Int J Clin Exp Med 2019;12(11):13023-13031 www.ijcem.com /ISSN:1940-5901/IJCEM0098885 Original Article A risk factor analysis and prediction model of placental abruption Kehua Huang1*, Jianying Yan1, Xiaoling Li2*, Xiaoqian Lin1, Qinjian Zhang1, Jinying Luo1, Tingting Yang1 1Department of Obstetrics and Gynecology, Fujian Provincial Maternity and Children’s Hospital, Affiliated Hospital of Fujian Medical University, Fuzhou, Fujian, China; 2Graduate School of Fujian Medical University, Fuzhou, Fujian, China. *Co-first authors. Received June 24, 2019; Accepted September 4, 2019; Epub November 15, 2019; Published November 30, 2019 Abstract: Objective: To investigate the risk factors for the incidence of placental abruption and establish a model for placental abruption prediction. Methods: We retrospectively analyzed the clinical data of 303 placental abruption patients who were hospitalized and delivered in the Department of Maternity of Fujian Maternal and Child Health Hospital from January 2017 to December 2017. Univariate and multivariate logistic regression analyses were used to analyze the risk factors for placental abruption. The multivariate logistic regression OR method was used to preliminarily establish the “placental abruption prediction model”. Results: (1) The multivariate logistic regression analysis showed that advanced maternal age, multigravida, irregular prenatal care, hypertensive disorder compli- cating pregnancy (HDCP), premature rupture of membranes (PROM), fetal growth restriction (FGR) and anemia are independent risk factors for the incidence of placental abruption. (2) The “placental abruption prediction model” predicted that the AUC of the placental abruption under the ROC curve is 0.777. When the total score reached the critical value of 5 points, the prediction of placental abruption sensitivity was 67.3%, and the specificity was 74.3%. Conclusions: Placental abruption is closely related to advanced maternal age, multigravida, irregular prenatal care, HDCP, PROM, FGR and anemia. The “placental abruption prediction model” has a certain predictive value for the incidence of placental abruption. Keywords: Placental abruption, predictive model, risk factors, maternal and child outcomes Introduction mainly based on the degree of clinical manifes- tations and laboratory test results that can Placental abruption is the premature separa- be measured in the mother and fetus [3]. The tion of part of the placenta after 20 weeks of clinical manifestations and signs of grade II gestation or at normal parturition before the and III placental abruptions are more typical, delivery of the fetus [1]. Placental abruption and the diagnosis is generally no more diffi- can cause maternal hemorrhagic shock, dis- cult, but grade 0 and I placental abruption seminated intravascular coagulation, postpar- symptoms and ultrasound images are not typi- tum hemorrhage, and maternal death, which cal and can easily lead to missed diagnoses or can increase adverse neonatal outcomes. The misdiagnoses [4]. The current need for con- cause of placental abruption may be related to firmed placental abruptions is based on the decreased trophoblastic invasion, poor placen- pathological diagnosis seen during or at the tal function and hypoperfusion, chronic hyper- time of delivery or delayed delivery. The pro- tension, decreased intrauterine pressure, intra- gnosis of placental abruption is closely relat- uterine hypoxia, and external force injuries, ed to its early diagnosis and timely manage- which are associated with placental decidual ment. Therefore, it is particularly important to hemorrhage [2]. China clinically uses the pla- look for early risk factors for placental abru- cental abruption classification criteria for the ption, prediction or early detection, and to assessment and judgment of the condition. provide timely treatment of placental abrup- The grading standards from grade 0-III are tion. Analysis and model of placental abruption Methods placental abruption, clinical grade, gestational weeks and methods, labor or duration of sur- Patient information gery, prevention of postpartum hemorrhage, medications, placental and fetal membrane The subjects of the study were all selected fr- residuals, etc.), postpartum-related conditions om among the pregnant women at the Fujian (range of placental margin pressure area, uter- Provincial Maternal and Child Health Hospital ine bleeding, postpartum hemorrhage, blood affiliated to Fujian Medical University (Fujian transfusion, admission to ICU, pathological Severe Critical Referral and Newborn Care features), and perinatal outcomes. Center) from January 1, 2017, to December 31, 2017. The information of 16,397 people was The criteria for the diagnosis of placental ab- retrieved from the patients’ electronic medical ruption (cf. Williams Obstetrics (24th edition)) records and examination results from the hos- [1] included the placental abruption grading pital information center. Inclusion criteria: cas- standard (refer to the clinical diagnosis and es of placental abruption that were clinically or treatment specification of placental abruption pathologically confirmed, cases that met the (first edition) [unified standard]) [5]. criteria for the diagnosis of placental abrupti- on, and the provision of informed consent. The analysis used the dependent variable for Exclusion criteria: 1) Systemic factors: coagu- patients with placental abruption, followed by a lopathy, vascular disease, autoimmune diseas- univariate analysis to screen for significant fac- es, etc.; 2) Obstetric hemorrhagic disease: pla- tors affecting placental abruption. A univariate centa previa, placenta accreta, threatened ut- analysis was used to screen the factors that erine rupture, rupture of anterior vasculature, had a significant effect on placental abruption, placental tumor, etc.; 3) Non-obstetric-induced and a multivariate logistic regression analysis bleeding: vaginal disease, cervical lesions, sub- was carried out to obtain the independent in- mucosal uterine fibroids, etc.; 4) Incomplete fluencing factors and the OR value of each fac- records of available case data. A total of 303 tor. According to the rounding of the OR values, patients with placental abruption were includ- a prediction model of placental abruption was ed in this study as a case group (excluding 7 established. Then, the total score was calcu- patients with placental abruption who met the lated for each patient based on the predictive exclusion criteria: pre-placental status in 3 model, which was the sum of the scores for cases, presumed placenta in 1 case, and miss- each category. Using this score, the receiver ing data in 3 cases). Using a 1:1 retrospective operating characteristic (ROC) curve was dr- case-control study design, another 303 ran- awn, the best demarcation value was desig- domly selected patients who did not have pla- nated, and the sensitivity and specificity were cental abruption during the same period were calculated to evaluate the validity of the pre- used as controls. diction model. Research design Statistical method The maternal clinical data included the follow- All data were processed using the SPSS 18.0 ing: general maternal status (age, parity, num- statistical software package; normal distribu- ber of caesarean sections, family history), pre- tion measurement data were expressed_ as the natal examination (ART, number of prenatal mean ± standard deviation ( x ± s); the mea- care visits, first gestational weeks, and prena- surement data were compared using a group t tal related laboratory tests), maternal compli- test or a t’test, and the count data were com- cations (chronic hypertension, diabetes, hyper- pared using a chi-square test or a corrected thyroidism, hypothyroidism, anemia, uterine chi-square test. Univariate and multivariate fibroids, uterine malformations, ovarian mass- logistic regression analyses were used to ana- es, twin pregnancy, HDCP, gestational diabe- lyze the risk factors for placental abruption. The tes, polyhydramnios, oligohydramnios, FGR, AUC of ROC curve evaluates the prediction ef- PROM, intrauterine infections, ruptured uterus, ficacy of the “placental abruption prediction peripartum hysterectomy, acute renal failure, model”. The Youden Index determines the opti- PPH, etc.), childbirth (clinical manifestations of mal predictive value of the “placental abruption 13024 Int J Clin Exp Med 2019;12(11):13023-13031 Analysis and model of placental abruption Table 1. Demographic and obstetric characteristics of the study participants Maternal characteristics Placental abruption n (%) No placental abruption n (%) P value Maternal age (years) <35 231 (76.2) 258 (85.1) 0.005 ≥35 72 (23.8) 45 (14.9) Parity (number) 0 135 (44.6) 171 (56.4) 0.003 ≥1 168 (55.4) 132 (43.6) Abortions (number) 0 289 (95.4) 291 (96) 0.688 ≥1 14 (4.6) (4 (4) Hospitalization time (days) <5 214 (70.6) 236 (77.9) 0.041 ≥5 89 (29.4) 67 (22.1) Medical expense (thousand RMB) <8 127 (41.9) 178 (58.7) 0.000 ≥8 176 (58.1) 125 (48.3) Delivery method 0.001 Vaginal delivery 173 (57.1) 214 (70.6) Caesarean section 130 (42.9) 89 (29.4) Weeks gestation at childbirtha 34.86±4.71 38.93±1.68 0.00 Umbilical cord length (cm)a 51.02±11.51 58.54±12.57 0.00 Neonatal body mass (g)a 2404.72±897.07 3214.80±441.00 0.00 aMean and standard deviation. prediction model”. P<0.05 was considered sta- ization time, and medical expense. The preg- tistically significant. nancy duration, length of the umbilical cord, and body mass of the neonates in the case