Xu Yingying, Hu Qiongyuan, Pei Dandan, Zhang Yin, Zhu Huanhuan, Hui Yan, Guan Wenxian, Xu Meiling, Chen Li
Division of Gastric and Hernia Surgery, Department of General Surgery, Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School, Nanjing, China.
Department of Urology, Nanjing Drum Tower Hospital, Affiliated Hospital of Nanjing University Medical School, Nanjing, China.
Front Nutr. 2024 Nov 26;11:1480390. doi: 10.3389/fnut.2024.1480390. eCollection 2024.
The incidence of enteral feeding intolerance (ENFI) in the early postoperative period is high in patients after gastric cancer resection due to the characteristics of surgical traumatic stress and changes in the physiological structure of the digestive tract, and the current evaluation of ENFI after gastric cancer resection mostly depends on the symptoms and complaints of patients after gastric cancer resection, which is lagging and subjective. Early accurate and objective prediction of the risk of early ENFI after gastric cancer resection is critical to guide clinical enteral nutrition practice.
This study included 470 patients who underwent radical gastric cancer surgery at the Division of Gastric Surgery of a tertiary hospital in Nanjing, Jiangsu Province, between November 2021 and October 2022. The patients were divided into a training set ( = 329) and a validation set ( = 141) in a 7:3 ratio. The predictors were first screened through Lasso regression. Subsequently, multifactorial logistic regression analysis was used to establish a model for predicting patients' early ENFI column charts after gastric cancer resection. Internal and external validation of the model were performed on the training set and validation set data, respectively, including plotting the area under the curve (AUC) of the receiver operating characteristic (ROC) curve and calibration curves to assess the differentiation and calibration of the prediction model. The Hosmer-Lemeshow test was also used to assess the fit of the model.
The incidence of early ENFI in postoperative patients with gastric cancer was 44.68% in the training set and 43.97% in the validation set. The final predictors entered into the model were enteral nutrition solution type (OR1 = 1.31/OR2 = 7.23), preoperative enteral nutrition pre-adaptation technique (OR = 0.29), surgical approach (OR = 2.21), preoperative Profile of Mood State-Short Form score (OR = 5.07), and intra-abdominal pressure (OR = 6.79). In the internal validation, the AUC was 0.836, the 95% CI ranged from 0.792 to 0.879, the Hosmer-Lemeshow test showed χ = 4.368 and = 0.737, the sensitivity was 0.775, and the specificity was 0.741. In the external validation, the AUC was 0.853, the 95% CI ranged from 0.788 to 0.919, the Hosmer-Lemeshow test showed χ = 13.740 and = 0.089, the sensitivity was 0.785, and the specificity was 0.823.
The Nomogram model of early ENFI in postoperative patients with gastric cancer, constructed on the basis of Lasso-logistic regression, had good predictive efficacy and may serve as a reference for healthcare professionals to identify high-risk patients with early ENFI after gastrectomy.
由于手术创伤应激的特点和消化道生理结构的改变，胃癌切除术后患者术后早期肠内营养不耐受（ENFI）的发生率较高，目前对胃癌切除术后ENFI的评估大多依赖于胃癌切除术后患者的症状和主诉，具有滞后性和主观性。早期准确、客观地预测胃癌切除术后早期ENFI的风险对于指导临床肠内营养实践至关重要。
本研究纳入了2021年11月至2022年10月期间在江苏省南京市某三级医院胃外科接受胃癌根治术的470例患者。患者按7:3的比例分为训练集（n = 329）和验证集（n = 141）。首先通过Lasso回归筛选预测因素。随后，采用多因素logistic回归分析建立胃癌切除术后患者早期ENFI柱状图预测模型。分别对训练集和验证集数据进行模型的内部和外部验证，包括绘制受试者工作特征（ROC）曲线的曲线下面积（AUC）和校准曲线，以评估预测模型的区分度和校准度。还采用Hosmer-Lemeshow检验评估模型的拟合度。
训练集中胃癌术后患者早期ENFI的发生率为44.68%，验证集中为43.97%。最终纳入模型的预测因素为肠内营养溶液类型（OR1 = 1.31/OR2 = 7.23）、术前肠内营养预适应技术（OR = 0.29）、手术方式（OR = 2.21）、术前简式情绪状态量表评分（OR = 5.07）和腹内压（OR = 6.79）。在内部验证中，AUC为0.836，95%CI为0.792至0.879，Hosmer-Lemeshow检验显示χ² = 4.368，P = 0.737，灵敏度为0.775，特异度为0.741。在外部验证中，AUC为0.853，95%CI为0.788至0.919，Hosmer-Lemeshow检验显示χ² = 13.740，P = 0.089，灵敏度为0.785，特异度为0.823。
基于Lasso-logistic回归构建的胃癌术后患者早期ENFI列线图模型具有良好的预测效能，可为医护人员识别胃癌切除术后早期ENFI的高危患者提供参考。
