摘要
目的基于围术期资料和麻醉因素建立骨科手术神经阻滞后爆发痛风险的预测模型及检验。方法选取2022年10月至2024年7月周口市中心医院收治的298例骨科手术神经阻滞患者纳入研究,按照7:3比例分为建模集208例和验证集90例,对爆发痛进行单因素分析、随机森林分析和最小绝对收缩与选择算子(LASSO)回归分析,构建风险预测模型,并加以验证。结果单因素分析结果显示,年龄≥60岁(OR=0.504)、神经阻滞药物为布比卡因脂质体(OR=0.502)、超前镇痛(OR=0.492)、局麻药复合使用地塞米松(OR=0.485)、女性(OR=1.354)、术前VAS评分≥4分(OR=1.601)、上肢手术患者(OR=1.861)、手术时间≥2 h(OR=2.428)、合并骨质疏松(OR=2.192)为骨科手术神经阻滞后发生爆发痛的影响因素;随机森林-LASSO回归结果显示,年龄、术前VAS评分、手术部位、神经阻滞药物、超前镇痛、局麻药复合使用地塞米松、合并骨质疏松是骨科手术神经阻滞后爆发痛的独立预测因子(P<0.05);基于随机森林-LASSO回归构建的风险预测模型一致性指数(C-index)为0.806;受试者工作特征(ROC)曲线、校准曲线结果显示,该模型在建模集和验证集中的ROC曲线下面积(AUC)值分别为0.806(95%CI:0.747~0.864)和0.836(95%CI:0.748~0.924);Hosmer-Lemeshow拟合优度检验显示,该模型拟合良好。结论骨科手术神经阻滞后爆发痛的独立预测因子为年龄、术前VAS评分、手术部位、神经阻滞药物、超前镇痛、局麻药复合使用地塞米松、合并骨质疏松,基于此构建的风险预测模型的区分度、准确性较高,预测能力较强,有助于辅助早期捕获、识别骨科手术神经阻滞后爆发痛患者,为临床治疗提供参考依据。
Objective To establish and validate a risk predictive model for breakthrough pain after nerve block in orthopedic surgery based on perioperative data and anesthetic factors.Methods A total of 298 patients undergoing nerve block for orthopedic surgery admitted to Zhoukou Central Hospital from October 2022 to July 2024 were included and divided into a modeling cohort(n=208)and a validation cohort(n=90)in a ratio of 7:3.Univariate analysis,random forest analysis,and least absolute shrinkage and selection operator(LASSO)regression analysis were conducted to identify predictors of breakthrough pain,on which a risk-prediction model was built and subsequently validated.Results Univariate analysis indicated that age≥60 years(OR=0.504),use of bupivacaine liposomes as the nerve block drug(OR=0.502),preemptive analgesia(OR=0.492),combined use of dexamethasone with local anesthetics(OR=0.485),being female(OR=1.354),preoperative VAS score≥4 scores(OR=1.601),patients undergoing upper limb surgery(OR=1.861),surgery duration≥2 hours(OR=2.428),and comorbid osteoporosis(OR=2.192)were influencing factors for the occurrence of breakthrough pain after nerve block in orthopedic surgery.Random forest-LASSO regression analysis revealed that age,preoperative VAS score,surgical site,nerve block drug,preemptive analgesia,combined use of dexamethasone with local anesthetics,and comorbid osteoporosis were independent predictors of breakthrough pain after nerve block in orthopedic surgery(P<0.05).The concordance index(C-index)of the risk prediction model constructed based on random forest-LASSO regression was 0.806.The results of the receiver operating characteristic(ROC)curve and calibration curve indicated that the area under the ROC curve(AUC)values for this model in the modeling and validation cohorts were 0.806(95%CI:0.747-0.864)and 0.836(95%CI:0.748-0.924),respectively.The Hosmer-Lemeshow goodness-of-fit test demonstrated that the model fit well.Conclusion The independent predictors for breakthrough pain after nerve block in orthopedic surgery include age,preoperative VAS score,surgical site,nerve block drug,preemptive analgesia,local anesthetic compounded with dexamethasone,and combined osteoporosis.The risk prediction model based on these factors demonstrates high discrimination,accuracy,and strong predictive ability,offering valuable support for the early capture and identification of patients at risk of breakthrough pain after nerve block in orthopedic surgery,thereby providing a reference basis for clinical treatment.
作者
刘曙光
张春燕
刘晶晶
韩劲松
LIU Shu-guang;ZHANG Chun-yan;LIU Jing-jing;HAN Jing-song(Department of Anesthesiology,Zhoukou Central Hospital,Zhoukou 466000,Henan,CHINA)
出处
《海南医学》
2025年第11期1569-1575,共7页
Hainan Medical Journal
基金
河南省周口市二〇二三年第二批科技发展计划(编号:2023ptgg118)。
关键词
骨科手术
神经阻滞
爆发痛
风险预测模型
影响因素
模型验证
Orthopedic surgery
Nerve block
Breakthrough pain
Risk prediction model
Influencing factors
Model validation
