Despite advances in medicine and surgery, postpartum haemorrhage remains one of the leading causes of maternal morbidity and mortality [29]. EPH is performed in the setting of life threatening haemorrhage and its incidence is considerably variable among countries and institutions. The incidence is very high in LIC countries, compared to HIC [2, 5]. The overall incidence of EPH at our hospital was 2/1000 deliveries, which is higher than those reported in HIC but is lower than the study done by Zeterogluet al in Turkey and in low-income countries [2, 5, 12, 22, 23].
In our study, the rate of EPH was higher among older women (age≥30years) and in multiparous and grand multiparous women. This is in agreement with a study done in Norway [11], Nigeria [12] and Queensland [10]. A similar trend was observed in study done in Portugal [18] and Turkey [26]. Other risk factors for EPH include, current Caesarean delivery, uterine atony and abnormal placentation such as placenta previa and placenta accreta, were similar to the findings in the literature [1–3, 7–10]
Current delivery by CS is associated with increased EPH. In the present study, the rates were 20 fold higher among women, who delivered by CS than those who delivered normally. The association between current CS delivery and EPH is consistent with the findings of other studies [1–3, 6–10]. The relationship between CS and EPH may be due to the increased risk of severe post partum haemorrhage associated with this method of delivery or a CS may be performed due to high risk complications, such as ruptured uterus or placenta accreta, a condition that can deteriorate and necessitate hysterectomy. The decision to perform EPH may also be more easily made after Caesarean delivery than after vaginal delivery [1, 10]. However, prior CS delivery seems to be unrelated with EPH.
There is a significant difference in indications related to EPH in HIC and LIC. The most frequent indication in the present study was uterine atony followed by uterine rupture and morbidly adherent placenta with or without placenta previa. This could be attributed to the high incidence of multiparity and present Caesarean delivery, which may have contributed to the impairment of uterine contractility. Atony was also the most common reason for EPH in LIC [2, 19, 20], Turkey [17, 26] and UK [27]. Conversely, in HIC, abnormal placentation has been found as the primary etiological causes for EPH [1–3, 6–10].
Uterine rupture is the second most frequent indication for EPH in this study. Advanced maternal age, multiparity and late referrals may be considered as the underlying factors for the relatively higher rate of uterine rupture in this study. Uterine rupture has been reported as the second most common indication for EPH in similar studies done in Nigeria [23] and Turkey [17]. However, statistics reported from Nigeria by Omiole-Ohons et al [12] and from Pakistan by Korejo et al [28], the commonest cause of EPH was uterine rupture.
Although EPH is lifesaving, it carries a high risk of maternal and fetal mortality and procedural complications. The case fatality rate of 13% and the still birth rate of 34% is evidence to this. Maternal mortality in our series is very high as compared to the studies in HIC [1–3], whereas it is very low as compared to studies done in Africa [5, 23], India [20] and Turkey [26]. Although a high proportion of women who underwent EPH were delivered by CS, 34% of them incurred fetal lose. This outcome is lower than studies done in LIC [2, 23, 28] and comparable to results of studies done India [20], but it is significantly higher than the studies done in HIC [2, 18, 21]. These could be explained by high rates of ruptured uterus in these studies, which is known to have a detrimental effect on perinatal outcome [20, 23, 28]. The other complications reported in this study were similar to other reports [12, 13, 23, 26].
There is often a debate regarding the benefits of subtotal vs. total hysterectomy. The present study has found a preference for subtotal hysterectomy. This could be because most of the patients were hemodynamically unstable at the time of the operation and it is safer, faster, and easier to perform as compared to total hysterectomy [13, 25]. Comparable incidences have also been reported in findings of various studies [12, 13, 17, 23, 26].
The major limitation of this study was being a retrospective study. The study was also limited in that the data collected were from a single institution and relatively its small cohort size.