Archives of Gynecology and Obstetrics

, Volume 283, Issue 4, pp 723–727

Emergency peripartum hysterectomy

Authors

    • Obstetrics and GynecologyZekai Tahir Burak Women’s Health Education and Research Hospital
  • Sarp Özcan
    • Obstetrics and GynecologyZekai Tahir Burak Women’s Health Education and Research Hospital
  • Şebnem Özyer
    • Obstetrics and GynecologyZekai Tahir Burak Women’s Health Education and Research Hospital
  • Leyla Mollamahmutoğlu
    • Obstetrics and GynecologyZekai Tahir Burak Women’s Health Education and Research Hospital
  • Nuri Danışman
    • Obstetrics and GynecologyZekai Tahir Burak Women’s Health Education and Research Hospital
Materno-fetal Medicine

DOI: 10.1007/s00404-010-1451-z

Cite this article as:
Karayalçın, R., Özcan, S., Özyer, Ş. et al. Arch Gynecol Obstet (2011) 283: 723. doi:10.1007/s00404-010-1451-z

Abstract

Purpose

To determine the incidence, indications, risk factors and complications of peripartum hysterectomy in a tertiary teaching hospital.

Methods

The medical records of 73 patients who had undergone emergency peripartum hysterectomy between 2003 and 2008 were reviewed retrospectively. Maternal characteristics and characteristics of the present pregnancy and delivery, hysterectomy indications, operative complications, postoperative conditions and maternal outcomes were evaluated.

Results

There were 73 emergency peripartum hysterectomies out of 114,720 deliveries, a rate of 0.63 per 1,000 deliveries. Eleven hysterectomies were performed after vaginal delivery (0.12/1,000 vaginal deliveries) and the remaining 62 hysterectomies were performed after cesarean section (2/1,000 cesarean sections). The most common indication for hysterectomy was placenta previa and/or accreta (31 patients, 42.4%), followed by uterine atony (26 patients, 35.6%). In this study, 22 of 29 patients (75.8%) with placenta previa and 12 of 16 patients (75%) with placenta accreta had previously had cesarean sections. Cesarean section is associated with placenta previa and accreta, which are the most common causes of emergency peripartum hysterectomy.

Conclusion

The increase in the cesarean delivery rate is leading to an increase in the rate of abnormal placentation (placenta previa and accreta), which in turn give rise to an increase in the peripartum hysterectomy rate. Cesarean section itself is also a risk factor for emergency peripartum hysterectomy. Therefore, every effort should be made to reduce the cesarean rate by performing this procedure only for valid clinical indications. The risk factors for peripartum hysterectomy should be identified antenatally. The delivery and operation should be performed in appropriate clinical settings by experienced surgeons when risk factors are identified.

Keywords

Emergency peripartum hysterectomyPostpartum hemorrhageCesarean sectionPlacenta previa accreta

Introduction

Postpartum hemorrhage is one of the leading causes of maternal mortality and morbidity. Severe postpartum hemorrhage was reported to occur in 6.7/1,000 deliveries in the UK [1]. A recent report noted that this number has increased from 7 to 17 [2]. The risk of maternal death, which is the most severe complication of hemorrhage, is estimated to be approximately 1 in 100,000 deliveries in developed countries and has been increasing [3, 4]. This risk is as high as 1 in 1,000 deliveries in developing countries [3]. Other maternal complications of postpartum hemorrhage include hypovolemic shock, disseminated intravascular coagulopathy, renal failure, hepatic failure and adult respiratory distress syndrome (ARDS) [3]. Conservative treatment of postpartum hemorrhage includes uterotonics (oxytocin, ergotamine), uterine massage, uterine artery embolization, uterine packing, pelvic vessel ligation, B-Lynch suture, multiple square sutures and recombinant activated factor VII [5]. The incidence of peripartum hysterectomy in the literature is reported as 0.24, 2.3 and 5.09 per 1,000 deliveries by Sakse et al. [6], Bai et al. [7] and Zeteroğlu et al. [8], respectively. The operation itself carries a high maternal mortality and morbidity risk. Thus, it is only performed as a life-saving procedure when all conservative measures fail to achieve homeostasis.

The objectives of this retrospective study are to examine the incidence, risk factors, indications, outcomes and complications of emergency peripartum hysterectomy performed in a tertiary teaching hospital between 2003 and 2008 and to compare the results with other reports in the literature.

Materials and methods

This study was a case series study. Medical and pathological records of the patients who had undergone emergency hysterectomy following vaginal or cesarean delivery due to postpartum hemorrhage between 2003 and 2008 in a tertiary teaching hospital were reviewed retrospectively. Emergency peripartum hysterectomy was defined as a hysterectomy performed in a life-threatening condition of postpartum hemorrhage. All deliveries were performed after 24 weeks of gestation, and the hysterectomy was performed shortly (within hours) after delivery. Cesarean sections were performed by obstetrics and gynecology specialist, whereas the hysterectomies were performed by consultant senior surgeons. Both medical and surgical modalities were used to control the hemorrhage before hysterectomy. Information obtained from the medical records included demographic details, previous obstetric history, details of the current pregnancy and delivery, postpartum hemorrhage, indications for peripartum hysterectomy, outcomes of hysterectomy as intraoperative and postoperative complications, length of hospital stay, amount of blood transfused and neonatal outcomes. The time and the duration of operation were obtained from maternal anesthetic records. The pathology reports were also obtained. Maternal complications, including maternal death and serious hemorrhagic, and neurological, urological, infectious, respiratory, renal and thromboembolic complications were also checked.

Statistical analysis

All statistical analysis was done using the SPSS for Windows release 11.5 packet program. A chi-squared test was used to compare the prevalence of hysterectomy between vaginal and cesarean deliveries. Frequencies and percentages were given as descriptive statistics. Alpha was set at 0.05 for the statistical significance level.

Results

Between 2003 and 2008, there were 84,890 (73.9%) vaginal and 29,830 (26.1%) cesarean deliveries, for a total of 114,720 deliveries. Seventy-three emergency hysterectomies were performed during this time period. Ten of the patients were referred from other centers due to postpartum hemorrhage. Maternal characteristics are shown in Table 1. The mean maternal age was 30.2 ± 5.44 years. There were 13 nulliparous and 60 multiparous women. Twenty-seven (36.9%) of the patients were admitted to the hospital at least once for antenatal care, whereas 46 (63.1%) patients never had antenatal care. Among the referrals, five of them delivered vaginally and the rest by cesarean section. The most common indication for cesarean section was repeat cesarean (Table 2).
Table 1

Maternal characteristics

Characteristic

n = 73

Age (years, mean ± SD)

30.2 ± 5.44

Gravidity (median, range)

3 (1–10)

Parity (median, range)

2 (0–5)

Antenatal follow-up (n, %)

 Present

27 (36.9)

 None

46 (63.1)

Previous uterine surgery (n, %)

 Cesarean

38 (52)

 Dilatation and curettage

7 (9.5)

 Hysteroscopic septum resection

1 (1.3)

Table 2

Characteristics of the current pregnancy and labor

Characteristics

n = 73

Multiple pregnancy (n, %)

 Singleton

68 (93.1)

 Twin

4 (5.4)

 Triplet

1 (1.3)

Mode of delivery (n, %)

 Spontaneous vaginal

9 (12.3)

 Operative delivery (vacuum)

1 (1.3)

 Vaginal birth after cesarean

1 (1.3)

 Planned cesarean

2 (2.7)

 Emergency cesarean

60 (82.1)

Indications for cesarean delivery (n, %)

 Repeat cesarean

17 (23.2)

 Repeat cesarean and placenta previa

14 (19.1)

 Fetal distress

8 (10.9)

 Repeat cesarean and abruptio

4 (5.4)

 Placenta previa

4 (5.4)

 Dystocia

4 (5.4)

 Abruptio

3 (4.1)

 Repeat cesarean and fetal distress

2 (2.7)

 IVF twin pregnancy

2 (2.7)

 Twin pregnancy

2 (2.7)

 IVF triplet pregnancy

1 (1.3)

 Abnormal presentation (transverse lie)

1 (1.3)

To avoid hysterectomy, pharmacological agents and surgical procedures were used to control hemorrhage. All patients received oxytocin and ergotamine derivative. The hypogastric artery was ligated in 17 patients. All patients received blood transfusions, with the median number of units of blood transfused being 8 (range 3–30) units.

Seventy-three emergency hysterectomies were performed due to intractable obstetric hemorrhage that was unresponsive to conservative management, representing an incidence of 0.63 per 1,000 deliveries. Eleven hysterectomies were performed after vaginal delivery (0.12/1,000 vaginal delivery) and the remaining 62 hysterectomies were performed after cesarean section (2/1,000 cesarean section). Thirty-eight patients (52%) previously had at least one cesarean section. One patient had previously a cesarean section, but required a hysterectomy after the vaginal birth of the child in this study. We found a significantly increased incidence of peripartum hysterectomy following cesarean section as compared to the incidence following vaginal delivery (p < 0.001).

Indications for cesarean and postpartum hysterectomy are shown in Table 3. The most common indication for hysterectomy was placenta previa and/or accreta (31 patients, 42.4%), followed by uterine atony (26 patients, 35.6%). In this study, 22 of 29 patients (75.8%) with placenta previa and 12 of 16 patients (75%) with placenta accreta had previously had cesarean sections. This shows that cesarean section is associated with placenta previa and accreta, which are the most common causes of emergency peripartum hysterectomy.
Table 3

Indications for cesarean hysterectomy and postpartum hysterectomy

Indication

Cesarean hysterectomy (n = 62)

Postpartum hysterectomy (n = 11)

p

Abnormal placentation

29 (46.7)

2 (18.1)

0.072

Placenta previa

15 (24.1)

0 (0.0)

 

Placenta accreta

0 (0.0)

2 (18.1)

 

Placenta previa and accreta

14 (22.5)

0 (0.0)

 

Atony

20 (32.2)

6 (54.5)

0.140

Abruptio

7 (11.2)

0 (0.0)

0.307

Rupture

3 (4.8)

3 (27.2)

0.04*

Laceration

3 (4.8)

0 (0.0)

0.608

p < 0.05

The operative complications, postoperative conditions and maternal outcomes are shown in Table 4. There were four cases of intraoperative bladder injury. Two of these patients had previously had three cesarean sections and the present pregnancy was complicated with placenta previa. The other two had two previous cesarean sections. Postoperative complications occurred in 23 patients for a rate of 31.5%. The mean duration of the hysterectomy was 169.73 ± 115.4 min. The median postoperative hospitalization stay was 6 (range 4–20) days. There were three relaparotomies because of persistent intra-abdominal bleeding. The rate of re-exploration following hysterectomy was 4.1%. There were two maternal deaths. Both maternal deaths were complicated with hypertensive disorders of pregnancy. The first was a 31-year-old woman at 32 weeks’ gestation. Her first pregnancy was terminated by cesarean section because of severe preeclampsia 2 years ago. She was performed cesarean section in another center because of placental abruption due to HELLP syndrome. She was referred because of intractable postpartum hemorrhage. The patient was taken to the operating theater and hysterectomy was performed. She was massively transfused. After 3 h, relaparotomy had to be performed because of persistent intra-abdominal bleeding. She was died due to disseminated intravascular coagulopathy and cardiac failure. The second woman, 42-year-old multigravida at 33 weeks’ gestation with a previous cesarean, was operated because of severe preeclampsia superimposed upon chronic hypertension. During the operation hysterectomy had to be performed due to placenta accreta in the old scar. In the postoperative period, her blood pressure remained high unresponsive to antihypertensive treatment. She had respiratory failure with a sudden drop in arterial oxygen saturation requiring ventilation. The patient was died after 2 days in the intensive care unit because of ARDS.
Table 4

Operative complications, postoperative conditions and maternal outcomes of the patients who underwent hysterectomy

Characteristicsa

n = 73

Bladder injury (n, %)

4 (5.4)

DIC (n, %)

9 (12.3)

Renal failure (n, %)

6 (8.2)

Infection (n, %)

5 (6.8)

Respiratory complications (n, %)

3 (4.1)

Wound dehiscence (n, %)

1 (1.3)

Multiorgan failure (n, %)

1 (1.3)

Re-exploration (n, %)

3 (4.1)

Maternal mortality (n, %)

2 (2.7)

Total blood transfusion (units, median, range)

8 (3–30)

Hospital stay (days, median, range)

6 (4–20)

aSome patients had more than one complication

Neonatal outcomes are shown in Table 5. The mean gestational age at delivery was 36.1 ± 5.8 weeks. The mean birth weight of the 79 infants was 2,664 ± 1,074 g. There were 13 intrauterine fetal demise and 7 neonatal deaths giving a perinatal mortality rate of 25.3% and for a neonatal mortality rate of 8.86%. The neonatal deaths were all due to prematurity.
Table 5

Neonatal outcomes

Characteristics

n = 79

Gestational age at delivery (weeks, mean ± SD)

36.1 ± 5.8

Birth weight (g, mean ± SD)

2664 ± 1,074

Apgar

 1st min (median, min–max)

7 (0–9)

 5th min (median, min–max)

9 (0–10)

Perinatal mortality (n, %)

20 (25.3)

Discussion

Despite advances in medicine and surgery, postpartum hemorrhage remains one of the leading causes of maternal mortality and morbidity. The risk factors for postpartum hemorrhage include coagulopathies, uterine atony, retained products of conception, precipitous or prolonged labor, fetal macrosomia or multiparity, maternal obesity, and previous primary postpartum hemorrhage; although a specific risk factor may not be present in many patients [9]. Although surgery cannot always be prevented, recognizing and assessing patients at risk and appropriate management of hemorrhage is important.

Hysterectomy following cesarean section was first described by Porro [10] and was used to prevent maternal mortality due to postpartum hemorrhage. Since then, hysterectomy has been widely used as a life-saving procedure for life-threatening postpartum hemorrhage that cannot be controlled by conservative measures. The indications for an emergency hysterectomy are uterine atony, uterine rupture, abnormal placentation and vessel injury due to the extension of a uterine incision. Selective cesarean hysterectomies have been performed to treat uterine leiomyomas, endometriosis and other gynecologic diseases in the past [7]. However, hysterectomies were not as often performed for this purpose due to their high complication rate.

The reported incidence of peripartum hysterectomy varies from 0.24 to 5.09 per 1,000 deliveries in the literature [6, 8, 11]. Our incidence of 0.63 per 1,000 deliveries for emergency peripartum hysterectomy is in agreement with the recent studies reported in the literature. Zeteroğlu et al. [8] reported the incidence of peripartum hysterectomy in a teaching hospital as 5.09/1,000 deliveries, which is the higher than that of other studies. Zorlu et al. reported the incidence of peripartum hysterectomies in two different time periods. The reported incidence between 1979 and 1985 was 0.40/1,000 deliveries, while during 1985–1989 it was 0.24/1,000 deliveries [12]. The decline in the ratio was attributed to the availability of standard obstetric care. In our study, most of the patients (63.1%) never had antenatal care. If the patients had had routine antenatal follow-up, preventive measures could be taken to reduce emergency peripartum hysterectomy rate.

Traditionally, uterine atony was the most common indication for hysterectomy [13]. Recent studies have indicated that abnormal placentation is replacing uterine atony as the most common indication for emergency peripartum hysterectomy [13]. In 1984, Clark et al. [14] reported that 43.4% of their emergency hysterectomies were done because of uterine atony, while 33.9% were due to placenta previa with accreta. A study from the same institution in 1993 stated that their primary indication was placenta accreta, the problem in 45% of cases, followed by uterine atony, with 20% [15]. Baskett et al. [16] reported that the main indications for hysterectomy were abnormal placentation (50%) and atonic postpartum hemorrhage (32.8%). Our study, in agreement with the abovementioned studies, shows that abnormal placentation (placenta previa and/or accreta) is the most common indication for emergency postpartum hysterectomy followed by uterine atony, causing 42.4 and 35.6% of hysterectomies, respectively.

In recent years, abnormal placentation has become more common due to the greater number of pregnant women who have previously delivered by cesarean section. Kastner et al. [17] analyzed 47 cases from 1991 to 1997 and showed that placenta accreta accounted for 48.9% of the cases, 51.1% of whom had previously had a cesarean delivery. Zelop et al. [18] analyzed adherent placentation, which accounted for 64% of cases, 59.8% of whom had previous delivered babies by cesarean section. In our study, 22 of 29 patients (75.8%) with placenta previa and 12 of 16 patients (75%) with placenta accreta had previously had a cesarean section. This shows that cesarean section is associated with placenta previa and/or accreta, which are the most common causes of emergency peripartum hysterectomy. Given the association between abnormal placentation and cesarean delivery, it has been suggested that the high incidence of emergency peripartum hysterectomy can be attributed to the increasing number of cesarean sections. In addition, the cesarean section itself tends to lead to more blood loss and increases the risk of peripartum hysterectomy. On the other hand, Glaze et al. [9] did not find an increase in the peripartum hysterectomy rate, despite a rise in the rate of cesarean delivery over the same period, from 20.7% in 1999 to 28.4% in 2004. They offered two possible explanations for this; first, their study was limited by the small number of emergency hysterectomy cases, and second, they suggested that the rate of peripartum hysterectomy had stayed constant. However, clinical experience suggests that a scarred uterus is more likely to lead to abnormal placentation during pregnancy [19].

To avoid surgery, direct vessel ligation may be performed, but can be ineffective in controlling severe hemorrhage. In case of uterine atony, subtotal hysterectomy is safer and quicker than total hysterectomy. However, it is not appropriate in cases of bleeding from the lower uterine segment associated with placenta previa and/or accreta [16].

Our results confirm previous observations that emergency hysterectomies are associated with high operative and postoperative complication rates. Postoperative complications occurred in 23 (31.5%) patients. The maternal mortality rate in our study was 2.7%, which is comparable with the maternal mortality rates found by Zorlu (4.5%), and Yücel (5.8%) [12, 20].

The increase in the cesarean delivery rate is leading to an increase in the rate of abnormal placentation (placenta previa and accreta), which in turn gives rise to an increase in peripartum hysterectomy rate. Cesarean section itself is also a risk factor for emergency peripartum hysterectomy. Therefore, every effort should be made to reduce the cesarean rate by performing this procedure only for valid clinical indications. For this purpose, counseling about the family planning should be given especially for the patients with repeat cesareans. In case of placenta previa, antenatal care is important for the identification of placental localization, prevention of cervical examination and coitus, counseling about the symptoms of labor and identification of anemia especially in the third trimester.

In conclusion, the risk factors associated with peripartum hysterectomy, such as the high maternal mortality and morbidity and the loss of future fertility, should be identified antenatally. Many cases of placenta previa can be identified or guessed with high probability before delivery in women who had previous cesarean sections and anterior placenta previa. The delivery and operation should be performed in appropriate clinical settings by experienced surgeons when such risk factors are identified.

Conflict of interest statement

None.

Copyright information

© Springer-Verlag 2010