产后出血的影响因素分析及预测  

The analysis and predict of postpartum hemorrhage

作　　者：赵相娟[1] 张梅娜[1] 张涛[2] 贲崴[1] 郑建华[1] 

机构地区：[1]哈尔滨医科大学附属第一医院妇产科,150001 [2]哈尔滨医科大学卫生统计学教研室

出　　处：《中国妇产科临床杂志》2012年第2期108-111,共4页Chinese Journal of Clinical Obstetrics and Gynecology

摘　　要：目的分析产后出血的相关影响因素,寻找预测产后出血的方法,为预防产后出血提供依据。方法选择2008年6月至2010年12月在哈尔滨医科大学附属第一医院住院分娩发生产后出血的67例患者为研究组,按照1:2的比例以随机数的方式随机抽取院内同期分娩非产后出血患者139例为对照组。采取病例对照研究方法分析产后出血的影响因素,建立多因素logistic模型,并使用ROC曲线评估logistic模型的预测效果。结果产后出血发生率约为1.39%(67/4808)。研究组和对照组产程异常、双胎、宫缩乏力、产道裂伤、胎盘因素(胎盘粘连、胎盘植入、前置胎盘)、血小板减少、职业和会阴切开比较,差异具有统计学意义(P<0.01)。对所有因素进行多因素logistic回归分析,最终进入最优模型的因素有6个:宫缩乏力、血小板减少、会阴切开、分娩方式、前置胎盘和职业;并且宫缩乏力、血小板减少和前置胎盘是产后出血的危险因素,OR值分别为60.415(95%CI:15.86～230.17)、36.216(95%CI:4.68～280.35)和19.175(95%CI:2.78～132.33);而分娩方式、职业和会阴切开是保护因素,OR值分别为0.219(95%CI:0.082～0.586)、0.292(95%CI:0.100～0.851)和0.092(95%CI:0.014～0.59)。ROC曲线下面积AUC=0.9049,灵敏度为75.9%,特异度为89.9%。结论产后出血为多因素影响的结果,本研究建立的logistic回归模型可有效地预测产后出血的发生。Objective To detect methods of indicating postpartum hemorrhage and to analyze the related risk factors.Methods We performed a retrospective analysis including 67 postpartum hemorrhage cases from the department of Obstetric and Gynecology in the First Affiliated Hospital of Harbin Medical University between June 2008 and December 2010,and 139 puerperas not suffering from postpartum hemorrhage were randomly selected as controls in the same hospital.The univariate analysis and multiple logistic regression model were established,and the ROC analysis was employed to evaluate the predictive ability of multiple models.Results The incidence of postpartum hemorrhage was about 1.39%.The Chi-square test illustrated that abnormal labor,twin pregnancy,uterine atony,laceration of the lower genital tract,placenta factors（placenta accrete,placenta increta,placenta previa）,thrombocytopenia,occupation and episiotomy had statistically significances between the two groups（P0.01）.Using multivariate logistic regression with stepwise variable selection strategy,6 factors were selected into the optimal model ultimately：uterine atony,thrombocytopenia,episiotomy,mode of delivery,placenta previa,occupation.Uterine atony,thrombocytopenia and placenta previa were risk factors of postpartum hemorrhage,the OR（95%CI） values were 60.415（15.86~230.17）,36.216（4.68~280.35） and 19.175（2.78~132.33）;while delivery method,occupation and episiotomy were protective factors,OR values were 0.219（0.082~0.586）,0.292（0.100~0.851） and 0.092（0.014~0.59）.The area under ROC curve was 0.904 9,the sensitivity was 75.9% and the specificity was 89.9%.Conclusions Postpartum hemorrhage is influenced by many factors.The multiple logistic regression model constructed in our study could be effectively applied to predict the occurrence of postpartum hemorrhage.

关 键 词：产后出血 危险因素 LOGISTIC回归 预测 

分 类 号：R714.461[医药卫生—妇产科学]