mg/dL [0.43-3.38] [0.49-2.29] Transfusion of any blood 35 (51.5) 128 (62.1) 0.120 1.55 1.55 product(s) [0.89-2.69] [0.85-2.80] Transfusion of 4 or more 10 (14.7) 35 (17.0) 0.659 1.19 1.28 units PRBC [0.55-2.55] [0.56-2.91] Estimated blood loss 39 (57.4) 136 (66.0) 0.197 1.44 1.29 after balloon placement [0.82-2.53] [0.71-2.35] >200 mL Uterine artery 11 (16.2) 31 (15.1) 0.823 0.91 0.76 embolization [0.43-1.94] [0.34-1.72] Hysterectomy 1 (1.5) 0 (0) 0.248 --Maternal ICU admission 5 (7.4) 18 (8.7) 0.721 1.21 1.30 [0.43-3.38] [0.44-3.88] IUBT, intrauterine balloon tamponade; PRBC, packed red blood cells; mg/dL, milligrams per deciliter; mL, milliliter; ICU, intensive care unit. Data are listed as n (%) or median (interquartile range). *Model adjusted for baseline differences in cesarean section and night time IUBT placement.
ACCEPTED MANUSCRIPT Einerson
17
FIGURE
AC C
EP
TE D
M AN U
SC
RI PT
Figure. Histogram of intrauterine balloon tamponade duration in two-hour time periods.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT