| 摘要: |
| ［摘要］ 目的 基于血浆总胆固醇（TC）水平构建非高脂血症食管癌切除患者发生术后谵妄（POD）的列线图预测模型。方法 回顾性分析2019年1月至2023年12月在中国人民解放军联勤保障部队第九六〇医院行食管癌切除术的187例非高脂血症患者的临床资料。根据术后1周内谵妄发生情况将患者分为谵妄组（24例）和非谵妄组（163例）。采用多因素logistic回归分析食管癌切除患者发生POD的影响因素，构建列线图预测模型。通过Hosmer-Lemeshow拟合优度检验、Bootstrap重抽样法、受试者工作特征（ROC）曲线及决策曲线分析（DCA）验证模型的效能。结果 与非谵妄组比较，谵妄组年龄较大，使用右美托咪定的患者占比较低，术中最低平均动脉压（MAP）以及TC、低密度脂蛋白胆固醇（LDL-C）水平较低，麻醉时间、术后住院时间较长，冠心病患者占比、术后转入重症监护病房（ICU）患者占比、术后30 d内死亡患者占比较高，差异有统计学意义（P<0.05）。多因素logistic回归分析结果显示，较高的TC水平、较高的术中最低MAP以及使用右美托咪定是抑制食管癌切除患者发生POD的独立保护因素（P<0.05）。基于3个影响因素构建列线图预测模型。Hosmer-Lemeshow拟合优度检验结果显示该模型拟合良好（χ2=3.763，P=0.878）。Bootstrap重抽样法得出C-index值为0.897，模型具有较高的预测一致性。ROC曲线显示模型预测食管癌切除患者发生POD的AUC（95%CI）为0.919(0.870～0.969），特异度和灵敏度分别为79.10%和91.70%。DCA曲线显示模型在风险为0.05～0.85的阈值范围内净获益较高且稳定，具有较好的临床价值。结论 较高的血浆TC水平、较高的术中最低MAP以及使用右美托咪定是抑制非高脂血症食管癌切除患者发生POD的独立保护因素，基于以上3个因素构建的列线图预测模型对POD具有较好的预测价值。 |
| 关键词: 总胆固醇 预测模型 食管癌 术后谵妄 老年人 血脂 |
| DOI：10.3969/j.issn.1674-3806.2025.10.07 |
| 分类号: |
| 基金项目:山东省自然科学基金青年项目（编号：ZR2021QH031） |
|
| Construction of a nomogram model for predicting postoperative delirium in non-hyperlipidemia patients undergoing esophageal cancer resection based on plasma total cholesterol levels |
|
GAO Shengrun1, LI Yun2, LI Yan1, MA Guodong1, GAO Chengjie1
|
|
1.Department of Anesthesiology, the 960th Hospital of the PLA Joint Logistics Support Force, Jinan 250031, China; 2.Department of Pain Management, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan 250022, China
|
| Abstract: |
| ［Abstract］ Objective To construct a nomogram model for predicting postoperative delirium(POD) in non-hyperlipidemia patients undergoing esophageal cancer resection based on plasma total cholesterol(TC) levels. Methods A retrospective analysis was conducted on the clinical data of 187 non-hyperlipidemia patients who underwent esophageal cancer resection in the 960th Hospital of the PLA Joint Logistics Support Force from January 2019 to December 2023. According to the occurrence of delirium within 1 week after the operation, the patients were divided into delirium group(24 patients) and non-delirium group(163 patients). Multivariate logistic regression was used to analyze factors influencing the incidence of POD in the patients undergoing esophageal cancer resection, and a nomogram prediction model was constructed. The predictive efficacy of the model was verified by using Hosmer-Lemeshow goodness-of-fit test, Bootstrap resampling method, receiver operating characteristic(ROC) curve and decision curve analysis(DCA). Results Compared with the non-delirium group, the delirium group was older, and had a lower proportion of patients receiving dexmedetomidine treatment, lower levels of intraoperative minimum mean arterial pressure(MAP) and total cholesterol(TC), and low-density lipoprotein cholesterol(LDL-C), longer duration of anesthesia and postoperative hospital stay, and a higher proportion of patients with coronary heart disease, a higher proportion of patients transferred to intensive care unit(ICU) after surgery and a higher proportion of patients who died within 30 days after surgery, with statistically significant differences(P<0.05). The results of multivariate logistic regression analysis showed that a higher TC level, a higher intraoperative minimum MAP, and use of dexmedetomidine were independent protective factors for inhibiting the incidence of POD in the patients undergoing esophageal cancer resection(P<0.05). A nomogram prediction model was constructed based on the above three influencing factors. The results of Hosmer-Lemeshow goodness-of-fit test showed that the model fitted well(χ2=3.763, P=0.878). The C-index value obtained by the Bootstrap resampling method was 0.897, indicating that the model had a relatively high prediction consistency. The ROC curve showed that area under the curve(AUC)(95%CI) of the model for predicting POD in patients undergoing esophageal cancer resection was 0.919(0.870-0.969), with specificity of 79.10% and sensitivity of 91.70%. The DCA curve showed that the net benefit of the model was relatively high and stable within the risk thresholds ranging from 0.05 to 0.85, indicating that the model had good clinical value. Conclusion Higher plasma TC level, higher intraoperative minimum MAP and use of dexmedetomidine are independent protective factors for inhibiting the incidence of POD in non-hyperlipidemic patients undergoing esophageal cancer resection. The nomogram prediction model constructed based on the above three factors has good predictive value for POD. |
| Key words: Total cholesterol Prediction model Esophageal cancer Postoperative delirium(POD) The elderly Blood lipids |
