脑卒中伴吞咽障碍患者行胃造瘘术后营养状况的影响因素及列线图分析  

Identification of influencing factors and development of a nomogram for nutrition status after percutaneous endoscopic gastrostomy in stroke patients with swallowing disorders

作　　者：高艳芳 康春博 邓海连 高雅 Yan-Fang Gao;Chun-Bo Kang;Hai-Lian Deng;Ya Gao(Gastrointes-tinal Rehabilitation Center,Beijing 100144,China;Neurology Rehabilitation Center,Beijing Rehabilitation Hospital Affiliated to Capital Medical University,Beijing 100144,China)

机构地区：[1]首都医科大学附属北京康复医院胃肠康复中心,北京市100144 [2]首都医科大学附属北京康复医院神经康复中心,北京市100144

出　　处：《世界华人消化杂志》2023年第21期904-912,共9页World Chinese Journal of Digestology

基　　金：北京市残疾人联合会项目,No.GFTC18AA0252-01;首都医科大学附属医院青年项目,No.2021-039.

摘　　要：背景营养不良是影响脑卒中伴吞咽障碍患者预后质量及增加死亡率的因素之一.内镜下经皮胃造瘘管术(percutaneous endoscopic gastrostomy,PEG)为此类患者提供了持续改善营养状况的重要途径.但是PEG术后患者营养状况及风险因子尚不完全明确,值得关注.目的了解脑卒中伴吞咽障碍患者行PEG术后的营养状态,并探讨相关影响因素,为加强脑卒中伴吞咽障碍患者的营养干预提供参考.方法纳入2018-06/2022-06间于首都医科大学附属北京康复医院接受PEG手术并完成12 wk随访的108位脑卒中伴吞咽障碍患者为研究对象.利用营养风险筛查2002(nutrition risk screening 2002,NRS2002)评分结果将研究对象分为高营养风险组和低营养风险组.基于随机森林的机器学习算法和Logistic回归分析脑卒中伴吞咽障碍患者PEG术后12 wk营养风险的影响因素并构建列线图预测模型,通过受试者工作特征曲线(receiver operating curve,ROC)、校准曲线和决策曲线评价列线图模型的准确性与效能.结果108名患者中,低营养风险组有70例(占64.8%)、高营养风险组有38例(占35.2%).Logistic回归结果表明,年龄≥60岁、家庭经济条件差、存在感染与高营养风险存在显著正相关(均β>0,P<0.05);接受了进一步医治、高血清白蛋白水平、体质量指数与高营养风险存在显著负相关(均β<0,P>0.05).列线图预测营养风险的曲线下面积为0.819,校准曲线贴合良好.结论由年龄(≥60 vs<60岁)、家庭经济条件(良好vs差)、有无接受进一步医疗治疗(有vs无)、感染情况(有vs无)等变量构建的列线图能够预测脑卒中伴吞咽障碍患者行PEG术后的营养风险,对加强行PEG脑卒中伴吞咽障碍患者的营养管理有一定指导意义.BACKGROUND Malnutrition is one of the factors that affect the prognosis and increase mortality in stroke patients with swallowing disorders.Percutaneous endoscopic gastrostomy(PEG)provides an important way to continuously improve the nutritional status of these patients.However,the nutritional status and risk factors of PEG patients after operation are not completely clear.AIM To investigate the nutritional status after PEG in stroke patients with swallowing disorders and to explore the related influencing factors,in order to provide reference for nutritional intervention in such patients.METHODS From June 2018 to June 2022,a total of 108 stroke patients with swallowing disorders who underwent PEG and completed 12-week follow-up at Beijing Rehabilitation Hospital Affiliated to Capital Medical University were included as the research subjects.According to the results of Nutrition Risk Screening 2002(NRS 2002),the patients were divided into either a high nutritional risk group or a low nutritional risk group.Random Forest machine learning algorithm and Logistic regression method were used to evaluate the influencing factors of nutritional status of stroke patients with swallowing disorders after PEG,and a nomogram prediction model was established.The accuracy and efficiency of the nomogram model were evaluated by receiver operating characteristic(ROC)curve,calibration curve,and decision curve analyses.RESULTS Among the 108 patients,there were 70 cases(64.8%)in the low nutritional risk group and 38(35.2%)in the high nutritional risk group.Logistic regression results showed that age≥60 years,poor family economic conditions,and local/systemic infection were significantly positively correlated with high nutritional risk(β>0,P<0.05).Further medical treatment,high serum albumin level,and body mass index were negatively correlated with high nutritional risk(β<0,P>0.05).The area under the ROC curve of the developed nomogram was 0.819,and the calibration curve fitted well.CONCLUSION The nomogram constructed based on var

关 键 词：脑卒中 吞咽障碍 经皮内镜下胃造瘘置管术 营养状况 

分 类 号：R656.61[医药卫生—外科学]