Investigation of malnutrition in stroke patients with dysphagia and construction of a risk prediction model

Wei Dandan; Xie Dongqin; Yan Lanyun; Shen Feifei; Yu Ling; Gu Suliang; Wang Peiying

doi:10.11986/j.issn.1673-873X.2025.09.06

Abstract：Objective To investigate the incidence of malnutrition in stroke patients with dysphagia and to construct a risk prediction model.Methods Convenience sampling was used to select 200 stroke patients with dysphagia treated at the First Affiliated Hospital with Nanjing Medical University from February 2021 to March 2024.The incidence of malnutrition was recorded,and the patients were assigned to a training set(n=140)and a validation set(n=60)in a 7∶3 ratio.In the train-ing set,univariate and multivariate logistic regression analyses were used to identify risk factors for malnutrition in stroke patients with dysphagia,and a nomogram model was constructed based on these factors.The model's discrimination and calibration were assessed using receiver operating characteristic(ROC)curves and the Hosmer-Lemeshow test.The model was further validated using data from the validation set.Results In this study,the incidence of malnutrition among the 200 patients was 41.00％(82/200).In the training set,the malnutrition group had a higher proportion of patients aged＞60 years,with a body mass index(BMI)＜18.5 kg/m2 at admission,lung infections,severe dependence in daily life,and a Grade Ⅴ result in the water swallow test.They also had lower serum albumin(Alb)levels compared to the well-nourished group,with all differences being statistically significant(P＜0.05).Multivariate logistic regression analysis indicated that age＞60 years(OR=3.312,95％CI:1.317-8.326),severe dependence in daily life(OR=12.110,95％CI:1.343-109.206),Grade V in the water swallow test(OR=4.610,95％CI:1.298-16.379),BMI＜18.5 kg/m2 at admission(OR=3.458,95％CI:1.054-11.347),and decreased Alb levels(OR=0.629,95％CI:0.522-0.757)were independ-ent predictors of malnutrition after stroke.Based on these five variables,a risk prediction model for malnutrition in stroke patients with dysphagia was developed.The logistic regression equation was:Logit(P)=1.198 × Age(＞60 years)+2.494 × Barthel Index(severe dependence)+1.528 × Water Drinking Test Grade(Ⅴ)—0.464 × Alb+1.241 × BMI at ad-mission(＜18.5 kg/m2)+14.959.The area under the curve(AUG)for the model in the training set was 0.865(95％CI:0.796-0.923),with a sensitivity of 74.39％and specificity of 86.21％.The Hosmer-Lemeshow goodness-of-fit test for the training set model was x2=10.535,P=0.229.In the validation set,AUC was 0.830(95％CI:0.724-0.936),with a sensitivity of 61.11％and specificity of 87.50％,and the Hosmer-Lemeshow goodness-of-fit test was x2=11.280,P=0.186.Conclusion The incidence of malnutrition in stroke patients with dysphagia remains high.Age＞60 years,se-vere dependence in daily life,Grade V in the water swallow test,BMI＜18.5 kg/m2 at admission,and decreased Alb levels are independent risk factors for malnutrition.The risk prediction model based on these indicators has good discrimination and calibration,and can be used for early screening of dysphagia in stroke patients in clinical practice.