摘要
目的:探讨颌骨囊肿患者术后发生感染的危险因素,并构建和验证列线图(nomogram)预测模型。方法:回顾2020年5月至2024年1月期间于本院就诊的257例颌骨囊肿患者临床资料,根据感染情况将患者分为感染组54例和未感染组203例。采用多因素Logistic回归分析筛选独立危险因素,采用R软件构建颌骨囊肿患者术后感染的nomogram预测模型,采用霍斯默-莱梅肖检验模型拟合度,校正曲线检验模型校准度,受试者工作特征(ROC)曲线检验模型区分度。结果:54例并发术后感染患者中分离出病原菌92株,其中革兰阴性菌41株(44.57%)、革兰阳性菌48株(52.17%)、真菌3株(3.26%)。未感染组和感染组颌骨囊肿患者在糖尿病、手术时间、术中出血量、未无菌操作、术后白细胞计数(WBC)、术后降钙素原(PCT)方面比较差异有统计学意义(P<0.05)。有糖尿病、手术时间较长、术中出血量较大、未无菌操作、术后WBC较高是颌骨囊肿患者术后发生感染的独立危险因素(P<0.05)。nomogram预测模型霍斯默-莱梅肖检验结果χ^(2)=5.745,P=0.146,模型拟合度高。校正曲线结果显示模型预测感染概率和实际感染概率基本一致,模型校准度高。ROC曲线结果显示模型预测AUC为0.915(95%CI:0.868~0.962),模型区分度高。当高风险阈值概率范围在0.04~0.98时,该nomogram预测模型有较高的临床净获益。结论:以危险因素糖尿病、手术时间、术中出血量、未无菌操作、术后WBC构建的nomogram预测模型有较高的预测效能,区分度和校准度均较高。
Objective:To explore the risk factors for postoperative infection in patients with jaw cysts,and to construct and validate a nomogram prediction model.Methods:The clinical data of 257 patients with jaw cysts who visited our hospital from May 2020 to January 2024 were reviewed.According to the infection status,the patients were grouped into an infected group of 54 cases and an uninfected group of 203 cases.Multivariate Logistic regression analysis was applied to screen for independent risk factors.R software was applied to construct a nomogram prediction model for postoperative infection in patients with jaw cysts.Hosmer-Lemeshow test was applied to test the model fit,the calibration curve was used to test the model calibration,and the receiver operating characteristic(ROC)curve was used to test the model discrimination.Results:92 strains of pathogenic bacteria were isolated from 54 patients with postoperative infections,including 41 strains of gram-negative bacteria(44.57%),48 strains of gram-positive bacteria(52.17%),and 3 strains of fungi(3.26%).There were significant differences between the uninfected group and the infected group in diabetes,operation time,intraoperative bleeding,non sterile operation,postoperative white blood cell count(WBC),and postoperative procallcitonin(PCT)(P<0.05).Diabetes,long operation time,large amount of intraoperative blood loss,non sterile operation,and high postoperative WBC were independent risk factors for postoperative infection in patients with jaw cysts(P<0.05).The Hosmer-Lemeshau test results of the nomogram prediction model showedχ^(2)=5.745 and P=0.146,the model had a high degree of fit.The calibration curve results showed that the predicted infection probability of the model was basically consistent with the actual infection probability,indicating a high degree of model calibration.ROC curve results showed that the model had an AUC of 0.915(95%CI:0.868-0.962),indicating high model discrimination.When the probability range of the high-risk threshold was 0.04-0.98,the nomogram prediction model had a high clinical net benefit.Conclusion:The nomogram prediction model based on the risk factors,including diabetes,operation time,intraoperative blood loss,non sterile operation,and postoperative WBC,has a high prediction efficiency,and a high degree of differentiation and calibration.
作者
邢晓伟
刘彬
饶圆
程增遂
张江云
XING Xiaowei;LIU Bin;RAO Yuan;CHENG Zengsui;ZHANG Jiangyun(Department of Oral and Maxillofacial Surgery,Handan Stomatological Hospital,Handan 056000,China;Department of Implantology,Handan Stomatological Hospital,Handan 056000,China;Department of Restoration,Handan Stomatological Hospital,Handan 056000,China;Department of Dentistry and Endodontics,Handan Stomatological Hospital,Handan 056000,China)
出处
《东南大学学报（医学版）》
CAS
2024年第6期868-874,共7页
Journal of Southeast University(Medical Science Edition)
基金
邯郸市科学技术研究与发展计划(23422083072ZC)。
关键词
颌骨囊肿
术后感染
危险因素
列线图预测模型
jaw cyst
postoperative infection
risk factors
nomogram prediction model
