Prevalence and Risk Factors of Primary Postpartum Hemorrhage after Vaginal Deliveries in the Bonassama District Hospital, Cameroon

Backgrou nd: Approximately 800 women die from pregnancy-related conditions daily with 99% of these maternal deaths occurring in low resource countries. Primary postpartum haemorrhage accounts for 25.7% of maternal deaths in Africa. In Cameroon, primary postpartum hemorrhage Original Research Article Halle-Ekane et al.; IJTDH, 13(2): 1-12, 2016; Article no.IJTDH.23078 2 remains the leading cause of maternal death. Objective: The aim of this study was to determine the prevalence, risk factors, and the maternal and fetal outcomes of primary postpartum hemorrhage in the Bonassama District Hospital, Douala, Cameroon. Methods: This study was a cross-sectional study. Of the 556 women interviewed regarding their willingness to participate in the study, 550 consented. Quantification of blood loss was done over the 24 hours using a 700cc measuring cylinder, an electronic weighing balance and the Pathfinder International wall chart for visual estimation of blood loss. The well-being of the neonates was evaluated by a pediatrician. Data was collected using a structured questionnaire and analyzed using EPI-INFO version 7.1.4.0. A P-value <0.05 used to determine association between variables was considered statistically significant. Results: The prevalence of primary postpartum hemorrhage was 23.6%. The risk factors of primary postpartum hemorrhage were: previous postpartum hemorrhage (aO.R=6.50; 95% C.I=4.60-9.20), multiparity (aO.R=7.02; 95% C.I=5.27-8.02), prolonged labor (aO.R= 1.1; 95% C.I=1.01-6.61) and macrosomia (aO.R=8.18; 95% C.I= 5.53-8.51). The main causes of PPH were uterine atony and obstetrical lacerations (P-value<0.01). Two (1.5%) maternal and 2 (0.4%) perinatal deaths were recorded. Conclusion: The prevalence of postpartum hemorrhage was high. The main cause of primary post-partum hemorrhage was uterine atony. Primary PPH was associated with uterine hypotonia, placenta retention and genital lacerations. The maternal and perinatal mortality were high. These findings highlight the pressing need for good quality emergency obstetric care and the availability of more accurate techniques of postpartum blood loss measurement. Secondly, hospitals need blood banks to manage patients with severe hemorrhage.


INTRODUCTION
Worldwide, women of reproductive age die from the complications associated with pregnancy and child birth. More than half of these deaths occur within the first 24 hours following childbirth and are associated with excessive bleeding [1][2][3].
In Cameroon, the maternal mortality ratio has increased over the recent years, from 690 deaths /100 000 in 2010 to 789 deaths/100 000 live births in 2014 [5]. The main cause of maternal death was hemorrhage (38%) with primary postpartum hemorrhage being the leading cause of death accounting for 4.1% of cases [5]. Studies conducted in the General Hospital and the Yaoundé Central Teaching Hospitals in 2008 and 2013 reported prevalence of primary postpartum hemorrhage of 1.68% and 4.1% respectively [5,6]. Recently, the incidence of primary postpartum hemorrhage in the Yaoundé Central Teaching Hospital was reported at 13.9% in 2014 [7]. This indicates that the prevalence of primary postpartum hemorrhage has significantly increased over the past years.
According to World Health Organization, of the 275,000 women who die every year from pregnancy and childbirth, one-quarter of these deaths are due to primary postpartum hemorrhage. As such, primary postpartum hemorrhage remains a major health issue because the majority of maternal deaths (88%) occur within four hours of delivery despite all the progress made in the active management of the third stage of labor (AMTSL) [8].
It is important to know the circumstances under which hemorrhage occurs and the risk factors associated with it [3]. The risk of dying from PPH depends not only on the amount and rate of blood loss, but also the health status of the woman. Poverty, lifestyle and malnutrition are some of the broad issues that have unfortunately come to be accepted as inevitable and unchangeable but which influence the outcome of a patient with PPH [9]. The reality about PPH is that two-thirds of the women who experience it have no identifiable risk factors such as multiple births or fibroids [10]. The aim of this study was to determine the prevalence, risk factors, and maternal and fetal outcomes of primary postpartum hemorrhage in the Bonassama District Hospital (BDH), Cameroon.

Study Setting
This cross-sectional analytic study was carried out in the Maternity unit of the Bonassama District Hospital from 1 st December 2014 to 15 th April 2015. The Bonassama District Hospital is one of seven health districts hospitals that serve the cosmopolitan town of Douala with a population of 1,088,026 inhabitants [11].
The BDH is located in Bonassama Health District which is one of the thirty health districts in Littoral Region, Cameroon. The hospital was chosen because it offers obstetrics and gynecological services to the population. It has a surgical theatre, a blood bank and attends to women of all social classes. An average of 1790 deliveries is carried out in the maternity annually. The hospital has eight general practitioners and nine specialists including a gynecologist. The personnel of the maternity of BDH during the period of study was made up of a gynecologist, eight midwives, three student midwives and six paramedical personnel. The paramedical personnel are not trained in providing obstetric care but are responsible for the cleanliness of the wards, taking samples and collecting results from the laboratory.

Sample Size Calculation
The sample size was calculated using the formula for comparing two proportions [12]. For a confidence level of 95%, Z crit = 1.96, Z pwr= standard normal deviate for the desired statistical power of 80% = 0.80 The pre-study prevalence of adverse outcome of primary PPH was estimated at 13.4% [4] while the pre-study estimate of the prevalence of PPH in naïve parturient with no risk factors in Cameroon was 50% [4]. A minimum sample size of 456 pregnant women was required for the study. However, five hundred and fifty parturients were enrolled to make our results more justifiable.

Data Collection
For each potential participant, after selfintroduction by principal investigator or research assistant, the study was carefully explained to the participant, with emphasis on the importance, risks and benefits associated with participating in the study. Furthermore, the choice of participating was voluntary after the briefing. A convenient sampling procedure was used to select all participants who met the inclusion criteria (pregnant women who delivered at BDH, were >28 weeks and had a vaginal delivery). Those referred from other maternities with PPH were excluded. During the study period, 556 women were interviewed but only 550 were enrolled. The others declined for personal reasons.

General measures
The third stage of labor was actively managed in all the participants. Uterine massage, which is part of active management of the third stage of labor (AMTSL) was systematically performed on all the women over a period of 30 minutes to 1 hour in order to avoid uterine atony. The participants who had symptoms and signs of cardiovascular decompensation were initially resuscitated with crystalloids or colloids.

Specific measures
Uterine evacuations were performed in cases of PPH due retention of placenta debris. This procedure was performed either by a midwife or a gynecologist. Conservative management involved the administration of a utero-tonic. Other surgical interventions like the repair of genital lacerations were carried out depending on the extent of the lesions with absorbable sutures. All the 550 participants were prescribed hematinic. Furthermore, all the patients who had a uterine evacuation or obstetrical lacerations were administered antibiotics in order to prevent sepsis.

Diagnosis of primary PPH
Immediately after the cord was clamped and sectioned, the blood collection was started by putting a bed pan under the buttocks of the woman. Blood collection and measurement continued until cessation of the third stage of the labor. The parturient was subsequently transferred to the postnatal ward after surveillance for about an hour. Three main steps were involved in the quantification of blood loss immediately after delivery and within 24 hours following delivery.
a) Measuring the Mass of Blood (m b ) Lost: Firstly, it was ensured that the weighing units of our electronic machine were displayed and the weighing machine was calibrated. An empty jar was weighed, blood collected from the bed pan was subsequently poured into it, and it was reweighed. The mass of blood loss (mb) was the difference between the two weights. b) Calculating the volume of blood lost during the third stage and within one hour after delivery (V 1 ): The volume of blood lost from our direct measurements was calculated using the formula: where: ρ= density of blood (1,060kg/m 3) [13]. The blood lost by the participants while in the ward (V 2 ) was quantified using the visual estimation of blood loss by the Pathfinder International wall chart [14].
The participants had their sanitary pads on during the first 24hours and the visual estimation was done before they changed their pads or beddings. c) The total blood lost (V T ) was calculated by adding the intra-partum blood loss to that obtained from visual estimation. That is, The participants were also examined for symptoms and signs of cardiovascular decompensation and for possible causes of primary PPH. The diagnosis of PPH was based on a vaginal blood loss of ≥ 500cc within the first 24 hours after delivery.
The parameters of neonates; birth weight and Apgar score were obtained after the first, fifth and tenth minutes of life. Those whose conditions were unsatisfactory were reevaluated by the pediatrician. Neonates with APGAR scores <7/10 after the fifth minute were transferred to the neonatology unit for further management. .

Data Management
A pretested questionnaire was used to record data on socio-demographic variables, risk factors of PPH (previous caesarean section, past history of PPH, pregnancy induced hypertension, HIV etc.), obstetric history (parity, polyhydramnios, uterine over-distension, duration of labor etc.) from the medical records, and fetal and maternal outcomes.
Data entered was doubled checked for errors. It was cleansed and saved in the hard drive of the computer as a back-up, transcribed into EpiInfo TM 7.1.4.0 software and analyzed. In describing the socio-demographic and obstetric characteristics, measures of central tendencies (means and medians) were used. For categorical data, frequencies were computed. Chi-squared or Fisher's tests were used to compare proportions where appropriate to determine the strength of association between variables. The odds ratio of patients with PPH and risk factors compared with those without risk factors was also determined. A multiple logistic regression analysis controlling for confounding factors was also done. Risk factors were significantly associated with PPH if P-value<0.05.

Limitation of Study
Some of the paturients did not allow their used sanitary pads to be assessed subsequently in order to quantify the blood loss within the first 24 hours. This probably resulted in an under estimation of the post-partum blood loss in some cases. This error was minimized by reiterating to the participants the aim of this study and how their contribution will actually improve on patient care. Furthermore, direct quantification of the amount of blood loss during a vaginal delivery might not be accurate because we ignored the maternal blood in the placenta and the blood lost while transferring blood from the bed pan into the measuring cylinders. This was minimized by calculating the volume of blood lost in the form of clots using the density of blood. Though the results of the study cannot be generalized to what obtains in other regions, because of the small sample size, they however, partially reflect the picture of primary PPH in a typical health district hospital in Cameroon.

Prevalence of Primary Post-partum Hemorrhage
During this study period, 130 (23.6%) cases of primary postpartum hemorrhage were diagnosed.

Socio-demographic Characteristics the Study Population
The mean age of the 130 participants with primary postpartum hemorrhage was 24.5 (SD 4.6) years. The majority of the women 111 (85.4%) who had primary PPH were between 20-30 years. Most of the women who had PPH were self-employed 93 (71.5%). They were mainly traders, hair-dressers and business women. Although the occupation of the patient was significantly associated with primary PPH (P=.001), it was not a risk factor for PPH after controlling for confounders. Other socio demographic characteristics are shown in Table 1.

Clinical (Medical, Surgical and Gynecological) Characteristics
Based on the past medical history, 5 (3.9%) participants had chronic hypertension, 8 (6.2%) HIV and 3 (2.3%) had other medical conditions like asthma. Of the 130 women who had primary PPH, 13 (10%) had malaria during pregnancy. This was significantly associated with PPH (P<.001). Those who had malaria during pregnancy were 6 times more likely to develop primary PPH than those women who did not. Pregnancy induced hypertension (PIH) accounted for 5 (3.9%) of the cases of PPH while 9 (6.9%) were associated with premature of membranes. Unlike malaria, these were not significantly associated with primary PPH. The risk of primary PPH was higher in patients with a history of abortion, myomectomy and postpartum hemorrhage (P =.002). There were two cases of uterine fibroids in the current pregnancies.

Causes of uterine over -distension
As shown in Table 4, the causes of large for dates uterus or uterine over distension based on the ultra-sonographic findings during pregnancy were; multiple pregnancy, hydramnios, uterine fibroids and macrosomia. There were 10 (1.8%) cases of multiple gestations; six (4.6%) of them developed PPH. Furthermore, out of the eight cases of uterine over-distension due macrosomia, 5 (3.8 %) developed PPH. There was no significant association between these variables and primary PPH.

Induction of labor
Of the 550 women who gave birth at the BDH, 10 (1.8%) had induced labor indicated for post-term or premature rupture of membranes (PROM). Among these 10 women, 7 (70%) developed PPH. After multiple logistic regression analysis adjusting for confounding factors, women who had an induced labor were 4 times more likely to develop PPH than those who were not induced (O.R=4.12; aO.R 4.02; 95% C.I=3.2 -11.09).

Duration of labor
As shown inTable 5, of the 218 (39.64%) women who delivered within 12 hours, 12 (9.2%) developed PPH, with the extremes of the duration of labor ranging between 2 and 26 hours. Prolonged labor was associated with PPH, participants whose labor lasted more than 24 hours having a higher risk of developing PPH than when labor lasted less than 12 hours.

Qualification of the birth attendant
One hundred and twenty three (94.9%) of the 130 deliveries that resulted in PPH were carried out by midwives compared to 2 (3.7%) deliveries conducted by the gynecologist. The cases of labor conducted by the gynecologist included breech presentations and locked twins, conditions that are usually associated with genital trauma. There was a significant association between primary PPH and the qualification of the birth attendant (P<.001).

Multiple pregnancies, Birth weight and PPH
Of the eight (1.5%) women who had twin deliveries, 4 (50%) were diagnosed of PPH but there was no significant association between PPH and multiple pregnancy (P = .45). Unlike preterm delivery which occurred in 18 (4.3%) women without primary PPH, 10 (7.7%) participants who gave birth to macrosomic babies were predisposed to PPH (P = .02) ( Table 6).

Fig. 2. Qualification of birth attendant
Of the 558 neonates delivered in BDH, there were 2 (0.4%) perinatal deaths; a still birth and a neonatal death from neonatal asphyxia.

Maternal and Neonatal Outcomes of Patients with Primary PPH
One hundred and four (95.4%) women diagnosed of PPH were administered 1-1.5 liters of Ringer Lactate or 0.9% normal saline because of mild hypotension while 6 (4.6%) participants were clinically stable. Colloids were only administered to 17 (13.1%) women who presented with moderate to severe primary PPH. Forty-seven (33.8 %) participants had uterine evacuations, 40 (30.8%) because of retention of placenta debris. Conservative management involved the administration of a utero-tonic. One hundred and thirty-five (24.6%) women without underlying co-morbidity were administered an infusion of oxytocin flowing at 10IU/h for 4hours and then re-evaluated. Twelve (9.2%) patients, who did not respond to the first line of treatment with intravenous oxytocin, were administered 500 mcg of Ergometrine intramuscularly. Twenty one (16.2%) participants were administered 800 mcg of misoprostol rectally.
Among the 49 (37.7%) participants who had obstetrical lacerations, 40 (7.3%) had lacerations of the genital tract which were sutured. No hysterectomy was performed for primary PPH.

Fig. 3. Relationship between maternal outcome and PPH
Of the 130 (23.6%) cases of PPH, 4 (3.1%) were transfused whole blood while 2 (1.5%) women died from intractable bleeding during referral to a tertiary center giving a maternal mortality ratio of 361 deaths/100,000 live births.
Neonatal outcomes were; uneventful deliveries, a still birth and a neonatal death. The case of birth asphyxia was due to prolonged labor.

DISCUSSION
The aim of this study was to determine the prevalence, risk factors, and the maternal and neonatal outcomes of primary postpartum haemorrhage in the Bonassama District Hospital, Cameroon.

Prevalence of Primary PPH
The prevalence of primary PPH of 23.6% is higher than the 4.1% reported by Tebeu and colleagues in Yaoundé [5]. This shows that there has been a significant increase in the prevalence of postpartum hemorrhage in some hospitals in Cameroon. This also reflected by the significant rise in the maternal mortality ratio from 690 to 782 deaths /100 000 from 2006 to 2010 with PPH being the leading cause of maternal mortality [3]. The prevalence of Primary PPH in this study was comparable to the 25.7% reported by Calvert et al. [6] in Africa, but higher than the 21.5% and 14.3% reported in East Africa and Northern Nigeria respectively [16,17]. However, lower values were reported by some researchers in West Africa in similar studies [18]. This difference in the prevalence of PPH is probably due to the methods used in estimating blood loss after delivery and the setting of these studies. Prasertcharoensuk reported that visual evaluation of blood loss could lead to an underestimation postpartum blood loss by 89% [19].
The clinical importance of this underestimation cannot be overemphasized as it could explain why PPH is still associated with a high maternal morbidity and mortality worldwide because of late intervention strategies [18].

Uterine atony
Uterine atony was the most common cause of PPH in 81 (62.5%) cases. This value was lower than the 70% reported by Anderson et al. [20] but higher than that of Sango et al. in HGY where uterine atony accounted for 51.06 % of cases of primary PPH. On the other hand, our result was comparable to the 60.7% obtained by Coulibaly in Mali [21]. Uterine atony remains a major public health issue as it is the main cause of primary postpartum hemorrhage. The risk factors associated with uterine atony were: previous (myomectomy, PPH, abortion), multiparity, malaria in pregnancy, induced labor, prolonged labor and macrosomia as documented in other studies [4][5][6][7].
There was no significant association between PPH and advanced maternal age, polyhydramnios, HIV, pregnancy induced hypertension, post term pregnancy and multiple gestations. Furthermore, polyhydramnios was not associated with PPH which contrasted the findings of the study conducted in a tertiary hospital in Pakistan by Anderson [20]. He suggested that, this could have resulted from the increased incidence of polyhydramnios and increased birth weight in pregnancies complicated by gestational diabetes. Uterine surgeries such myomectomy were risk factors for PPH because a scared uterus forms a background for placental tissue adherence which led to placenta retention and subsequently atony [23].
In this study, of the 130 cases of PPH, 13 (12.9%) women presented with uterine atony and retained placenta. This was lower than the 21.3% obtained by Adelusi et al. [24] in Scandinavia.

Trauma
Obstetrical lacerations were the primary cause of PPH in 37.7 % of cases as compared to 20% in Anderson et al. [19] study. This is because the age related changes in the connective tissue decrease the ability of cervical, vaginal and perineal muscles to stretch as needed during delivery resulting in greater trauma to tissues, prolonged labor and diminished uterine contractility after delivery [24]. Genital tract trauma at delivery was associated with an odd ratio of 1.7 (95% C.I 1.4-2.1) for blood loss >1000 ml [6]. Similar results were found in a Dutch study with a reported odd ratio of 1.82 (95% C.I 1.01-3.28) for PPH (>1000 ml) with perineal tears (second and third degree tears) being the most common cause [25]. Trauma to the broad ligament, uterine rupture, cervical and vaginal tears and perineal tears are all associated with an increased blood loss at normal vaginal delivery. The most frequent cause of obstetric hemorrhage was cervical tears 25 (51.0%) followed by perineal tears 15 (10.6%). This result was higher than the 21.5% and 18.6% of cervical and perineal tears respectively reported in a health district in Mali in 2006 [20]. No case of episiotomy was associated with primary PPH in this study population. The risk factors for obstetric laceration were precipitate labor and macrosomia or neonates babies presenting with congenital malformations such as hydrocephalus. However, women who gave birth to fetuses with birth weight less than 2500 g were less prone to having primary PPH.

Maternal and Neonatal Outcomes
Of the 130 participants who were diagnosed with primary PPH, 119 (91.5%) had a normal recovery. Five (3.8%) had hypovolemic shock, 4 (3.1%) were transfused whole blood and 2 (1.5%) died from uncontrollable bleeding during referral to a tertiary center (giving a MMR of 361 per 100.000 live births). The prevalence of hypovolemic shock due to severe PPH was lower than the 7.1% reported by Sango in Hopital General Yaoundé [23]. Death probably resulted from the delay in blood transfusion because of the difficulty in obtaining compatible blood since the blood bank had limited reserves or from disseminated intravascular coagulopathy due to consumption of clotting factors as reported by Sango [23]. The 1.5% maternal death recorded in this study is high and doubles the 0.7% reported by Sango in Hopital General Yaoundé [23] because the latter study was carried out in a tertiary center with better facilities and trained personnel. Other reasons which could account for the high mortality in this study were; late referral, insufficient blood reserves in the blood bank and the poor conditions under which the patients were transported during referral.
Neonatal outcomes were mostly uneventful although a stillbirth and early neonatal death due to birth asphyxia occurred.

CONCLUSION
The prevalence of primary postpartum hemorrhage in Bonassama District Hospital was high. It was a major cause of maternal and neonatal morbidity, and mortality. Though the occurrence of primary postpartum hemorrhage is unpredictable, its prevention lies in the identification of the associated risk factors and proper management of the third stage of labor. Furthermore, maternal and fetal morbidity and mortality can be reduced by availability of trained personnel, surgical theatres and blood transfusion services.

CONSENT
Participants signed a consent form before delivery. For those who were not literate, consent was given after the participant thumb -printed a consent form after the study and what it entailed was explained to her in the language she best understood by a research assistant and interpreter. For those pregnant women who were less than 21 years (legal age at which a patient can validly provide consent in Cameroon) assent was obtained from their parents or guardians.

ETHICAL ISSUES
Ethical clearance for the study was obtained from Institutional Review Board of the Faculty of Health sciences, University of Buea. While administrative approval was obtained from the Regional Delegation of Public Health, Littoral Region and the Chief Medical Officer of the hospital.