老年早期胃癌患者内镜治疗术后伴发营养不良的风险预测模型构建

Construction of a risk prediction model for postoperative malnutrition in elderly patients with early gastric cancer after endo-scopic therapy

中文关键词: 老年  胃癌  营养不良  影响因素  人格特质  低蛋白血症

英文关键词: elderly  gastric cancer  malnutrition  influencing factor  personality trait  hypoalbuminemia

中文摘要:

      目的 分析老年早期胃癌患者行内镜黏膜下剥离术(ESD)、内镜黏膜切除术(EMR)居家伴发营养不良的影响因素.方法 选取2022年1月-2023年12月在江苏省中医院接受早期胃癌行ESD或EMR治疗的80例老年患者作为研究对象,根据术后在居家环境中是否发生营养不良,将患者分为发生营养不良组(n=45)和未发生营养不良组(n=35).采用单因素和多因素分析方法,探讨影响营养不良发生的高危因素.结果 单因素分析显示,2组高血压、血钾、血糖、血红蛋白检测值差异无统计学意义(P＞0.05).2组年龄、文化程度、焦虑得分、抑郁得分、早饱感、人格特质、术前低蛋白血症、术前血钙等均与胃癌患者术后发生营养不良有关(均P＜0.05).Logistic回归分析显示,年龄＞75岁、抑郁得分＞35分、有早饱感、焦虑得分≥29分、术前低蛋白血症等均为导致胃癌患者术后发生营养不良的独立危险因素(P＜0.05).以模型新生成概率值为依据,对该模型进行数据验证,绘制预测模型ROC曲线,可见AUC=0.862(95％CI:0.795～0.859)＞0.75,最佳临界值为0.554,提示模型具有一定的预测能力.Hosmer-Lemeshow检验评价模型校准能力(x2=0.879,P=0.761＞0.05),提示模型有良好的拟合优度.DCA结果显示,当营养不良临界概率值在9％～95％时,使用该模型能够获得较高的净收益.结论 本研究构建的老年早期胃癌患者术后营养不良风险预测模型,通过多因素Logistic回归分析发现,年龄＞75岁、抑郁得分＞35分、有早饱感、焦虑得分≥29分、术前低蛋白血症等可能为其独立危险因素.该模型经数据验证,具有较好的预测能力和校准度,且在营养不良临界概率值范围内显示出较高的净收益,可为临床实践中老年早期胃癌患者的术后营养管理提供参考.

英文摘要:

      Objective To analyze the influencing factors of malnutrition in elderly patients with early gastric cancer un-dergoing endoscopic submucosal dissection(ESD)and endoscopic mucosal resection(EMR)at home.Methods 80 elderly patients who received ESD or EMR treatment for early gastric cancer in Jiangsu Provincial Hospital of Traditional Chinese Medi-cine from January 2022 to December 2023 were selected as the observation subjects.According to whether malnutrition occurred in the home environment after surgery,they were divided into malnutrition group(n=45)and non-malnutrition group(n=35).The high risk factors for malnutrition were investigated by univariate and multiple factor analysis.Results Univariate analysis showed that there was no statistically significant difference in the detection values of hypertension,blood potassium,blood glucose,and hemoglobin between the two groups(P＞0.05).The age,educational level,anxiety score,depression score,early satiety,personality traits,preoperative hypoalbuminemia,preoperative blood calcium,etc.of both groups were all associated with postoperative malnutrition(all P＜0.05).Logistic regression analysis showed that age＞75 years old,depres-sion score＞35 points,early satiety,anxiety score ≥29 points and preoperative hypoproteinemia were independent risk factors for postoperative malnutrition in gastric cancer patients(P＜0.05).Based on the newly generated probability values of the model,data validation was conducted on the model,and the ROC curve of the prediction model was plotted.It was found that AUC=0.862(95％CI=0.795-0.859)＞0.75,and the optimal critical value was 0.554.It indicated that the model had cer-tain predictive ability.Hosmer-Lemeshow test evaluated the calibration ability of the model(x2=0.879,P=0.761＞0.05),and the evaluated result suggested that the model had a good goodness of fit.The DCA results showed that when the critical probability of malnutrition was between 9％and 95％,the application of this model could achieve higher net benefits.Conclusion This study established a risk prediction model for postoperative malnutrition in elderly patients with early gastric cancer.Multivariate Logistic regression analysis shows that age＞75 years old,depression score＞35 points,early satiety,anx-iety score ≥29 points,preoperative hypoproteinemia may be independent risk factors.This model has good predictive ability and calibration,and shows a high net benefit within the critical probability range of malnutrition.It can provide reference for postoperative nutrition management of early gastric cancer patients in clinical practice.