混合痔行吻合器痔切除术术后出血预测模型的建立

Establishment of a bleeding prediction model for mixed hemorrhoids after PPH

目的联合多项检查指标,发现吻合器痔切除术(PPH)术后出血的危险因素,制定其出血预测模型,以期为临床减少PPH术后出血,改善患者预后,减轻患者痛苦。方法本研究为回顾性研究。分析2016年12月至2021年5月于上海交通大学医学院附属松江医院(筹)就诊的258例行PPH患者的临床资料,其中,男115例,女143例,平均年龄45.64岁。观察指标为有无高血压、糖尿病并发症,肛管放置时间、血红蛋白、白细胞、白蛋白、丙氨酸转移酶(ALT)、天冬氨酸转移酶(AST)、肌酐、尿酸、血小板、活化部分凝血活酶时间(APTT)、国际标准化比值(INR)、凝血酶原时间(PT)、凝血酶时间(TT)、肠道准备是否充分、大便是否干结、是否应用抗生素,运用t检验及χ^(2)检验进行单因素分析,运用logistic回归分析进行多因素分析并建立术后出血的预测模型,同时运用受试者工作特征曲线(ROC)对该模型进行预测效力验证。结果在多因素分析中,大便干结(P=0.01)、肛管放置时间(P=0.02)、抗生素应用情况(P=0.02)、肠道准备(P<0.01)、TT(P=0.01)在预测PPH术后出血方面差异均有统计学意义。PPH术后出血预测模型为Y=1[/1+EXP(2.16×肛管放置时间-2.24×大便干结+1.97×不使用抗生素+2.92×肠道准备不充分+1.64×TT-35.29)],利用ROC对该模型预测效力进行检验,曲线下面积(AUC)为0.987。结论大便干结、肛管放置时间、抗生素应用情况、肠道准备、TT均为PPH术后出血的独立危险因素,与出血密切相关,该预测模型的建立对帮助临床减少该并发症的发生有一定的帮助。

Objective To find out the risk factors of postoperative bleeding after procedure for prolapse and hemorrhoids(PPH)combined with several examination indicators,and to formulate its bleeding prediction model in order to reduce the postoperative bleeding after PPH,improve the patients'prognosis,and alleviate the patients'pain.Methods This study was a retrospective observational study.The clinical data of 258 patients who underwent PPH in Songjiang Hospital Affiliated to Shanghai Jiaotong University School of Medicine(Preparatory Stage)from December 2016 to May 2021 were retrospectively analyzed.Among them,there were 115 males and 143 females,with an average age of 45.64 years old.The main observation indicators were whether there were complications of hypertension and diabetes,anal tube placement time,hemoglobin,leukocytes,albumin,alanine aminotransferase(ALT),aspartate aminotransferase(AST),creatinine,uric acid,platelets,activated partial thrombin time(APTT),international normalized ratio(INR),prothrombin time(PT),thrombin time(TT),whether the intestinal preparation was adequate,whether the stool was dry,and whether the antibiotics are used.Univariate analysis was performed using t test andχ^(2) test,and multivariate analysis was performed using logistic regression analysis.A predictive model for postoperative bleeding was established,and the receiver operating characteristic curve(ROC)was used to verify the predictive efficacy of the model.Results In the multivariate analysis,dry stool(P=0.01),anal tube placement time(P=0.02),antibiotic use(P=0.02),intestinal preparation(P<0.01),and TT(P=0.01)all had statistically significant differences in predicting postoperative bleeding after PPH.The predictive model of postoperative bleeding after PPH was Y=1/[1+EXP(2.16×anal tube placement time-2.24×dry stool+1.97×no antibiotic use+2.92×inadequate intestinal preparation+1.64×TT-35.29)],the predictive power of the model was tested by ROC,and the area under the curve(AUC)was 0.987.Conclusions Dry stool,anal tube placement time,antibiotic use,intestinal preparation,and TT are independent risk factors for postoperative bleeding after PPH,which are closely related to bleeding.The establishment of this predictive model is helpful to reduce the occurrence of this complication in clinical practice.