天津医药 ›› 2026, Vol. 54 ›› Issue (5): 484-488.doi: 10.11958/20252863

• 临床研究 • 上一篇    下一篇

结直肠癌术后患者肠内营养喂养不耐受的影响因素及预警模型的构建

彭静1(), 余小倩2△()   

  1. 1 南京医科大学第四附属医院胃肠外科(邮编 210009)
    2 南京医科大学第四附属医院放疗科(邮编 210009)
  • 收稿日期:2025-09-05 修回日期:2026-03-03 出版日期:2026-05-15 发布日期:2026-05-13
  • 通讯作者: E-mail:372496859@qq.com
  • 作者简介:彭静(1985),女,主管护师,主要从事普外科疾病营养管理方面研究。E-mail:313570092@qq.com
  • 基金资助:
    江苏省重点研发计划专项资金项目(BE2020723)

Influencing factors and risk prediction model of enteral nutrition intolerance in postoperative colorectal cancer patients

PENG Jing1(), YU Xiaoqian2△()   

  1. 1 Department of Gastrointestinal Surgery, the Fourth Affiliated Hospital of Nanjing Medical University, Nanjing 210009, China
    2 Department of Radiotherapy, the Fourth Affiliated Hospital of Nanjing Medical University, Nanjing 210009, China
  • Received:2025-09-05 Revised:2026-03-03 Published:2026-05-15 Online:2026-05-13
  • Contact: E-mail:372496859@qq.com

摘要:

目的 分析结直肠癌术后患者肠内营养喂养不耐受(EFI)发生的影响因素并构建预测模型。方法 选取结直肠癌术后患者392例,根据7∶3的比例分为建模组(274例)和验证组(118例)。根据建模组在术后是否发生肠内营养喂养不耐受分为EFI组(88例)和正常组(186例);收集患者的基本资料、药物使用情况、营养状况、术后首次实验室指标。Logistic回归分析结直肠癌术后患者发生EFI的影响因素,建立风险预测评分模型,采用受试者工作特征(ROC)曲线评价模型的区分度,校准曲线评价模型的校准度。结果 EFI组年龄、共病指数、白细胞计数、使用镇痛药物、营养不良、血清白蛋白<35 g/L、肿瘤类型为溃疡型、浸润型、肿瘤分期Ⅲ—Ⅳ期比例均高于正常组(P<0.01)。多因素Logistic回归分析发现,年龄≥70岁、共病指数≥2、肿瘤类型溃疡型/浸润型、肿瘤分期Ⅲ—Ⅳ期、使用镇痛药物、营养不良、血清白蛋白<35 g/L及白细胞计数>10×109/L为结直肠癌术后发生EFI的危险因素。验证组ROC曲线下面积为0.834(95%CI:0.756~0.920),敏感度为92.2%,特异度为76.4%。验证组校准曲线接近参考曲线,提示该模型有很好的预测效果。结论 本研究构建的结直肠癌术后患者EFI预警模型可为及时实施预防策略,降低营养不耐受风险提供参考。

关键词: 结直肠肿瘤, 肠道营养, 列线图, 喂养不耐受, 预测模型

Abstract:

Objective To analyze the influencing factors of enteral nutrition feeding intolerance (EFI) in patients after colorectal cancer surgery and construct a predictive model. Methods A total of 392 patients who underwent colorectal cancer surgery were selected and divided into the modeling group (274 cases) and the validation group (118 cases) at a ratio of 7∶3. The modeling group was further divided into the EFI group (88 cases) and the normal group (186 cases) based on whether EFI occurred after surgery. Basic patient information, drug use, nutritional status and the first postoperative laboratory indicators (serum albumin, hemoglobin and white blood cell count) were collected. Logistic regression was used to analyze the influencing factors of EFI in patients after colorectal cancer surgery, and a risk prediction scoring model was established. The discrimination of the model was evaluated by the receiver operating characteristic (ROC) curve, and the calibration of the model was evaluated by the calibration curve. Results The age, comorbidity index, white blood cell count, use of analgesic drugs, malnutrition and the proportion of patients with serum albumin < 35 g/L in the EFI group were all higher than those in the normal group. Moreover, there were statistically significant differences in tumor type and tumor stage between the two groups (P<0.01). Multivariate Logistic regression analysis revealed that age≥70 years, Charlson comorbidity index ≥2, ulcerative/infiltrative tumor type, tumor stage Ⅲ-Ⅳ, use of analgesic drugs, malnutrition, serum albumin <35 g/L and white blood cell count > 10×109/L were risk factors for EFI after colorectal cancer surgery. The area under the ROC curve of the validation group was 0.834 (95%CI: 0.756 - 0.920), with a sensitivity of 92.2% and a specificity of 76.4%. The calibration curve of the validation group was close to the reference curve, suggesting that this model has a very good predictive effect. Conclusion The constructed EFI warning model for patients after colorectal cancer surgery can help identify high-risk patients for feeding intolerance in clinical practice and provide a reference for the timely implementation of preventive strategies and reduction of the risk of nutritional intolerance.

Key words: colorectal neoplasms, enteral nutrition, nomogram, feeding intolerance, prediction modeling

中图分类号: