Contribution to the Study of Hemorrhages in the Third Trimester of Pregnancy, Etiology and Management

Third trimester bleeding is a common concern in obstetrics. The main objective of this work was to study the management of hemorrhages in the third trimester of pregnancy in the maternity ward of the Sominé Dolo hospital in Mopti. Our prospective descriptive cross-sectional survey type study conducted at the maternity ward of Sominé Dolo hospital in Mopti over a period from January 1, 2017 to December 31, 2017 included 94 cases collected. During this period we had performed 1485 deliveries including 94 cases of pregnancies complicated by 3rd trimester hemorrhage, a frequency of 6.33%. The main cause of hemorrhage in the third trimester was represented by placenta preavia 42.6% followed by retroplacental hematoma 28.7%, uterine rupture born. In 55.3% of cases neonatal mortality occurred antenatally. Neonatal morbidity was represented by prematurity, i.e. 20.2% (n = 19/94) and low birth weight, i.e. 22.3% (n = 21/94).


Introduction
Third trimester hemorrhage is external bleeding through the vagina during the third trimester of pregnancy. It occurs in less than 5% of pregnancies and constitutes an obstetric emergency with a risk of maternal and fetal morbidity and mortality. The urgency is first to assess the bleeding and its maternal-fetal impact and to stabilize the general condition before looking for an etiology [1].
Thus, in France, third trimester hemorrhage represented immediate postpartum hemorrhage as the first cause of maternal death with a rate of 17% and for the same period in the United Kingdom and the United States the 4th cause of death with a rate of 5% [2] [3].
The incidence of bleeding during the third trimester of pregnancy varies from 3% to 5% in developed countries: in France, AYOUBl noted 5% in 2000. In LONDON, a study carried out by DEREK in 1981 found a rate ranging from 3 to 6% [2] [3] [4].
In Africa, Akpovi in Benin noted a lower rate of 2.42% [5]. In Mali, a study carried out at the Gabriel Touré University Hospital Center by F Kané concluded at a rate of 2.6%, or 67 cases of hemorrhage in the third trimester of pregnancy out of a total of 2568 deliveries carried out in this hospital. This same author finds that bleeding during the third trimester of pregnancy represents 18.7% of the causes of maternal death with respective maternal and fetal lethality rates of 4.54% and 5.57% [6] [7] [8] [9] [10].
In 2009 at the Sominé Dolo hospital in Mopti, they represented the 2nd cause of maternal death after eclampsia, and the first cause in 2008 with postpartum haemorrhages [11].
These hemorrhages of the third trimester of pregnancy are not only a daily concern in current obstetric practice but also they encompass a range of obstetric pathologies (placenta previa, retro-placental hematoma and uterine rupture) whose delay in management could be detrimental to mother and fetus.
However, their brutal nature, the insufficiency and/or the lack of prenatal follow-up, the delay and/or the absence of a diagnosis, the insufficiency or the unavailability of the means of resuscitation (blood and blood products) confer on the hemorrhages of the third quarter all their severity.
In Mali, even if the free caesarean section had greatly improved the prognosis of these haemorrhages, it was clear that at the Sominé Dolo hospital in Mopti, the only 2nd level reference structure, haemorrhages in the third trimester of Open Journal of Obstetrics and Gynecology pregnancy constituted in 2017 a public health problem. The main objective of our work was to study the maternal-fetal prognosis of hemorrhage in the 3rd trimester of pregnancy in the maternity ward of the Sominé Dolo hospital in Mopti.

Material and Methods
Our study was conducted at the maternity ward of the gynecology-obstetrics department of the Sominé Dolo hospital in Mopti, which was the only 2nd level reference health structure of our national health pyramid in the 5th administrative region of the center of the country. Our study was a descriptive study of the cross-sectional survey type and extended over a period of 12 months from Janu- Grand multigravesta > to 6 pregnancies, Nulliparous = 0, childbirth, Primiparous = 1 childbirth; Pauciparous = 3 to 4 deliveries; Multipara = 5 to 6 deliveries; Grand multipara > to 6 deliveries; Full term newborn = ≥37 weeks of amenorrhea, Premature = 28 -36 weeks of amenorrhea. Cesarean in first intention = absolute cesarean, Cesarean in second intention = relative cesarean; Poor general condition was defined in our context by a state of shock with coma and or blood pressure less than or equal to 8/4 cm hg, hemoglobin level ≤ 6 g/dl, Glasgow score varying from 6 to 7; Fair general condition = blood pressure between 9/5 cm Hg and 10/6 cm Hg, hemoglobin level varying from 7 g/dl to 10g/dl with obnubilation, Glasgow score varying from 9 to 12; General condition good: blood pressure ≥ 11/7 cm Hg, hemoglobin level ≥ 11 g/dl, good conscience, Glasgow score varying from 13 to 15; Coma is defined as the suppression of alertness and consciousness; Obnubilation: is the state of drowsiness interspersed with a period of consciousness with reduced attention and memory impairment; Severe anemia = hemoglobin level ≤ 6 g/dl.
Moderate anemia = hemoglobin level between 7 to 9 g/dl; Evacuation: reference carried out in an emergency context; Referral: Mechanism by which a health facility directs a case that exceeds its skills to a more specialized and better equipped structure outside of any emergency situation.

Results
In total, during our study period from January 1, 2017 to December 31, 2017, we collected at the obstetrics gynecology department of the Sominé Dolo hospital in Mopti 1485 deliveries including 94 cases of pregnancies complicated by 3rd trimester hemorrhages, i.e. a frequency of 6.33%.
The epidemiological profile consisted of a 20 -35 year old age group which was the most represented (73.4%). In addition, the majority of the women were married (95.7%), living in rural areas (59.6%) and were housewives (95.7%).
These data are presented in Table 1 below.
About 1/4 of our patients had a non term pregnancy (23.4%) and about 2/3 (66.6˚) of the patients had a more or less severe anemia. The main cause of hemorrhage was placenta previa (42.5%) followed by retroplacental hematoma (28.7%) and uterine rupture (26.6%). Maternal morbidity was marked by hypovolemic shock (48.9%) followed by infections (28.8%) and coagulopathy. This information is presented in Table 2.
Most of the patients had not had any prenatal consultation, i.e. 48% of the cases ( Figure 1).
More than 8/10 (87.2%) of the patients had delivered by caesarean section, and the route of delivery and type of intervention depended on the condition of the mother and the fetus. The main causes of neonatal mortality were uterine rupture and retro placental hematoma with respectively 48.08% and 38.47% (Table 3).

Discussion
The limits of our study were: the absence of an ultrasound device in the delivery room for the etiological diagnosis and the problem of staffing the department with emergency drugs but also the insufficiency of blood products at the level of our laboratory lack of regular donors. Thus, these last factors had a serious impact on the adequate care of our patients.  Coulibaly Y [13] in his study had reported a similar age group 19 to 35 years but at a lower proportion (52%) than that of our study (73.4%). The 19 -35 age group would correspond to the optimal fertility period and the percentage difference could be related to the sampling. In our study, placenta previa was the main cause of these hemorrhages with 42.6% (n = 40) followed by retro placental hematoma 28.7% (=27), uterine rupture 26.6% (n = 25) and an association of placenta previa and retro placental hematoma 2.1% (n = 2) ( Table 2). Our result was comparable to that of Coulibaly Y [13] in terms of etiological frequency of hemorrhages with respectively Placenta preavia 42.2%, retroplacental hematoma 35.3% and uterine rupture 18.6% of cases. The factors associated with the high incidence of uterine rupture in our study compared to the literature would be mainly related to the three delays in the evacuation referral system. This would be the delay in making the decision to use local health structures, the delay in geographical accessibility, and the delay in receiving first aid at our community health centers (CSCOM). In addition to these delays, other factors may contribute to poverty, including the lack of access to local health facilities for prenatal consultations [14] [15] [16].
He also noted that 21.8% of bleeding in the third trimester of pregnancy was unexplained in his study. The indication for treatment depended on the etiology of the haemorrhage but also on the maternal-fetal state. Our caesarean section rate was 87.5% (35/40), which was much higher than that of IZRAR N [17] who reported 60% of cases in his study. This discrepancy could be related to the study methodology in terms of period duration and sample size (Table 3). Regarding uterine rupture, which represented a significant proportion in our study, 52% of our patients had undergone hysterorrhaphy against 48% hysterectomy for hemostasis. This high hysterectomy for haemostasis rate could be explained by the high morbidity linked to infections, hypovolemic shock and coagulopathy (Table 3). We deplored 11/94 cases of death, i.e. 12% maternal mortality in our study. This rate was much higher than those of IZRAR N. [17] 2% (2/110). Our high rate of maternal death would be related to the delay in referral-evacuation, transport difficulties due to isolation and the permanent non-availability of whole blood at the Sominé Dolo hospital in Mopti and above all the inability to extract fresh frozen plasma which is essential to correct coagulation disorders despite hysterectomy for haemostasis. In our specific case, the causes of these maternal deaths were retro placental hematoma 64% (n = 7/11), uterine rupture Open Journal of Obstetrics and Gynecology 27% (3/11) and ultimately the association of placenta previa and retroplacental hematoma 9% (1/11). The fetal prognosis, like that of the mother, was poor marked by prematurity and a high rate of neonatal deaths with respectively 22.3% (n = 21) and 55% of cases (n = 52). Our mortality rate was much higher than that reported by IZRAR N [17] which was 18.6%. Our high death rate would be related, as in the case of maternal death, to the delay in evacuation referrals.

Conclusion
It appears from our descriptive study of cross-sectional survey type that the main cause of bleeding in the third trimester of pregnancy was placenta previa 42.6% followed by retroplacental hematoma 28.7% and uterine rupture 26.6%.
The maternal-fetal prognosis had been impacted by the late evacuations in relation to the difficult geographical accessibility of the region but also the severe lack of blood products for the management of these third trimester hemorrhages.

Authors' Contribution
Our thanks go to Tioukani Augustin Thera and Pierre Coulibaly who were the designers of our study. The questionnaire was developed and validated by these two authors. The drafting of the manuscript was done by Seydou Mariko. All authors have contributed substantially to the development of this work. All authors have read and approved the manuscript until submission.