2022 年第 2 期 第 17 卷

老年全身麻醉手术患者术后认知功能恢复情况的预测模型构建

Construction of a predictive model for postoperative cognitive function recovery in elderly patients undergoing general anesthesia surgery

作者：仲勇项金慧汪静李兴

英文作者：Zhong Yong Xiang Jinhui Wang Jing Li Xing

单位：上海中医药大学附属曙光医院麻醉科，上海201203

英文单位：Department of Anesthesiology Shuguang Hospital Shanghai University of Traditional Chinese Medicine Shanghai 201203 China

关键词：全身麻醉；认知功能；白细胞介素6；白细胞介素10；白细胞介素1β

英文关键词：Generalanesthesia;Cognitivefunction;Interleukin-6;Interleukin-10;Interleukin-1β

• 摘要：

• 目的  探讨老年全身麻醉手术患者术后认知功能恢复情况的预测模型构建。方法  选取2018年4月至2021年4月上海中医药大学附属曙光医院收治的361例老年全身麻醉手术患者，根据术后认知功能恢复情况分为延迟组和非延迟组，比较2组一般资料、术前情况、术中情况及术后情况，应用Logistic回归模型分析影响老年全身麻醉手术患者术后认知功能恢复的相关因素。随机选取80%（289例）的病例作为训练集，20%（72例）作为测试集，基于训练集建立列线图预测模型，并在测试集评估模型效能。结果  训练集289例患者中，术后认知功能恢复延迟60例（20.8%），非延迟229例（79.2%）。2组患者年龄、手术史、是否吸烟、受教育年限、术前蒙特利尔认知评估量表（MoCA）评分、术前白细胞介素6（IL-6）水平、麻醉深度、主要麻醉方式、麻醉时长、术中IL-6、IL-10、IL-1β水平比较差异均有统计学意义（均P＜0.05）。Logistic回归分析显示，术前MoCA评分（比值比=0.630，95%置信区间：0.487～0.815）、麻醉深度（比值比=0.906，95%置信区间:0.833～0.986）、麻醉方式（比值比=0.411，95%置信区间:0.191～0.883）、术中IL-10水平（比值比=0.843，95%置信区间:0.744～0.954）为全身麻醉老年患者术后认知功能延迟恢复的保护因素，术前IL-6水平（比值比=1.358，95%置信区间:1.168～1.580）、术中IL-6（比值比=1.027，95%置信区间:1.003～1.050）、术中IL-1β水平（比值比=1.803，95%置信区间:1.422～2.286）为全身麻醉老年患者术后认知功能延迟恢复的危险因素（均P＜0.05），上述因素预测术后认知功能延迟恢复的一致性指数为0.871（95%置信区间：0.818～0.923），校正曲线显示列线图模型预测与实际观察的相关性较好，平均绝对误差为0.018。外部验证受试者工作特征曲线显示敏感度为82.1%，特异度为87.5%，曲线下面积为0.890。结论  术前IL-6水平和术中IL-6、IL-10、IL-1β水平与老年全身麻醉手术患者术后认知功能恢复相关，可作为老年全身麻醉患者术后认知功能恢复的预测指标。

• Objective To explore the predictive model of postoperative cognitive function recovery in elderly patients undergoing general anesthesia surgery. Methods Totally 361 elderly patients undergoing general anesthesia surgery who were admitted to Shuguang Hospital, Shanghai University of Traditional Chinese Medicine from April 2018 to April 2021 were enrolled. According to the postoperative cognitive recovery, they were divided into delayed group and non-delayed group. The general information, and preoperative, intraoperative and postoperative conditions of the two groups were compared. Logistic regression model was used to analyze the related factors affecting the cognitive function recovery in elderly patients undergoing general anesthesia surgery. The training set was composed of 80% (289 cases) of the cases randomly, and another 20% (72 cases) was divided into the test set. A nomogram prediction model was constructed based on the training set, and the effectiveness of the model were evaluated on the test set. Results Among 289 cases of training test, there were 60 cases (20.8%) in the delayed group and 229 cases (79.2%) in the non-delayed group. There were significant differences in age, surgical history, smoking, years of education, preoperative Montreal Cognitive Assessment(MoCA) score, preoperative interleukin-6(IL-6) level, depth of anaesthesia, main anesthesia Methods , duration of anesthesia, and intraoperative IL-6, IL-10 and IL-1β levels between the two groups (all P＜0.05). Logistic analysis showed that preoperative MoCA score ［odds ratio (OR)=0.630, 95% confidence interval (CI):0.487-0.815］, depth of anesthesia (OR=0.906, 95%CI:0.833-0.986), anesthesia method (OR=0.411, 95%CI:0.191-0.883), intraoperative IL-10 (OR=0.843, 95%CI:0.744-0.954) were the protective factors for delayed recovery of cognitive function in elderly patients under general anesthesia surgery. The preoperative IL-6 level (OR=1.358, 95%CI:1.168-1.580), intraoperative IL-6 (OR=1.027, 95%CI:1.003-1.050), intraoperative IL-1β level (OR=1.803, 95% CI:1.422-2.286) were risk factors for delayed recovery of cognitive function in elderly patients undergoing general anesthesia (all P＜0.05). The consistency index of the above factors predicting delayed recovery of cognitive function after surgery was 0.871 (95%CI:0.818-0.923). The calibration curve showed that the nomogram model prediction had a good correlation with actual observations, and the average absolute the error was 0.018. The externally verified receiver operating characteristic curve showed that the the sensitivity was 82.1%, the specificity was 87.5% and the area under the curve was 0.890. Conclusions Preoperative IL-6 levels, intraoperative IL-6, IL-10 and IL-1β levels are correlated with postoperative cognitive function recovery in elderly patients undergoing general anesthesia surgery. Those can be used as predictors of postoperative cognitive function recovery in elderly patients undergoing general anesthesia.