ICU患者肠内营养喂养不耐受风险预测模型的系统评价

Systematicevaluation of risk prediction models for enteral feeding intolerance in ICU patients

目的系统评价ICU患者肠内营养喂养不耐受(feeding intolerance,FI)风险预测模型。方法计算机检索中国知网、万方数据库、中国生物医学文献数据库、PubMed、CINAHL、Embase、Web of science、Cochrane Library、Medline数据库中与主题相关的文献,检索时限为建库至2023年7月1日。由2名研究者独立筛选文献及提取数据,并对文献进行质量评价。结果纳入9篇危重患者肠内营养FI风险预测模型构建研究,共13个预测模型。13个模型的受试者工作特征曲线下面积或C指数为0.70~0.921。9项研究的整体偏倚风险较高,4项研究的适用性较好,偏倚风险主要源于未选择合适的数据来源、样本量不合理、未报告缺失数据及处理方法、基于单变量分析筛选预测变量、缺少模型性能评估。多变量模型中包含较多的预测因子为年龄、白蛋白水平、APACHE Ⅱ评分、机械通气、空腹血糖及肠内营养开始时间等。结论ICU患者肠内营养FI风险预测模型研究仍处于初步发展阶段,模型仍存在高偏倚风险。未来研究应规范报告建模过程并进一步验证模型在临床实践中的适用性及可行性,为制定相关预防策略提供参考依据。

Objective To systematically evaluatethe risk prediction model of enteral feeding intolerance(FI)in ICU patients.Methods The literatures related to the topic were searched by computer inCNKI,Wanfang Data,CBM,PubMed,CINAHL,Embase,Web of science,Cochrane Library,and Medline,and the search period was from establishment of the databaseto July 1,2023.Two researchers independently screened the literature and extracted the data,and then evaluated the quality of the literature.Results A total of 9 studies on the risk prediction model of enteral FI in ICU patientswere included,with a total of 13 prediction models.The area under the ROC curve o of the models or index C ranged from 0.70 to 0.921.The overall risk of bias in 9 studies were high,and the applicability in 4 studies wasgood,which was mainly due to failure to select appropriate data sources,unreasonable sample sizes,failure to reportmissing data and processing methods,screening predictors based on univariate analysis,and lack of model performance evaluation.The predictors included in the multivariate model were mostly age,albumin,APACHEⅡscore,mechanical ventilation,fasting blood glucose,and start enteral nutrition time,etc.Conclusion The study of risk prediction model of enteral FI in ICU patients is still in the preliminary development stage,and the models still have a high risk of bias.In the Future,studies should standardize the report modeling process and further verify the applicability and feasibility of the model in clinical practice,so asto provide certain referencesfor the development of relevant prevention strategies.