Association between the Stress Hyperglycemia Ratio and Multiple Organ Dysfunction Syndrome Incidence in Trauma Patients: A Retrospective Cohort Study and Predictive Model Establishment Based on Machine Learning

Abstract

BackgroundTrauma patients experience significant stress states, leading to physiological and pathological changes. Severe trauma may result in multiple organ dysfunction syndrome (MODS). This study aims to evaluate the association between the stress hyperglycemia ratio (SHR) and MODS in trauma patients.MethodsClinical data from 784 trauma patients were extracted from the MIMIC-IV (3.1) database. Based on clinical diagnoses, trauma patients were divided into a diabetic trauma group and a non-diabetic trauma group. Each group was further stratified into three subgroups according to SHR tertiles. The outcome was the development of MODS within 7 days of ICU admission. The association between SHR and MODS was analyzed using restricted cubic splines and logistic regression, with further validation through subgroup analyses. The Boruta algorithm was employed to assess the predictive capability of SHR, and machine learning algorithms were utilized to develop predictive models.ResultsData from 784 trauma patients were analyzed. In the non-diabetic trauma group, restricted cubic spline curves revealed a U-shaped association between SHR and the risk of MODS development. Compared to lower SHR, higher SHR was significantly associated with an increased risk of MODS in non-diabetic trauma patients (OR > 1, P < 0.05). Boruta feature selection demonstrated that SHR in non-diabetic trauma patients exhibited a higher Z-score, and the model constructed using the Support Vector Machine (SVM) algorithm showed optimal performance (AUC = 0.908). In the diabetic trauma group, no significant relationship was observed between elevated SHR and MODS occurrence.ConclusionIn the non-diabetic trauma group, elevated SHR showed a significant association with MODS occurrence, whereas no significant association was observed between high SHR and MODS in the diabetic trauma group.

Note:
Funding declaration: None.

Conflict of Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Keywords: Trauma, MODS, Stress hyperglycemia ratio, Machine learning