摘要
目的 分析超声引导下经皮肾穿刺活检(PRB)术后出血的影响因素,构建风险预测模型并探讨其临床价值.方法 选取于南京医科大学附属南京医院和滁州市来安家宁医院接受PRB术的72例患者作为建模组,并根据是否发生术后出血分为出血组(32例)和未出血组(40例),收集患者基线资料、实验室检查和超声检查资料,采用多因素Logistic回归分析筛选预测PRB术后出血的影响因素,构建风险预测模型.另选符合纳入标准的64例患者作为验证组,绘制受试者工作特征曲线分析该模型的预测效能;Hosmer-Lemeshow拟合优度检验评估该模型的拟合程度.结果 出血组与未出血组肾皮髓质分界、血肌酐、肾小球滤过率、血红蛋白(Hb)、血小板计数(PLT)、IgA肾病(IgA N)及高血压肾病(HTN)占比比较,差异均有统计学意义(均P<0.05).多因素Logistic回归分析显示,肾皮髓质分界、Hb、PLT、IgA N及HTN均为预测PRB术后出血的独立危险因素(均P<0.05).由此构建的风险预测模型在建模组的曲线下面积、灵敏度、特异度分别为 0.881、81.25%、92.50%,在验证组的曲线下面积、灵敏度、特异度分别为 0.837、71.87%、90.00%;Hosmer-Lemeshow拟合优度检验显示,该模型在建模组、验证组的预测概率与实际概率拟合均较好(χ2=4.3015、7.1549,P=0.6359、0.2094).结论 本研究构建的风险预测模型对PRB术后出血具有较好的预测效能,可为临床高风险人群的筛选和防治提供有价值的参考依据.
Abstract
Objective To analyze the influencing factors of postoperative bleeding under ultrasound-guided percutaneous renal biopsy(PRB),and to construct a risk prediction model and explore its clinical value.Methods Seventy-two patients who underwent PRB at Nanjing Hospital Affiliated to Nanjing Medical University and Lai'an Jia Ning Hospital of Chuzhou were selected as the modeling group.They were divided into a bleeding group(32 cases)and a non-bleeding group(40 cases)according to whether they had bleeding.Baseline data,laboratory tests and ultrasound examination data were collected from patients,and multiple Logistic regression analysis was used to screen the influencing factors of postoperative bleeding after PRB,and a risk prediction model was constructed.Another 64 patients who met the inclusion criteria were selected as the validation group,and the receiver operating characteristic curve was drawn to analyze the predictive efficacy of the model.Hosmer-Lemeshow goodness of fit test was used to evaluate the degree of fit of the model.Results There were statistically significant differences in the demarcation of renal cortex and medulla,serum creatinine,glomerular filtration rate,hemoglobin(Hb),platelet count(PLT),IgA nephropathy(IgA N),and the proportion of hypertensive nephropathy(HTN)between the bleeding group and the non-bleeding group(all P<0.05).Multivariate Logistic regression analysis indicated that the demarcation of renal cortex and medulla,Hb,PLT,IgA N,and HTN were independent risk factors for predicting postoperative bleeding after PRB(all P<0.05).The area under the curve,sensitivity,and specificity of the established risk prediction model in the modeling group were 0.881,81.25%,and 92.50%,respectively,and those in the validation group were 0.837,71.87%,and 90.00%,respectively.The Hosmer-Lemeshow goodnessoffit test showed that the predicted probabilities and actual probabilities of the model fitted well in both the modeling group and the validation group(χ2=4.3015,7.1549,P=0.6359,0.2094).Conclusion The risk prediction model constructed in this study has good predictive performance for postoperative bleeding after PRB,and can provide valuable reference for screening and prevention of high-risk populations in clinical practice.
维普
万方数据