Factors and predictive model for malnutrition in poststroke disabled patients: A multicenter cross-sectional study
Hongji Zeng1, Ang Cai2, Weijia Zhao1
1School of Public Health, Zhengzhou University, Zhengzhou, China.
Background:
Although malnutrition has been shown to influence the clinical outcome of poststroke disabled patients, the associated factors and the prediction model have yet to be uncovered.
Objectives:
This study aims to assess the current prevalence and factors associated with malnutrition in poststroke disabled patients and establish a prediction model.
Methods:
A multicenter cross-sectional survey among Chinese poststroke disabled patients (≥18 y old) was conducted in 2021. Information on patients' basic data, medical history, Barthel Index, dysphagia, and nutritional status was collected. A multivariable logistic regression model was used to identify the factors that influence malnutrition. Nomogram was developed and internal validation was conducted using 5-fold cross-validation. External validation was performed using the data from a preliminary survey. Receiver operating characteristic (ROC) analysis, calibration curves, and decision curve analysis (DCA) were used to analyze the predictive value of the nomogram.
Results:
Four hundred fifty-seven cases were enrolled, with the prevalence of malnutrition as 71.77%. Age (aOR = 1.039, 95% CI: 1.006-1.078), pulmonary infection (aOR = 4.301, 95% CI: 2.268-14.464), dysphagia (aOR = 24.605, 95% CI: 4.966-191.058), total intake volume (aOR = 0.997, 95% CI: 0.995-0.999), Barthel Index (aOR = 0.965, 95% CI: 0.951-0.980), and nasogastric tube (aOR = 16.529, 95% CI: 7.418-52.518) as nutrition support mode (compared to oral intake) were identified as the associated factors of malnutrition in stroke-disabled patients (P < 0.05). ROC analysis showed that the area under the curve (AUC) for nomogram was 0.854 (95% CI: 0.816-0.892). Fivefold cross-validation showed the mean AUC as 0.829 (95% CI: 0.784-0.873). There were no significant differences between predicted and actual probabilities. The DCA revealed that the model exhibited a net benefit when the risk threshold was between 0 and 0.4.
Conclusions:
Age, pulmonary infection, dysphagia, nutrition support mode, total intake volume, and Barthel Index were factors associated with malnutrition in stroke-related disabled patients. The nomogram based on the result exhibited good accuracy, consistency and values.
