Analysis of risk factors and development of predictive model for malnutrition in patients with traumatic brain injury.

Abstract

Malnutrition is a highly prevalent complication in patients with traumatic brain injury (TBI), and it is closely related to the prognosis of patients. Accurate identification of patients at high risk of malnutrition is essential. Therefore, we analyzed the risk factors of malnutrition in patients with TBI and developed a model to predict the risk of malnutrition. A retrospective collection of 345 patients with TBI, and they were divided into malnutrition and comparison groups according to the occurrence of malnutrition. Univariate correlation and multifactor logistic regression analyses were performed to determine patients' malnutrition risk factors. We used univariate and logistic regression (forward stepwise method) analyses to identify significant predictors associated with malnutrition in patients with TBI and developed a predictive model for malnutrition prediction. The model's discrimination, calibration, and clinical utility were evaluated using the receiver operating characteristic (ROC) curve, calibration plots, and decision curve analysis (DCA). A total of 216 patients (62.6%) developed malnutrition. Multifactorial logistic regression analysis showed that pulmonary infection, urinary tract infection, dysphagia, application of NGT, GCS score ≤ 8, and low ADL score were independent risk factors for malnutrition in patients with TBI (P < 0.05). The area under the curve of the model was 0.947. Calibration plots showed good discrimination of model calibration. DCA showed that the column line plot models were all clinically meaningful when nutritional interventions were performed over a considerable range of threshold probabilities (0-0.98). Malnutrition is widespread in patients with TBI, and the nomogram is a good predictor of whether patients develop malnutrition.