Peripartum hysterectomy: a ten-year experience at a tertiary care hospital in a developing country

Introduction

Acute bleeding after delivery is a complication which, although unpredictable, can usually be controlled by uterotonic medications. In life-threatening situations, however, an emergency hysterectomy may have to be undertaken as a last resort. ¹ An emergency hysterectomy has been defined as a hysterectomy performed at the time of delivery or in the immediate postpartum period. It was Heratio Storer who performed the first subtotal caesarean hysterectomy on a woman on 21 July 1868. His patient survived the operative procedure but died after 78 h. Eduardo Porri performed the first successful operation in 1876. ² The incidence of emergency hysterectomy is reported to range from 0.2 to 1.5 per 1000 deliveries. ^3,4

Previously, the most common indications for peripartum hysterectomy were uterine atony and uterine rupture. ⁵ However, more recently, as there has been an increase in the number of caesarean section (CS) deliveries, the most common indication has become morbidly adherent placenta. ⁶ Morbidly adherent placenta, uterine atony and a previous CS have been reported to increase the risk of emergency obstetric hysterectomy (EOH) by two- to sixteen-fold. ⁶ Although EOH is generally an uncommon procedure, it is associated with a relatively high incidence of maternal mortality and morbidity. ⁷ This study focuses on EOH for pregnancies over 24 weeks.

The purpose of this study was:

To estimate the frequency of emergency peripartum hysterectomy at a single tertiary care institution;

Materials and methods

This is a retrospective review of hospital charts identified from the hospital obstetric database of patients who underwent obstetric hysterectomy between January 1995 and December 2006 at the Aga Khan University Hospital. All of the hysterectomies in our series were undertaken for life-saving indications, such as ruptured uterus, massive bleeding from placenta previa or uterine atony, except for one case where the indication for procedure was predefined.

Data was abstracted from individual medical charts and laboratory records. In order to ensure that no cases were missed, the search was run twice using the keywords caesarean hysterectomy and peripartum hysterectomy. Forty-five obstetric patients had an emergency hysterectomy during this time. The individual charts were scrutinized and demographic data, as well as clinical details, were retrieved in order to determine the indications for and type of hysterectomy performed along with postoperative morbidity and mortality.

Data was available for all variables except the time spent on the conservative management of postpartum haemorrhage during the surgery. As the total operating time included both, we only know the total time spent in operating theatre and do not have the breakdown of how much time was spent in conservative management and how much was taken to perform the operation. The records of six cases could not be accessed (6/45, 13.3%). The data studied included age, parity, gestational age, birth weight, previous CS delivery, placenta previa, type of hysterectomy and operative complications. Indications for hysterectomy were classified as: uterine atony; morbidly adherent placenta with previa; placenta previa with bleeding bed; and uterine dehiscence or rupture. Outcomes included: the estimated blood loss; the amount of blood transfused; the time from delivery to the completion of the hysterectomy; the operating time; the type of hysterectomy performed; the length of stay in the intensive care unit and the hospital; and any associated complications.

The statistical analysis of the data was performed with SPSS 16.0. Summary statistics are presented as mean with standard deviations and P values. Dichotomous variables were analysed with chi-square or Fisher exact tests as applicable, and t-tests were used for continuous variables. A P value of <0.05 was considered statistically significant.

Results

There were 31,360 deliveries during the 10-year study period; 22,151 were vaginal deliveries and 9308 CSs. During this time, 45 women with peripartum hysterectomy were identified, representing an incidence of 1.4 per 1000 deliveries. The data for six cases could not be accessed.

Morbidly adherent placenta was the most common (18, 46.15%) indication for hysterectomy followed by uterine atony (9, 23.08%), placenta previa (7, 17.94%) and uterine dehiscence/rupture (5, 12.82%). About two-thirds of the hysterectomies were subtotal (STH; n = 26, 66.7%) and the rest (n = 13, 33.3%) were total hysterectomies (TH). Uterine atony was the only indication for hysterectomy in primiparous women. For multiparous women, morbidly adherent placenta (42%) was the most common indication followed by uterine atony (22%). Of the patients with morbidly adherent placenta, 93% had a prior history of CS and 40% had uterine curettage. Conservative measures were used in smaller number of patients before embarking on hysterectomy. The methods used include uterine packing (7, 17.9%), stitches on placental bed (6, 15.4%), internal iliac ligation (5, 12.8%) and balloon tamponade (1, 2.6%).

Our analysis of the outcomes of emergency peripartum hysterectomy (Table 1) shows that the duration of surgery was significantly shorter (P = 0.045) for total hysterectomy (136 minutes) in comparison with subtotal hysterectomy (173.4 minutes). Although the length of hospital stay was shorter for patients who underwent subtotal hysterectomy (mean = 7.08 days) compared to total hysterectomy (mean = 9.46 days), this difference was not statistically significant. Furthermore, there was no statistically significant relationship between the blood loss and the duration of the surgery or the time from delivery to the completion of the hysterectomy.

There was an average of 1.36 ± 0.49 complications post-operatively (Table 2). Patients who had undergone TH were found to have a statistically significant (P = 0.029) higher number of average postoperative complications (1.62 ± 0.51) than those who had subtotal hysterectomy (1.23 ± 0.43). However, there were no statistically significant differences in the type of complication between the two groups. There were two (5.1%) maternal deaths out of the 39 cases.

Discussion

Our results suggest that vaginal birth after primary CS and repeat CS deliveries are associated with an increased risk of peripartum hysterectomy. These findings may be of concern, given the increasing rate of CS deliveries even in developing countries, especially for avoidable indications such as maternal request and breech.

Recent studies have indicated that abnormal adherent placentation is replacing uterine atony as the most common indication for emergency peripartum hysterectomy. ^2,3,8 The advances in pharmacological modalities might explain the decrease in uterine atony as the cause for peripartum hysterectomy. At the same time there have been increasing CS rates because of placenta previa, with or without accrete, leading to decreased proportion of atony as an indication for EOH.

In our series, 42% of cases were for morbidly adherent placenta. While these cases had not been diagnosed prenatally, they were anticipated as the majority had previous uterine scars which were likely to increase the risk of defective implantation. In the past, an antenatal suspicion of morbidly adherent placenta on the basis of a history of CS was used when planning surgery. However, in recent years the use of color Doppler and magnetic resonance imaging has been gaining popularity when establishing the diagnosis preoperatively. ^8–9 The use of such modalities may facilitate in the planning of the surgery (timing and presence of a highly skilled surgeon) which will eventually reduce unnecessary blood loss and other possible intraoperative and postoperative complications.

The majority (76%) of the patients in our study were found to be anaemic (preoperative haemoglobin <11.0 g/dL). ¹⁰ As has been widely reported in literature, anaemia is a very common problem faced by the childbearing women of Pakistan. Poor nutrition, high parity and poor antenatal care are a few of the factors responsible for this. Eighty-one percent of the women who underwent elective CS were found to be anaemic prior to the procedure. Although this figure strongly reflects the standard of antenatal care provided, we cannot account for other factors such as patient compliance and socioeconomic status. Every possible effort should be made to maintain a reasonable haemoglobin level (above 12 gm/dL) in these high risk patients and they may need repeated counselling sessions.

In our study, the mean hospital stay (7.87 days) was shorter for both TH and STH in comparison with other studies. ² Patients who had undergone STH (7.07 days) had a speedier recovery than those who had undergone TH (9.46 days). This was found to be consistent with the results of other studies. In our study the difference was not statistically significant.

Most hysterectomies are performed as an emergency operation when all other methods have failed to control bleeding. When it becomes essential to perform a hysterectomy, it is better to execute it timely, before the patient's status deteriorates any further. The mean time from delivery to completion of hysterectomy in our study population was found to be shorter than that reported by Forna et al. ¹¹ This was contrary to the presumption that we tend to spend more time saving the uterus due to cultural taboos associated with hysterectomy (a procedure that ends the reproductive potential of a female). This is consistent with the recommendation given in the Confidential Enquiries into Maternal and Child Health UK ¹² that hysterectomy should be undertaken sooner rather than later, as literature suggests that earlier hysterectomy results in lower blood loss and, possibly, lower rates of additional morbidity. ¹³

It remains questionable whether we are making enough effort to reduce the third delay by skipping the conservative measures, even in situations where they might have been tried and hysterectomy could have been avoided. ¹⁴ However, as this was a study of peripartum hysterectomy, we do not have any information about the number of women who were treated with other therapies which allowed the women to avoid a hysterectomy and preserve their future fertility.

The mean blood loss during CS hysterectomy in our study was comparable to other studies. Similarly, the mean operating time for TH was also comparable with the current literature. However, it was observed that, in our setting, STH had a much longer operating time. This might be explained by the fact that the decision to perform a STH was actually taken in cases where technical difficulties were encountered for TH (e.g. adhesions).

Maternal mortalities in a tertiary care centre are a rare event in view of improved postoperative care, readily available blood products, ICU facilities and multidisciplinary input. However, we had two cases of maternal mortality – both the women were unbooked patients. Their death can be attributed to factors such as late referral, severe anaemia, disseminated intravascular coagulation and other co-morbids.

Our study had a lower rate of postoperative complications (both TH and STH) compared with other studies.

Conclusion

Although the protocol for the management of postpartum haemorrhage (PPH) encompass all possible management options, the utilization of this protocol occurs at a subjective level depending upon the cause, amount of bleeding and surgeon preference (which is dependent upon experience and exposure to different modalities of management). A protocol describing the management option for PPH during CS with time lines should be formulated. Sequence and time spent for different uterus conserving methods need to be investigated. Among these conservative methods, balloon tamponade with a Foley's or Sangestaken Blakemore tube has shown promising results in the conservation of the uterus in cases of atony and placental bed bleeding. ¹⁵ The hydrostatic condom has been introduced and used successfully to control PPH quickly and effectively. It is simple to use, inexpensive and safe. These procedures do not demand high skills and any health-care provider experienced in delivery may use this procedure for controlling massive PPHs. It should, therefore, be included as a part of PPH protocol, before embarking on an EOH. ¹⁶ Obstetric compressions sutures are effective and easily learned techniques, and are a promising alternative to hysterectomy for severe PPH due to uterine atony. The use of compression sutures does not appear to compromise or jeopardize subsequent pregnancies. ¹⁷

Once a timed sequence is introduced, an audit could be done in order to compare these sequences.

Women with recognized antenatal risk, such as previous CS, placenta previa and potential risk for morbidly adherent placenta, should be counselled for the possibility of a hysterectomy and booked for care at an obstetric unit with on-site transfusion services and skilled obstetricians who are available around the clock in order to avoid delayed transfer and the resultant mortality.