{"title":"Analysis of Risk Factors and Establishment of a Risk Prediction Model for Severe Postpartum Haemorrhage.","authors":"Jing Wang, Pin Hu, Ying Yang, Yu Zhang, Yihong Lu, Xiaoqin Wang","doi":"10.12968/hmed.2024.0455","DOIUrl":null,"url":null,"abstract":"<p><p><b>Aims/Background</b> Severe postpartum haemorrhage (PPH) is a dangerous condition, characterized by rapid progression and poor prognosis. It remains the leading preventable cause of maternal death worldwide. This study aimed to investigate the risk factors for severe PPH and establish a prediction model to identify severe PPH early, allowing for early intervention reduce maternal death. <b>Methods</b> Clinical data were collected from 784 patients diagnosed with PPH and delivered at the Second Affiliated Hospital of Anhui Medical University between December 2018 and December 2023. These cases were categorized into the training cohort. Severe PPH was diagnosed in 234 patients based on the criterion of the volume of vaginal bleeding volume exceeding 1000 mL within 24 hours after delivery; these patients were assigned to the experimental group. The remaining 550 patients with nonsevere PPH were assigned to the control group. Data from an additional 338 postpartum women from the same period were selected and classified into the validation cohort. Univariate and multivariate logistic regression analyses were performed to pinpoint the determinants associated with severe PPH. Additionally, these analyses were instrumental for developing and assessing a prediction model to forecast the risk of such complications. <b>Results</b> Most of the PPH cases in this study stemmed from uterine atony, the leading aetiology. The second most common factor was soft birth canal lacerations and haematoma formation, accounting for 7.26% of the subjects in experimental group and 6.55% of those in the control group. Uterine rupture, uterine inversion, and amniotic fluid embolism were exclusively observed in the experimental group. In the univariate analysis, notable disparities were identified across various parameters, including maternal age, gravidity, parity, abortion frequency, gestational week at delivery, prothrombin time (PT), activated partial thromboplastin time (APTT), <i>in vitro</i> fertilisation, foetal position, caesarean delivery, pregnancy with anaemia, and hypertensive disorders of pregnancy (<i>p</i> < 0.05). A multivariate logistic regression model revealed that a parity of two or more, two or more abortions, <i>in vitro</i> fertilisation, gestational weeks at delivery, foetal position, caesarean delivery, pregnancy with anaemia, and hypertensive disorders of pregnancy were independent predictors of severe PPH (<i>p</i> < 0.05). The predictive model demonstrated excellent calibration for the training and validation datasets, with the areas under the curve (AUC) for receiver operating characteristic (ROC) curves measuring 0.799 and 0.759, respectively. Clinical decision curve analysis (DCA) confirmed a significant threshold exceeding 0.9, signifying a substantial net benefit and model precision. <b>Conclusion</b> Parity ≥2 times, abortion ≥2 times, <i>in vitro</i> fertilisation, gestational week at delivery, abnormal foetal position, caesarean delivery, pregnancy with anaemia, and hypertensive disorders of pregnancy are independent risk factors for severe PPH. The predictive model established in this study accurately predicts the occurrence of severe PPH and has significant value for clinical application.</p>","PeriodicalId":9256,"journal":{"name":"British journal of hospital medicine","volume":"85 11","pages":"1-16"},"PeriodicalIF":1.0000,"publicationDate":"2024-11-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"British journal of hospital medicine","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.12968/hmed.2024.0455","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/11/13 0:00:00","PubModel":"Epub","JCR":"Q3","JCRName":"MEDICINE, GENERAL & INTERNAL","Score":null,"Total":0}
引用次数: 0
Abstract
Aims/Background Severe postpartum haemorrhage (PPH) is a dangerous condition, characterized by rapid progression and poor prognosis. It remains the leading preventable cause of maternal death worldwide. This study aimed to investigate the risk factors for severe PPH and establish a prediction model to identify severe PPH early, allowing for early intervention reduce maternal death. Methods Clinical data were collected from 784 patients diagnosed with PPH and delivered at the Second Affiliated Hospital of Anhui Medical University between December 2018 and December 2023. These cases were categorized into the training cohort. Severe PPH was diagnosed in 234 patients based on the criterion of the volume of vaginal bleeding volume exceeding 1000 mL within 24 hours after delivery; these patients were assigned to the experimental group. The remaining 550 patients with nonsevere PPH were assigned to the control group. Data from an additional 338 postpartum women from the same period were selected and classified into the validation cohort. Univariate and multivariate logistic regression analyses were performed to pinpoint the determinants associated with severe PPH. Additionally, these analyses were instrumental for developing and assessing a prediction model to forecast the risk of such complications. Results Most of the PPH cases in this study stemmed from uterine atony, the leading aetiology. The second most common factor was soft birth canal lacerations and haematoma formation, accounting for 7.26% of the subjects in experimental group and 6.55% of those in the control group. Uterine rupture, uterine inversion, and amniotic fluid embolism were exclusively observed in the experimental group. In the univariate analysis, notable disparities were identified across various parameters, including maternal age, gravidity, parity, abortion frequency, gestational week at delivery, prothrombin time (PT), activated partial thromboplastin time (APTT), in vitro fertilisation, foetal position, caesarean delivery, pregnancy with anaemia, and hypertensive disorders of pregnancy (p < 0.05). A multivariate logistic regression model revealed that a parity of two or more, two or more abortions, in vitro fertilisation, gestational weeks at delivery, foetal position, caesarean delivery, pregnancy with anaemia, and hypertensive disorders of pregnancy were independent predictors of severe PPH (p < 0.05). The predictive model demonstrated excellent calibration for the training and validation datasets, with the areas under the curve (AUC) for receiver operating characteristic (ROC) curves measuring 0.799 and 0.759, respectively. Clinical decision curve analysis (DCA) confirmed a significant threshold exceeding 0.9, signifying a substantial net benefit and model precision. Conclusion Parity ≥2 times, abortion ≥2 times, in vitro fertilisation, gestational week at delivery, abnormal foetal position, caesarean delivery, pregnancy with anaemia, and hypertensive disorders of pregnancy are independent risk factors for severe PPH. The predictive model established in this study accurately predicts the occurrence of severe PPH and has significant value for clinical application.
期刊介绍:
British Journal of Hospital Medicine was established in 1966, and is still true to its origins: a monthly, peer-reviewed, multidisciplinary review journal for hospital doctors and doctors in training.
The journal publishes an authoritative mix of clinical reviews, education and training updates, quality improvement projects and case reports, and book reviews from recognized leaders in the profession. The Core Training for Doctors section provides clinical information in an easily accessible format for doctors in training.
British Journal of Hospital Medicine is an invaluable resource for hospital doctors at all stages of their career.
The journal is indexed on Medline, CINAHL, the Sociedad Iberoamericana de Información Científica and Scopus.