食管癌术后患者肠内营养耐受不良的影响因素及其列线图风险模型建立

[中文引用][英文引用]

 【摘要】目的： 探究食管癌术后患者肠内营养（EN）耐受不良的影响因素并构建列线图风险模型。方法： 选取2020年9月至2022年10月淮安市第二人民医院收治的158例食管癌患者的临床资料，依据术后EN耐受情况分为耐受组（116例）和耐受不良组（42例），收集患者年龄、性别、高血压、高血糖、术后药物使用情况、血清白蛋白是否正常、EN途径、是否肠道损伤等资料；Logistic回归分析影响食管癌术后EN耐受不良的因素。R3.6.3软件及rms程序包构建预测食管癌术后EN耐受不良的列线图模型。预测食管癌术后EN耐受不良列线图模型的区分度及一致性采用ROC及校准曲线评估。结果： 耐受不良组高血糖、术后药物使用、血清白蛋白不正常及肠道损伤的比例均显著高于耐受组（P<0.05）；多因素Logistic回归分析结果显示，高血糖、术后药物使用、血清白蛋白不正常及肠道损伤是EN耐受不良的独立危险因素（P＜0.05）；列线图模型显示，高血糖增加87.5分的权重，术后药物使用增加100分的权重，血清白蛋白不正常增加69.8分的权重，肠道损伤增加80.00分的权重；H-L拟合度检验显示，χ2=8.318、P=0.304，校准曲线斜率趋近1，ROC曲线进行模型区分度评估，曲线下面积为0.840（95%CI 0.773~0.893），敏感度、特异性分别为80.95%、76.72%。结论：高血糖、术后药物使用、血清白蛋白不正常及肠道损伤均是食管癌患者术后EN耐受不良的影响因素，以此构建的列线图模型在用于个体化预测食管癌患者术后EN耐受不良风险具有较高的临床意义。

 [Abstract]  Objective  To explore the influencing factors of enteral nutrition (EN) intolerance in patients with esophageal cancer after operation and construct a nomogram risk model. Methods  The clinical data of 158 patients with esophageal cancer admitted to Huai'an Second People's Hospital from September 2020 to October 2022 were selected and divided into tolerance group (116 cases) and intolerance group (42 cases) according to EN tolerance after operation. The patient's age, sex, hypertension, hyperglycemia, postoperative drug use, whether serum albumin is normal, EN pathway, whether intestinal injury and other data were collected. Logistic regression analysis was used to analyze the factors influencing EN intolerance after esophageal cancer surgery. R3.6.3 software and rms package were used to construct a nomogram model to predict EN intolerance after esophageal cancer surgery. The differentiation and consistency of EN intolerance nomogram model were evaluated by ROC and calibration curve. Results  The proportions of hyperglycemia, postoperative drug use, abnormal serum albumin and intestinal injury in the intolerance group were significantly higher than those in the tolerance group (P<0.05). Multivariate logistic regression analysis showed that hyperglycemia, postoperative drug use, abnormal serum albumin and intestinal injury were independent risk factors for EN intolerance (P<0.05). The nomogram model showed that the presence of hyperglycemia accounted for 87.5 points, the presence of postoperative drug use accounted for 100 points, the presence of abnormal serum albumin accounted for 69.8 points, and the presence of intestinal injury accounted for 80.00 points. H-L fit test showed that, χ2=8.318, P=0.304, the slope of the calibration curve approached 1, the ROC curve was evaluated for model discrimination, the area under the curve was 0.840 (95%CI: 0.773-0.893), and the sensitivity and specificity were 80.95% and 76.72%, respectively. Conclusion  Hyperglycemia, postoperative drug use, abnormal serum albumin and intestinal injury are the influencing factors of EN intolerance in patients with esophageal cancer after surgery. The nomogram model constructed by this method has a higher clinical significance in predicting the risk of EN intolerance in patients with esophageal cancer after surgery.