麻醉后监测治疗室患者术后口渴风险预测模型的构建及验证  

Construction and validation of a nomogram model to predict postoperative thirst risk in patients in the post-anesthesia care unit

作　　者：郑敏 陆小英 张丽君 薄禄龙 彭琳 Zheng Min;Lu Xiaoying;Zhang Lijun;Bo Lulong;Peng Lin(Faculty of Anesthesiology,the First Affiliated Hospital of Naval Military Medical University,Shanghai 200433,China;Nursing Department,the First Affiliated Hospital of Naval Military Medical University,Shanghai 200433,China)

机构地区：[1]海军军医大学第一附属医院麻醉学部,上海200433 [2]海军军医大学第一附属医院护理部,上海200433

出　　处：《国际麻醉学与复苏杂志》2025年第3期266-272,共7页International Journal of Anesthesiology and Resuscitation

基　　金：2022年海军军医大学深蓝护理科研项目(2022KYG15);上海市护理学会优秀青年人才育苗计划([2023]35);海军军医大学第一附属医院长风人才工程([2024]1)。

摘　　要：目的探讨麻醉后监测治疗室(PACU)患者术后口渴的影响因素,构建其风险预测模型并检验预测效果。方法选取2023年12月至2024年3月入PACU苏醒的366例患者,按照2∶1比例采用随机数字表法分为建模组(244例)和验证组(122例)。记录患者一般资料(性别、年龄、体重指数、高血压病史、糖尿病史、冠心病史、吸烟史、饮酒史、心功能分级)、麻醉手术情况[禁食禁饮时长(NPO时长)、留置气管导管时长(置管时长)、美国麻醉医师协会(ASA)分级、手术分级、术中补液量、输血量、出血量、尿量、体液平衡、舒芬太尼用量]、使用麻醉药(抗胆碱能药、麻黄碱、肌松拮抗药)情况。采用数字分级评分法(NRS)进行口渴评分,建模组根据评分结果分为口渴组(183例)和非口渴组(61例)。应用logistic回归分析危险因素,利用R软件绘制列线图预测风险模型,Hosmer-Lemeshow(H-L)检测判断模型的拟合优度,采用受试者操作特征(ROC)曲线和决策曲线分析(DCA)评价模型预测发生口渴的价值,将验证组临床数据代入模型中,判断本预测模型的预测价值。结果建模组患者术后口渴的发生率为75.0%,主动汇报口渴患者口渴评分较高(P<0.05),主动汇报口渴患者年龄较小,但年龄差异无统计学意义(P>0.05)。口渴预测模型纳入风险因素分别为年龄、高血压病史、NPO时长、置管时长、使用抗胆碱能药(均P<0.05)。预测模型的H-L检测P=0.592;ROC曲线下面积(AUC)为0.820(95%置信区间0.770~0.880),最佳临界值为0.727,敏感度为0.851,特异度为0.810;模型外部验证显示正确率为82.79%;提示模型具有较好的拟合效果和较高的预测价值。结论PACU患者术后口渴发生率高。构建的模型具有较好的预测效果,能早期、高效预测PACU患者术后口渴发生的风险,为术后口渴的早期评估和预防提供科学依据。Objective To explore the influencing factors of postoperative thirst in patients in the post-anesthesia care unit(PACU),construct a prediction model and evaluate its predictive performance.Methods A total of 366 patients who emerged from anesthesia in the PACU from December 2023 to March 2024 were enrolled.According to the random number table method,the patients were divided into a modeling group(n=244)and a validation group(n=122)in a ratio of 2∶1.Data collected included general information(e.g.sex,age,body mass index,history of hypertension,history of diabetes,history of coronary heart disease,smoking history,alcohol consumption history,and cardiac function classification),anesthesia-related information[e.g.duration of fasting(NPO duration),duration of endotracheal intubation(intubation duration),American Society of Anesthesiologists(ASA)classification,surgical grade,intraoperative fluid volume,blood transfusion volume,blood loss,urine volume,fluid balance,and sufentanil dosage],and the use of anesthetic agents(e.g.anticholinergic drugs,ephedrine,and muscle relaxant antagonists).Thirst was scored using the Numeric Rating Scale(NRS).Based on the NRS scores,the modeling group was divided into a thirst group(n=183)and a non-thirst group(n=61).Logistic regression analysis was used to identify risk factors,and a nomogram was constructed by R software to predict the risk model.The model's goodness of fit was assessed by the Hosmer-Lemeshow(H-L)test,and its predictive value for thirst was evaluated using the receiver operating characteristic(ROC)curve and decision curve analysis(DCA).The clinical data of the validation group were applied to the model to determine its predictive value.Results The incidence of postoperative thirst in the modeling group was 75.0%.Patients who actively reported their thirst had significantly increased thirst scores(P<0.05).Younger patients were more likely to report thirst,but the difference in age was not statistically significant(P>0.05).Risk factors included in the thirst prediction

关 键 词：麻醉后监测治疗室 口渴 风险因素 预测模型 列线图 

分 类 号：R614[医药卫生—麻醉学]