Impact of Maternal Preoperative Hemoglobin Levels on Intraoperative Hemorrhage Risk in Placenta Accrete Spectrum Disorders: A Retrospective Cross-Sectional Study

Background

Placenta accreta spectrum (PAS) disorders are diseases in which all or part of the placenta tissue invades the maternal uterine myometrium to varying degrees, and the disorders encompass a heterogeneous group of anomalies characterized by an abnormal adhesion or invasion of chorionic villi through the myometrium and uterine serosa [1]. Irving and Hertig first described PAS disorders in 1937 as different degrees of villous tissue invading the muscle layer, and even the surrounding pelvic organs [2]. When placental tissue does not completely or partially spontaneously separate at delivery [3], it often causes life-threatening postpartum hemorrhage, resulting in multisystem organ failure, disseminated intravascular coagulation, need for admission to an intensive care unit, peripartum hysterectomy, and even death [4,5]. PAS disorders has always been a challenge for obstetricians to enhance clinical management, and their prevalence is increasing. Previous studies had suggested that planned delivery [6,7], reasonable time for pregnancy termination [8], delayed hysterectomy [9], and a variety of surgery approaches [10–13] for hemostatic control and minimizing adjacent organ injury reduced maternal mortality and morbidity in patients with PAS disorders.

A previous study demonstrated that women with gestational anemia are at an increased risk of severe postpartum hemorrhage during and after delivery [14]. However, there has been little research investigating the association between preoperative maternal hemoglobin levels and severe postpartum hemorrhage in PAS disorders. Therefore, we conducted a cross-sectional study to explore the relationship between the preoperative maternal hemoglobin level and the risk of intraoperative massive hemorrhage in pregnant women with PAS disorders, exploring potential etiological factors of intraoperative massive hemorrhage in PAS disorders.

Material and Methods

STUDY DESIGN AND PARTICIPANTS:

This cross-sectional retrospective study identified patients with discharge diagnosis of PAS disorders [15] (ICD-o72, ICD-o73), according to the codes of the International Classification of Disease, 10th revision. Data were extracted out of the electronic medical records at the Second Hospital of Hebei Medical University from January 1, 2014, to December 31, 2020. The basic clinical data were collected using a standard clinical audit protocol, and all data were fully anonymous before being submitted for analysis. This study was approved by the Research Ethics Committee of the hospital. The requirement for patient informed consent was waived because of the retrospective nature of this study.

EXCLUSION CRITERIA AND INCLUSION CRITERIA:

The patient exclusion criteria were as follows: (1) participants with internal and surgical diseases, such as diabetes mellitus or gestational diabetes mellitus, hypertensive disorder, thyroid diseases, hepatic diseases, heart disease, and acute pancreatitis; (2) stillbirth, fetal growth restriction, and placental abruption; (3) gestational age <28 weeks or >42 weeks; and (4) missing data on maternal preoperative hemoglobin level and intraoperative blood loss. The inclusion criteria of patients were as follows: (1) met the diagnose of PAS disorders according to FIGO classification for the clinical diagnosis of PAS disorders [15]; (2) gestational age 28 to 42 weeks; and (3) termination of pregnancy by cesarean delivery.

DATA COLLECTION AND MEASUREMENT:

EpiData v3.1 Software was used to extract and manage study data. The baseline characteristics of participants included age, occupation, region, body mass index (BMI), gravidity, and parity. Some potential confounders, such as weight changes, uterine anomalies, placental location, type of placenta previa, grade of PAS disorders, hemorrhage during pregnancy, assisted reproduction technology, and multiple pregnancies, were also collected. Intraoperative blood loss was measured according to cell protector and swab measurements from the surgical skin incision to the end of surgery within 2 h after delivery. Intraoperative massive hemorrhage was considered when the volume of postpartum hemorrhage exceeded more than 1500 mL within 2 h after surgery [16]. The level of preoperative maternal hemoglobin was determined from the last tested hemoglobin before delivery by a blood cell analyzer (UniCel DxH 800, Beckman Coulter). In addition, the location of placenta attachment and type of placenta previa were examined using the American GE-E10 color Doppler ultrasound by transabdominal or transabdominal combined with transvaginal scan before surgery [17]. BMI was calculated by the maternal height and the pregnant weight based on hospital inpatient report (kg/m2). The following variables were also collected: gestational age at delivery, as determined by last menstrual period and confirmed by ultrasound in all patients (week); uterine anomalies, including fibroid tumors or previous uterine surgery, such as dilatation and curettage, myomectomy, endometrial resection, or surgical abortion; hemorrhage during pregnancy with at least 1 episode of bleeding from the genital tract during the antenatal period; and grade of PAS disorders was defined according to the standardized terminology and clinical diagnosis on “placenta accrete spectrum disorders” in China [18].

STATISTICAL ANALYSIS:

Continuous variables were expressed as mean±SD (normal distribution) or median (quartile; skewed distribution), and categorical variables were expressed as a percentage. According to the recommendation of the STROBE statement [19], logistic regression was used to quantify unadjusted ORs and adjusted ORs and 95% CIs expressing the association between preoperative maternal hemoglobin level and the risk of intraoperative massive hemorrhage. We selected the potential confounders on the basis of their association with the outcomes of interest or a change in effect estimate of more than 10%. In addition, based on 5 replications and a chained equation approach method in the R MI procedure, the missing data of the covariates were imputed by multiple imputations. All of the analyses were performed with the statistical software package R (http://www.R-project.org, The R Foundation) and Empower Stats (http://www.empowerstats.com, X&Y Solutions, Inc, Boston, MA). P<0.05 was considered statistically significant.

Results

BASELINE CHARACTERISTICS OF PARTICIPANTS:

Among all pregnant women with caesarean delivery from January 1, 2014, to December 31, 2020, at the Second Hospital of Hebei Medical University, there were 1015 patients diagnosed with PAS disorders. Among them, 477 patients were excluded due to internal and surgical diseases (n=414), gestational age <28 weeks or >42 weeks (n=62), and missing data (n=1). Ultimately, a total of 538 consecutive patients met the inclusion criteria (Figure 1). The demographics and clinical characteristics of all participants are shown in Table 1. The average age of the participants was 31.12±4.68 years, and approximately 23.61% of them were older than 35 years. The mean level of preoperative maternal hemoglobin was 10.99±1.39 g/dL, the overall incidence of anemia (hemoglobin <11 g/dL) was 48.88%, and the incidence of intraoperative massive hemorrhage was 38.66%.

UNIVARIATE ANALYSIS:

The univariate analysis showed that age, parity, number of cesarean section, placental location (posterior, lateral, anterior or central), type of placenta previa, grade of PAS, hemorrhage during pregnancy, and antepartum anemia were positively associated with the risk of intraoperative massive hemorrhage. Univariate analysis also showed that the grade of PAS disorders was significantly associated with the risk of intraoperative massive hemorrhage. While, age ≥35 years, occupation, BMI, weight changes, gravidity, uterine anomalies or not, and multiple pregnancy were not associated with the risk of intraoperative massive hemorrhage. In contrast, urban or not (OR=0.60, 95% CI 0.41–0.87), gestational age at delivery (OR=0.82, 95% CI 0.76–0.89), vaginal delivery or not (OR=0.29, 95% CI 0.164.53), assistant reproduction technology (OR=0.45, 95% CI 0.21–0.97), and the level of preoperative maternal hemoglobin (OR=0.64, 95% CI 0.55–0.73) were significantly negatively correlated with the risk of intraoperative massive hemorrhage (Table 2).

UNIVARIATE LINEAR REGRESSION MODEL:

The univariate linear regression model was used to evaluate the association between the level of preoperative maternal hemoglobin and the risk of intraoperative massive hemorrhage. The original model, minimally adjusted model, and fully adjusted model are shown in Table 3. In the non-adjusted model, the level of preoperative maternal hemoglobin had a significant negatively correlation with the risk of intraoperative massive hemorrhage (OR=0.64, 95% CI 0.55–0.73, P<0.0001). In the minimally adjusted model (adjusted for gestational age at delivery, parity, type of placenta previa, and grade of PAS) and the fully adjusted model (adjusted for age, region, gestational age at delivery, gravidity, parity, vaginal delivery, uterine anomalies, assistant reproduction technology, placental location, type of placenta previa, hemorrhage during pregnancy, and grade of PAS), the changes of the effect were not obvious (OR=0.66, 95% CI 0.55–0.79, P<0.0001 and OR=0.65, 95% CI 0.53–0.79, P<0.0001, respectively). For sensitivity analysis, the level of preoperative maternal hemoglobin was a categorical variable; the fully adjusted ORs for patients in tertile 2 and tertile 3, when compared with those in tertile 1, were 0.35 (95% CI 0.19–0.64; P<0.001) and 0.31 (95% CI 0.16–0.58; P<0.001), respectively.

NON-LINEAR ANALYSIS AND 2-PIECEWISE LINEAR REGRESSION MODEL:

The relationship between the level of preoperative maternal hemoglobin and the risk of intraoperative massive hemorrhage was non-linear after adjusting for age, region, gestational age at delivery, gravidity, parity, vaginal delivery, uterine anomalies (fibroid tumors or previous uterine surgery or surgery abortion), assisted reproduction technology, placental location, type of placenta previa, hemorrhage during pregnancy, and grade of PAS (Figure 2). Moreover, a 2-piecewise linear regression model was used to calculate the inflection point (11.5 g/dL; Table 4). When the level of preoperative maternal hemoglobin was <11.5 g/dL, the preoperative maternal hemoglobin level increased by 1 g/dL, and the adjusted risk of intraoperative hemorrhage was reduced by 0.48 (OR=0.52; 95% CI 0.39–0.70; P<0.0001). However, there was no significant association between the level of preoperative maternal hemoglobin and the risk of intraoperative massive hemorrhage (OR=1.06, 95% CI 0.64–1.75 and P=0.7602) when the level of preoperative maternal hemoglobin was ≥11.5 g/dL.

Discussion

Our study showed that the median intraoperative blood lost was 1100 mL (range, 200–9850 mL) and the incidence of the intraoperative massive hemorrhage was 38.66%, which was slightly lower than that of reported studies [20]; the reason may be the current multidisciplinary approach and ongoing process improvement in the management of this disease. This study also provided some new insights. Lower preoperative maternal hemoglobin concentration, as a changeable factor, was found to be a predictor for intraoperative massive hemorrhage in PAS disorders. In this retrospective, cross-sectional study, we found that the correlation was non-linear between the level of preoperative maternal hemoglobin and the risk of intraoperative massive hemorrhage after adjusting for age, region, gestational age at delivery, gravidity, parity, vaginal delivery, uterine anomalies, assisted reproduction technology, placental location, type of placenta previa, hemorrhage during pregnancy, and grade of PAS disorders. Moreover, when the preoperative hemoglobin level of pregnant women was lower than 11.5 g/dL, there was a significant negative correlation between the preoperative hemoglobin level and risk of intraoperative massive hemorrhage in PAS disorders. However, when the preoperative maternal hemoglobin level was more than 11.5 g/dL, there was no significant correlation.

The presence of PAS disorders is correlated with the incidence of maternal diseases, such as life-threatening massive hemorrhage, which leads to secondary complications, including multisystem organ failure and hysterectomy [5,21–23]. Consequently, it is particularly important to optimize the preoperative hemoglobin levels in PAS disorders in order to improve the tolerance of intraoperative bleeding. According to evidence-based guidelines for the management of abnormally invasive placenta recommendations by the International Society for Abnormally Invasive Placenta, if the level of hemoglobin is lower than 11 g/dL before 28 weeks of gestation or lower than 10.5 g/dL after 28 weeks of gestation, iron supplementation should be given to optimize the patient’s hemoglobin level before surgery [24]. Retrospective cohort studies have reported that patients with anemia had a higher risk of postpartum hemorrhage than those without prenatal anemia [20,25]. Also, a meta-analysis suggested that severe prenatal anemia increased postpartum hemorrhage risk [26]. In the present study, we found similar results, namely that anemia was associated with a higher rate of intraoperative massive hemorrhage, and the correlation between the preoperative maternal hemoglobin level and the risk of intraoperative massive hemorrhage was a non-linear relationship [14]. However, a prospective cohort study by Parks et al [27] noted that only women with severe anemia had a significantly higher rate of postpartum hemorrhage, not found in mild and moderate anemia. The possible reason may be that our patients had lower tolerance for postpartum hemorrhage than those in the previous study. A slightly lower hemoglobin level has been found to be negatively associated with severe postpartum hemorrhage even without severe anemia.

The mechanisms underlying the association are not yet very clear. Soltan et al reported that women with anemia are prone to have postpartum hemorrhage, possibly due to uterine systolic fatigue caused by the raised nitric oxide levels [28]. A previous study also found a potential role that an impaired coagulation profile in anemic pregnant women could lead to postpartum hemorrhage [29]. Another possibility was the presence of low-grade prenatal fibrinolysis in severe anemic pregnant women, which could increase postpartum hemorrhage during delivery [30]. Further, more basic experimental studies are required to investigated if this connection is causal and to identify the underlying mechanism.

There were some limitations in this study. First, the study was a retrospective, cross-sectional observational study conducted in patients with PAS disorders at a single center; therefore, it reflected only the clinical characteristics and pregnancy outcomes of PAS disorders in this study population. Second, we did not review the use of medicines such as oxytocin, misoprostol, prostaglandin, and tranexamic acid in the operation and the surgical methods. It is noteworthy that the potential exposure factors resulting from such errors would bias the results toward a reduced occurrence of postpartum hemorrhage, and thus the results are an underestimation of the association between maternal preoperative hemoglobin concentration and massive intraoperative hemorrhage. Further, we should conduct multi-center clinical studies or prospective cohort studies to reduce selection bias in the future.

Conclusions

In conclusion, there was a non-linear relationship between preoperative maternal hemoglobin level and intraoperative massive hemorrhage; namely, the maternal preoperative hemoglobin level was significant negatively related to the risk of massive intraoperative hemorrhage when the hemoglobin was less than 11.5 g/dL. Given the high prevalence of PAS disorders, the result of the study might be of significance for clinicians to enhance antenatal care, and could help obstetricians in their evaluations and planning treatment strategies for patients with PAS disorders to reduce adverse pregnant outcomes.