Poor Uterine Contractility as Maternal Predisposition to Postpartum Hemorrhage Among Low-Risk Women: An Analysis of Large-Scale Database

Background Postpartum hemorrhage (PPH) is a major cause of maternal mortality. Even seemingly low-risk women may suddenly develop postpartum hemorrhage, and these cases are often unpredictable and threatens maternal life. We hypothesized that innate poor uterine contractility may be a risk factor of PPH. In this study, we examined the association between the innate poor uterine contractility, suggested by the characteristics of labor and PPH in obstetrically low-risk women. Methods We used the Japan Perinatal Registry database of the Japan Society of Obstetrics and Gynecology registered in 2013–2016. With exclusion of women with well-known risk factors for PPH (maternal basal disease of hematologic disease, uterine leiomyoma, pregnancy by assisted reproductive technology, placenta abruption, placenta accrete, low-lying placenta, hypertensive disorders of pregnancy, macrosomia, polyhydramnios, or epidural analgesia), we analyzed 174,082 primiparous women who had one live singleton birth via vaginal delivery in cephalic presentation at 37 weeks’ gestation. Information about abnormal labor patterns (hypotonic uterine dysfunction, prolonged labor, and arrest of labor) diagnosed by obstetricians were also used in this study. In order to focus on innate maternal poor uterine contractility, we classied subjects into four classes according to whether they were diagnosed with abnormal labor patterns and whether they had used uterotonics, and Odds ratios and their 95% CI were calculated as well.

70%-80% of PPH cases [13,14]. Seemingly obstetrically low-risk women who do not have any of the above-mentioned risk factors may suddenly develop atonic PPH. These cases are di cult to predict and can lead to maternal death, which is a worrisome concern for both clinical mothers and obstetricians.
They become especially critical in understaffed clinics with poor medical resources, where beginning blood transfusion immediately is di cult [15].
Many previous studies reported that induction and augmentation of labor are associated with PPH [16,17]. Furthermore, Grotegut et al. [18] found an association between women with severe PPH secondary to uterine atony and greater amounts of oxytocin exposure. Medical indications for induction of labor include prevention of the risks associated with the prolongation of pregnancy, such as preeclampsia, fetal growth restriction (FGR), oligohydramnios, premature rupture of membrane (PROM), or the absence of spontaneous labor near post-term pregnancy. Similarly, the medical indications for the augmentation of labor include hypotonic uterine dysfunction and prolongation of the latent phase. Thus, the absence of spontaneous labor at term, hypotonic uterine dysfunction, and prolongation of latent phase may be caused by innate poor uterine contractility in the pregnant women. Therefore, we hypothesized that innate poor uterine contractility may be a risk factor of PPH, even in seemingly low-risk cases without the abovementioned well-known risk factors. To date, no studies have been focused on innate uterine contractility to verify its association with the risk of PPH. In this study, we analyzed the association between the innate poor uterine contractility (suggested by the characteristics of labor) and PPH in obstetrically lowrisk women, using the data from the Japan Perinatal Registry database.

Methods
In this study, we used the Japan Perinatal Registry database, which is managed by the Japan Society of Obstetrics and Gynecology (JSOG). The database is a national collaboration of obstetric departments that started in 2001 and includes live births and stillbirths at 22 weeks or later. Standardized formats were given to obstetricians to enter the data, including information about maternal characteristics, maternal basal disease, complications of pregnancy, delivery characteristics, and neonatal outcomes, of all who gave birth at their own facility. In 2013, contents registered in the database were signi cantly changed to be more detailed clinically than before. Therefore, we used the data of 890,652 registered We used data of maternal characteristics (maternal age; history of treatment for infertility; height; body weight before pregnancy; gestational weight gain), delivery characteristics (gestational age and neonatal birth weight), medical interventions for delivery (mechanical methods for ripening cervix, which include balloon catheter or laminaria tents; uterotonics use; instrumental delivery, which include forceps or vacuum extraction) and pregnancy complications (gestational diabetes mellitus (GDM), FGR, nonreassuring fetal status (NRFS), PROM and intra amniotic infection (IAI) ). Furthermore, information of abnormal labor patterns (hypotonic uterine dysfunction, prolonged labor, and arrest of labor) diagnosed by obstetricians were also used in this study. These abnormal labor patterns were diagnosed according to the following de nition by JSOG. 1) Hypotonic uterine dysfunction: when the intrauterine pressure is less than 10 mmHg at the cervical dilatation of 4 to 8 cm and less than 40 mmHg at the cervical dilatation of 9 cm in the second stage of labor. It is also used as a name of clinical presentation and refers to a condition in which labor does not progress due to inadequate uterine contraction. 2) Prolonged labor: when the baby does not deliver within 30 hours in the nulliparous women and 15 hours in the multiparous women, after the contraction cycle is within 10 minutes. 3) Arrest of labor: after the onset of labor, labor is progressed once, but labor has not progressed (no cervix dilation or descent the fetal head) for more than 2 hours despite the same uterine contraction. These data conform to the uniform coding speci cations and have passed rigorous quality checks.
PPH was de ned as having a blood loss ≥ 1,000 mL within 2 hours after delivery or requiring the use of a blood transfusion.
Maternal characteristics, delivery characteristics, and pregnancy complications were compared between PPH cases and controls who were not PPH, which were statistically assessed by using chi-square tests for categorical variables and t-test for continuous variables. Odds ratios (ORs) and their 95% con dence intervals (95% CIs) were calculated. Moreover, multivariable ORs were also calculated by logistic regression adjusted for maternal age, treatment for infertility, pre-pregnancy BMI, gestational weight gain, gestational age, neonatal birth weight, cervical ripening, GDM, FGR, NRFS, PROM, and IAI. A p value ≤ 0.05 was considered to indicate statistical signi cance. Then, in order to focus on maternal innate poor uterine contractility, we classi ed subjects into four classes according to whether they were diagnosed abnormal labor patterns and whether they had used uterotonics, and ORs and their 95% CI were calculated as well. For multiple comparisons, a Bonferroni correction was made, and statistical signi cance was de ned as a value of P < 0.0167 (Bonferroni correction as 0.05 / 3 tests). Statistical analyses were performed using IBM SPSS Statistics, version 25.0 (Armonk, NY: IBM Corp.).

Results
There were 10,508 PPH cases (6.0%) among a total of 174,082 low-risk women analyzed. Maternal characteristics, labor/delivery characteristics, abnormal labor patterns and pregnancy complications are shown in Table 1. Chi-square tests were conducted to compare data of controls and PPH cases. Maternal age ≥ 35 years (25.2% vs 19.8%, P < 0.001), infertility treatment (9.7% vs 7.9%, P < 0.001) and prepregnancy BMI ≥ 30 (4.1% vs 1.8%, P < 0.001) were signi cantly more likely in women with PPH cases than controls. Except for PROM, there were signi cant differences between women with PPH and controls in all variables of labor/delivery characteristics and pregnancy complications. Among them, only FGR was signi cantly lesser in controls (1.3% vs 2.6%, P < 0.001). Table 2 shows the crude and adjusted ORs and 95% CI for the risk of PPH. In abnormal labor patterns, hypotonic uterine dysfunction (adjusted OR 1.28, 95% CI 1.22 to 1.34), prolonged labor (adjusted OR 1.41, 95% CI 1.30 to 1.52) and arrest of labor (adjusted OR 2.02, 95% CI 1.78 to 2.29) were signi cantly associated with the risk of PPH. The three medical interventions were also signi cantly associated. Table 3 shows the logistic regression analysis for risk of PPH associated with four classes classi ed by diagnosis of abnormal labor patterns and uterotonics use. Compared to women who were not diagnosed with any abnormal labor patterns and did not use any uterotonics, women who were diagnosed with abnormal labor patterns were at a signi cantly increased risk for PPH regardless of uterotonics use. These association were not eliminated by the adjusted multivariable logistic regression analysis (not use uterotonics: adjusted OR 1.23, 95% CI 1.10 to 1.37, and use uterotonics: adjusted OR 1.30, 95% CI 1.23 to 1.37). Women who were not diagnosed with any abnormal labor patterns, but used uterotonics, were also at an increased risk of PPH (adjusted OR 1.36, 95%CI 1.29 to 1.43).

Discussion
The present results suggest that women with abnormal labor patterns are at a risk of PPH, regardless of uterotonics use. We assumed that innate poor uterine contractility lead to the reported abnormal labor patterns: hypotonic uterine dysfunction, prolonged labor, and arrest of labor.
Previous studies have reported that longer duration of labor, particularly during the second stage of labor, induction or augmentation of labor, or instrumental delivery, were independent risk factors for PPH [4,7,16,19,20]. Another study reported that women with PPH secondary to uterine atony were exposed to oxytocin for a longer duration of time and required higher maximal oxytocin dosing for delivery [18]. Additionally, one study reported that repeated dinoprostone vaginal insets increase the risk of PPH [21].
These previous reports were, in part, consistent with our ndings. However no previous studies have focused on the innate uterine contractility and its role in PPH.
Our study is novel in that we examined the risk factor for PPH by separated according to the use of uterotonics and the diagnosis of abnormal labor patterns. In previous studies, it is unclear if the provided risk factors include both medical indications and medical interventions. For example, women with a longer duration of labor will be augmented by uterotonics. If the uterine contractility is still insu cient after the administration of uterotonics, more uterotonics may be required. It might be di cult to judge which one causes PPH. Therefore, medical indications and medical interventions should be considered separately.
The present results showed that women who were diagnosed with abnormal labor patterns in the rst or second stage of labor had an increased risk of PPH, regardless of uterotonics use. This suggests that the administration of uterotonics alone might not be responsible for PPH. We hypothesized that an innate uterine dysfunction was underlying the abnormal labor patterns that were associated with PPH.
Uterine contraction in the third stage for postpartum hemostasis depends on two kinds of hormones, oxytocin and prostaglandins, as in the rst and second stage. At the time of labor, oxytocin and prostaglandins bind to cell-surface receptors, causing depolarization of the uterine myocyte, which in turn promotes the opening of ligand-regulated calcium channels. Increased intracellular calcium activates myosin-light chain kinase, which activates myosin. Consequently, myocyte contractility is activated [23].
After delivery, myometrial bers surrounding the maternal spiral artery of the placental bed contract owing to the oxytocin and prostaglandin stimulations, and then myometrial contraction compresses the spiral arteries and veins. As a result of myometrial contraction, the uterine walls strongly oppose one another. It is thought that these mechanisms regulate hemostasis in the third stage of labor [24]. During labor, oxytocin is released in pulses, increasing in both the frequency and duration of pulse as labor progresses [25]. Additionally, myometrial oxytocin receptor (OXTR) density increases throughout pregnancy [26]. Prostaglandins are potent stimulators of myometrial contractility, and among many prostaglandin classes, PGE2 and PGF2 are involved in uterine contractions. Large amounts of prostaglandins are released in the third stage of labor, and plasma levels of PGF2 reach their maximum and start to decline within 10 minutes after placental separation [27]. These studies suggested that prostaglandins play an important role in securing hemostasis by way of myometrial contraction in the third stage of labor. In women with clinically abnormal labor patterns, we assumed that there is some dysfunction in the above described mechanisms. Namely, uterine contractility might be poor in the rst two stages of labor (sometimes uterotonics may be required). Subsequently, in the third stage of labor, the myometrium intrinsically may not be able to contract, fails to respond to hemostatic regulation, which may result in PPH. Some pathogenic dysfunction may also play a role, including such as the synthesis dysfunction of oxytocin and prostaglandins or decreased sensitivity or expression of OXTRs, FP receptors, EP1 receptors, or EP3 receptors. Reinl et al. [28] have suggested that some OXTR gene variants are associated with increased oxytocin responsiveness during labor. Grotegut et al. [29] reported that the genetic variation in OXTR and gene encoding G protein-coupled receptor kinase-6, which regulates desensitization of the OXTR, is associated with the amount of oxytocin required as well as the duration of labor.
Uterine fatigue and lactate accumulation in myometrium might also contribute to uterine dysfunction.
Under muscle contraction, lactic acid is produced by glycolysis in all human cells. Under myometrium contraction during labor, lactate is produced, which leads to intracellular acidi cation. Intracellular acidi cation inhibits the calcium channels in myometrial cells and decreased intracellular calcium will lead to a weakening of the myometrial contraction [30,31]. Thus, women diagnosed with abnormal labor patterns might include cases of lactate-induced uterine dysfunction.
Our ndings suggested that women used uterotonics even in the absence of diagnosis of abnormal labor patterns had increased the risk of PPH. Aside from abnormal labor patterns, indications for use of uterotonics include HDP, FGR, PROM, post-term pregnancy, or prevention of macrosomia. Unfortunately, there are no indications for use of uterotonics in this database. In the obstetric clinical situation, oxytocic agents are routinely used in the rst two stages. Previous studies have reported that prolonged or large amounts of oxytocin exposure, which may lead to the downregulation of OXTRs and failed uterine involution, results in PPH [6]. OXTR desensitization has been demonstrated in in vitro studies [32][33][34].
There is no consensus on whether prostaglandin use is a risk factor for PPH, however some studies have demonstrated an association between prostaglandin use and an increase risk for PPH [17,35,36]. One possible reason for increased PPH risk in women who used uterotonics without abnormal labor patterns may be due to the above-mentioned mechanisms.
In this study, we could use true low-risk subjects extracted from a large-scale clinical database certi ed by JSOG, de ned as primiparous women who had live birth to singleton fetus via vaginal delivery in cephalic presentation at 37 weeks' gestation. For caesarean delivery, the amount of bleeding has been reported to be almost twice that of vaginal delivery [12,14]. Some reasons for this result include: the effects of anesthesia, bleeding from a myometrial incision, active bleeding persisting after suture of the incision, or the effects of surgical procedures. Additionally, the amount of bleeding might include the amniotic uid volume. Therefore, PPH after vaginal delivery and PPH after caesarean delivery should be assessed separately. Furthermore, we excluded women who had well-established risk factors for PPH, including coagulopathy, uterine leiomyoma, placental abruption, placenta accreta, low-lying placenta, polyhydramnios and HDP, which could be determined within the antenatal period [6,9,14,17].
HDP is an independent risk factors for PPH, as previous studies have demonstrated [37,38]. Particularly in preeclampsia, patients are usually hypovolemic, so they are more likely to receive blood transfusions than in those with a normal pregnancy [39]. In addition, angiogenic factors, which are strongly associated with an increased risk of preeclampsia, also play an important role in the clotting system through the activation of coagulation and vascular thrombosis [40]. Positive association between angiogenic factors and postpartum bleeding has been reported [37]. This is because women with HDP were excluded from this study.
Additionally, we excluded women who were administered epidural analgesia during labor. A metaanalysis of randomized trials comparing epidural with non-epidural analgesia in women in labor reported that epidural analgesia prolongs the second stage of labor and increases the likelihood of requiring an instrumental vaginal delivery [41]. The use of epidural analgesia may change the course of labor from the course of a woman's natural uterine contractility. On the other hand, the prolonged labor itself is also an indication for epidural analgesia. It is di cult to judge which of the above factors individually caused PPH. In Japan, only a few obstetric hospitals perform epidural analgesia, thus the rate of women who receive epidural analgesia during labor in Japan was only 6.1% [42]. Therefore, for generalizability in obstetric hospitals, women received epidural analgesia were excluded from this study.
In this study, we were able to use a su cient sample size even after strict extraction and adjustment for several other risk factors for the analysis because of a large-scale database. There are no studies in Japan that have investigated the risk factors for PPH using such a large sample size. Furthermore, contrary to previous studies [18,43], detailed information about the type of uterotonics allowed us to investigate their effects on PPH according to the type and number of uterotonics required during delivery. after the birth progress regardless of route of delivery. In this database, however, the amount of bleeding within only 2 hours after delivery was registered and signs or symptoms of hypovolemia was not registered. Therefore, women with blood loss of ≥ 1,000 mL or women who received blood transfusion were de ned as PPH cases to compensate for the above missing information.

Conclusions
Our ndings suggest that women with innate poor uterine contractility are predisposed to PPH wven with no other well-known risk factors, such as multiple pregnancies, polyhydramnios, preeclampsia, and hematologic disease.
Further studies are necessary to identify practical methods to predict the innate poor uterine contractility, such as genetic variants, which may make it possible to reduce the maternal mortality due to PPH.

Declarations
Ethics approval and consent to participate The study protocol was reviewed and approved by the Bioethics Committee of Dokkyo Medical University (reference number: Daigaku 30014) and clinical research review subcommittee of JSOG (reference number: 92). The requirement for obtaining written informed consent from the patients was waived because the study utilized anonymized data from a database.

Consent for publication
Not applicable

Availability of data and materials
The data that support the ndings of this study are available from the Japan Society of Obstetrics and Gynecology (JSOG) under license for the current study. Restrictions apply to the availability of these data and thus are not publicly available. Data are, however, available from the authors upon reasonable request and with permission from the JSOG.

Competing interests
The authors declare that they have no competing interests.

Funding
This work was supported by JSPS KAKENHI Grant Number JP 17K11249.
Author' contributions