Maternal and perinatal outcomes of uterine rupture in Lubumbashi, Democratic Republic of Congo

Methods: Institution based cross sectional study was conducted from December 1st, 2012 to 31st, 2016 on uterine rupture. During the study selected 158 women were included by using exhaustive sampling method. Data were checked, coded and analyzed into STATA version 12. Chi-square test was used to identify the predictors of maternal and perinatal mortalities in women with uterine rupture and 95% Confi dence Interval of odds ratio at p value less than 0.05 was taken as a signifi cance level.


Introduction
Uterine rupture is de ined as a non-surgical continuity solution of the pregnant uterus reaching the uterine fundus and/ or the lower segment. It can be spontaneous or provoked [1]. It occurs most often during labor, but it can also occur during pregnancy, on a healthy or scarred uterus. Anatomically, there are two forms of uterine rupture: complete rupture (or intraperitoneal rupture) and incomplete rupture (or disunity when it occurs on a scar uterus) [1]. The frequency of the occurrence of uterine rupture varies in different regions of the world. It is rare in developed countries but remains a public health problem in developing countries, particularly in Africa where it occurs mainly as a result of prolonged or obstructed labor [2,3]. Uterine rupture is one of the major obstetric complications that contribute signi icantly to maternal and perinatal mortality and morbidity [2], especially in our underequipped countries where accessibility to quality health care is a problem. Its prevalence in Africa is between 0.3 and 2.4% [2,4] while in Europe and industrialized countries in general it is below 0.1% [5,6]. The majority of the cases in these countries occur in women with a history of cesarean section while in developing countries it is rather prolonged labor on non-scarred uterus [2,7]. In the Democratic Republic of Congo (DRC), in Lubumbashi, Kakudji, et al. [8] reported 12 (0.41%) cases of uterine rupture out of a total of 2,911 deliveries recorded in 10 reference maternities. Mwilenyi, et al. [9], in their study at Jason Sendwe Hospital in Lubumbashi, found a frequency of 1.1% and the maternal and perinatal mortalities were 15.1% and 76.1% respectively.
However, there are very few studies on this subject in DRC despite the observation made about the existence of this pathology and its consequences in our obstetric practice. In view of the situation described above, it is imperative to know the situation in our country and particularly in Lubumbashi city.
This study aimed to assess the prevalence and associated factors of mortality among women with uterine rupture in referral hospitals of Lubumbashi, in the south east part of the Democratic Republic of Congo.

Study setting
The study setting was Lubumbashi city, which is located in the south east part of the Democratic Republic of Congo. We selected 9 public referral maternities funded by Government of DRC through the Ministry of Health. (Provincial General Reference Hospital Jason Sendwe, University Clinics, Gécamines-Sud Hospital, SNCC Hospital, General Reference Hospital of Katuba, General Reference Hospital of Kamalondo, General Reference Hospital of Kampemba, General Hospital Reference of Rwashi and General Reference Hospital of Kenya). Lubumbashi is a city with a population of more than 2 million people.
Institution based cross sectional study was conducted from December 1 st , 2012 to 31 st , 2016 on uterine rupture.

Case defi nition and selection
Uterine rupture was de ined as tearing of the uterine wall either partially or complete during pregnancy and labor, diagnosed either clinically and later con irmed at laparotomy. The cases were retrospectively collected from the maternity ward and operating theatre registers as well as from the patients' case iles at the hospital medical records of ice.
In total, we included by using exhaustive sampling method 186 cases of uterine rupture which allowed us to calculate the frequency. But of the 186 cases, we excluded 28 cases in the statistical analyzes due to an incomplete record missing most of the study variables.

Data collection
Data was abstracted from the maternity ward and operating theatre registers as well as from the patients' case iles at the hospital medical records of ice using a pre tested case report form. Information on the patients' age, parity, previous caesarean section, antenatal care attendance, modality of admission at the hospital, gestational age, type of uterine rupture, site of uterine rupture, type of treatment, maternal and fetal outcomes, length of postoperative hospital stay and other relevant information were collected.
The data collectors were midwives who were trained to collect data from women's obstetric iles or charts and to validate the diagnosis of obstructed labor using admission, delivery and theatre registers. Thought the data collection period which lasted 4 weeks, the irst author was providing oversight and supervision to prevent under reporting.

Statistical analysis
The data were entered and analyzed using STATA software, version 12.0. Descriptive statistics were obtained through frequencies and cross tabulations. Chi square test or Fisher exact test when appropriate were used to identify the predictors of maternal and perinatal mortalities in women with uterine rupture and 95% Con idence Interval (95% CI) of odds ratio (OR) at p-value less than 0.05 was taken as a signi icance level.

Ethical considerations
Ethical approval was obtained from the Ethical Medical Committee of the University of Lubumbashi. Permission to access obstetric records was obtained from the medical director of the hospital and these were anonymously entered into the database.
The ages of mothers who got uterine rupture ranged between 15-44 years, with an average age of 29.5 ± 6.2 years. The average parity was 4.7 ± 2.5 with (range: from 1 to 11). Most of the cases 132 (83.54%) were referred from peripheral maternities, and 57 (36.06%) mothers had not performed any antenatal care (ANC) visit. The major part of uterine ruptures had occurred in term pregnancies (82.17%). The average gestational age was 38.23 ± 2.65 weeks (range: from 28 to 42 weeks).
Sixty-eight (43.04%) cases had occurred on a scarred uterus (history of the previous caesarean section) and 90 (56.96%) cases on a healthy uterus.
The rupture site was essentially the lower segment in 51.27% of the cases, and the rupture was complete in 81.17% and incomplete in 18.83%.
No parameter studied was associated with maternal death ( Table 2). As for perinatal death (Table 3), we found that parity ≥4 was signi icantly associated with perinatal death, indicating that newborns of mothers with parity ≥4 had 3.5 times the risk of dying compared to those of mothers having a parity <4 (OR = 3.5 [1.7-7.2]). Similarly, for newborns of mothers who had a complete rupture, they were 13.9 times more likely to die compared to those whose mothers had an incomplete rupture (OR = 13.9 [5.3-36.5]). As for the site of the rupture, the segmento-corporeal lesion was signi icantly associated with perinatal death (OR = 6.4 [1.9-21.3]).

Prevalence
The frequency of uterine rupture varies according to hospital statistics. We recorded an overall prevalence of 0.49%. This frequency is comparable to that recently reported in Lubumbashi (DRC) by Kakudji, et al. [8] which was 0.41%. It seems to be higher than that reported in 1994 in Nyankunde (in the north east part of the DRC) by Ahuka, et al. (0.35%) [10].
We think that these frequencies reported in African countries could be underestimated because few studies have been carried out in rural areas (where home deliveries have so far been noted), and most of the data come from urban centers, where the incidence of uterine rupture may be lower because services are more easily accessible [22].
This high frequency of uterine ruptures in developing countries is probably linked to lack of equipment, low number of quali ied and trained caregivers in obstetric and neonatal emergency care in maternities and high number of makeshift maternities (where diagnosis, management and timely referral of obstetric emergencies remain a serious problem) which are constantly emerging in some large African cities. To these are added the problem of lack of ANC visits in our settings where less than 30% of pregnant women reach 4 or more ANC [8,23].

Type and site of the uterine rupture
In our study, 81.17% of uterine ruptures were complete  and 18.83% were incomplete. This predominance of complete uterine ruptures over incomplete ones is also found in several studies [3,4,24,25]. Contrary to our observation, Guyot, et al. [7] (in France) found a higher proportion of incomplete ruptures (69.44%) than complete ruptures.
We noted that in only 10% of the cases did the uterine rupture sit in the fundus; 51% had involved the lower segment alone and nearly 38% of the uterine ruptures were segmentocorporeal. Kouakou, et al. [26], Astatikie, et al. [4] and Diallo, et al. [14] found that many uterine ruptures were located in the lower segment with 77.48%, 71.5% and 70.41% of the cases, respectively.
Fouedjio, et al. [24] had found 50% of segmento-corporeal lesions. In these segmento-corporeal lesions, the uterine breach would certainly have started in the lower segment and would have spread to the uterine fundus because of the delay in management. Because of its fragility, it is the lower segment which is the choice's site for rupture, thus constituting the starting point for most uterine ruptures [6]. Also, we must add the important role of low segment caesarean section as provider of segmental and segmento-corporeal ruptures. Corporeal ruptures are often secondary to corporeal caesarean sections which are rare, but also to myomectomies.

Management
Repair of the ruptured uterus (conservative surgery) was the most common (93.04%) in our series; it was performed whenever state of the uterus, site and extent of lesions had permitted it. This conservative tendency is found in several studies from African countries [14,27,28] because the procreative function is an important element in the balance of African households [14].
In our series the proportion of hysterectomy was 5.06%. Guyot, et al. [7] had not reported any hemostasis hysterectomy in their series. The rate of hemostasis hysterectomies varies in the literature between 5.6 and 61.5% in the event of complete ruptures [6,14,[29][30][31][32]. Mukasa, et al. [2], in their series of 83 Ugandan women, recorded 28.6% total hysterectomy and 37.7% subtotal hysterectomy. Astatikie, et al. [4], in their series of 242 Ethiopian women, reported 73.9% hysterectomy (57% total hysterectomy and 16.9% subtotal hysterectomy). These not very homogeneous rates seem to be linked to the difference in attitude adopted taking into account certain parameters, namely age, parity, extent of lesions, involvement of vascular pedicles, state of the uterine muscle, desire of the patient's subsequent pregnancy and presence or absence of infection.

Maternal outcome
Out of 158 cases of uterine rupture, we observed 14 (8.86%) maternal deaths or approximately one death for 11 cases of uterine rupture. This result could be explained by the delay in referral women from the irst level health centers to the reference hospitals and also by the delay in diagnosis and treatment. Nayama, et al. [31], in their series of 195 Niger women, no maternal death was recorded. The literature review shows that the maternal mortality found in African countries varies between 4.8 and 21.4% [11,12,14,27], and between 5 and 12.5% in Asian countries [3,16,[17][18][19][20][21][22][23][24][25]. It is still largely extended to that observed in the United States, which is 0.29% [33], to that observed in Europe, which is less than 1.5% [5,7,20,34,35].
Our study shows that none of the variables studied were associated with maternal death. Eze and Ibekwe noted that the maternal deaths recorded in their study were due to hypovolaemia secondary to massive hemorrhage and septicemia [27]. Astatikie, et al. [4] found that women whose birthing was done at home, those who developed hypovolemic shock or severe postoperative anemia were more likely to die from the uterine rupture. For Ahmed, et al. [30], maternal death was signi icantly associated with duration of surgery greater than 2 hours and they pointed out that this would be attributed to the complexity of the uterine lesion and the poor general condition of the patient. As Ahmed, et al [30], we believe that early admission of pregnant women to hospital, timely diagnosis of uterine rupture, adequate resuscitation of patients, availability of a blood transfusion, short delay between diagnosis and management as well as presence of an experienced surgeon have a reducing effect on maternal death after uterine rupture. Osemwenkha and Osaikhuwuomwan [36] emphasize that late admission to the hospital is a major cause of this poor outcome and is due to poverty, delays in referral, poor transport network and lack of ambulance services.

Perinatal outcome
Perinatal death rate in our study was 72.04%. The feticidal nature of uterine ruptures had been found in numerous African studies where the perinatal mortality linked to uterine rupture varied between 61.29% and 98% [4,13,[36][37][38]. These igures are comparable to those found in Asian countries where it varies between 61.90% and 87.50% [25]. It is very much higher than that observed in Europe, varying between 0% and 15.72% [7,34,35], and in the United States where it is 1,44% [33].
In our study, complete and segmento-corporeal uterine lesions were signi icantly associated with perinatal death. This could be explained by the fact that these lesions would be more hemorrhagic and therefore likely to quickly cause death of the newborn because maternal antepartum hemorrhage is a risk factor for perinatal death. The delay in diagnosis and management (most of patients being admitted after rupture) and an under-equipped neonatal resuscitation structure are believed to be the cause of this high rate of stillbirths in developing countries.

Conclusion
Uterine rupture is a devastating condition that carries grave risks to the newborn and the mother as well. Uterine rupture is an obstetric condition with catastrophic consequences. The resulting high perinatal mortality suggests a synergy of action between political decision-makers, health caregivers and the community itself to promote antenatal care, family planning, to improve the health system through periodic refresher courses. Likewise, the provision of medical transport at each health center is necessary to speed up medical referrals, in order to eliminate delays in management.