神经重症病人肠内营养不耐受风险预测模型构建及验证

    Construction and validation of a risk prediction model for enteral nutrition feeding intolerance in neurology intensive care unit patients

    • 摘要: 目的:探讨神经重症病人肠内营养不耐受(ENFI)的相关影响因素,建立ENFI列线图。方法:选取2022年1月至2024年11月在重症医学科进行治疗的109例神经重症病人为研究对象。采用一般情况资料调査表、ENFI状况调査表进行持续5 d的数据收集,采用二元logistic回归方程进行多因素分析,并构建列线图预测模型,使用受试者工作特征(ROC)曲线评估预测模型效能。结果:40例(36.70%)病人发生ENFI。ENFI和耐受组病人在格拉斯哥昏迷(GCS)评分、是否合并糖尿病、血K+水平、血清白蛋白水平、有无机械通气、是否使用药物和营养风险等方面比较,差异均有统计学意义(P<0.05)。logistic回归分析结果显示,合并糖尿病、使用血管活性药物和使用机械通气是神经重症病人发生ENFI的独立危险因素;使用质子泵抑制剂(PPIs)、GCS评分≥5分和血清白蛋白≥35 g/L是保护因素。列线图模型的区分度AUC(95%CI)为0.870(0.798~0.942),灵敏度为75.0%,特异度为91.3%。模型预测了良好的潜在临床效果。结论:合并糖尿病、使用血管活性药物和使用机械通气是神经重症病人发生ENFI的独立危险因素,使用PPIs、GCS评分≥5分、血清白蛋白≥35 g/L是保护因素。临床护理工作中可以通过关注以上因素来预测神经重症病人ENFI的发生,采取相关措施预防营养风险,改善病人预后。

       

      Abstract: Objective: To explore the related influencing factors of enteral nutrition feeding intolerance(ENFI) in neurology intensive care unit patients,and establish a nomogram of ENFI. Methods: From January 2022 to November 2024,109 neurology intensive care unit patients were selected as the research subjects.Data collection was conducted continuously for 5 days using the general information questionnaire and ENFI questionnaire.Multivariate analysis was performed using the binary logistic regression equation,and a nomogram prediction model was constructed.The receiver operating characteristic(ROC) curve was used to evaluate the efficacy of the prediction model. Results: ENFI occurred in 40 cases(36.70%) of patients.The differences in the Glasgow coma scale (GCS) score,combination with or without diabetes,blood K+ level,serum albumin level,having mechanical ventilation or not,using drugs or not and nutritional risk were statistically significant between the ENFI group and tolerance group(P<0.05).The results of logistic regression analysis showed that the coexisting diabetes,use of vasoactive drugs and use of mechanical ventilation were the independent risk factors of ENFI in neurology intensive care unit patients.The use of proton pump inhibitors(PPIs),GCS score of ≥5 points,and serum albumin level ≥35 g/L were the protective factors.The AUC(95%CI) of the nomogram model was 0.870(0.798-0.942),the sensitivity was 75.0%,and the specificity was 91.3%.The model predicted a good potential clinical effects. Conclusions: Diabetes mellitus and use of vasoactive drugs and mechanical ventilation are the independent risk factors of ENFI in neurology intensive care unit patients,while the use of PPIs,GCS score ≥5 points and serum albumin ≥35 g/L are the protective factors.In clinical nursing work,the occurrence of ENFI in neurology intensive care unit patients can be predicted by paying attention to the above factors,and the relevant measures can be taken to prevent nutritional risks,and improve the prognosis of patients.

       

    /

    返回文章
    返回