Stress hyperglycemia ratio as an important predictive indicator for severe disturbance of consciousness and all-cause mortality in critically ill patients with cerebral infarction: a retrospective study using the MIMIC-IV database.

Abstract

Background: Stress hyperglycemia ratio (SHR) has been linked to prognosis of cerebrovascular diseases. Nevertheless, the association between SHR and severe disturbance of consciousness (DC) and mortality among patients with cerebral infarction remains explored. This study seeks to assess the predictive potential of SHR for severe DC and mortality among patients with cerebral infarction.

Methods: We identified individuals diagnosed with cerebral infarction within the MIMIC-IV database. We employed logistic regression to examine the correlation between the SHR index and the severity of patients' consciousness disturbance, as well as in-hospital mortality. Furthermore, we employed restricted cubic spline curves to explore potential non-linear relationships between the SHR index and outcome measures. To assess the predictive performance of the SHR index and admission blood sugar level on outcome indicators, we compared receiver operating characteristic (ROC) curves.

Results: A non-linear relationship existed between SHR and the risk of severe disturbance of consciousness, while there was a linear relationship with all-cause mortality. The AUC value for predicting severe disturbance of consciousness by the SHR index is 0.5419 (95% CI: 0.5188-0.5661). The AUC value for predicting in-hospital mortality based on the SHR index is 0.6264 (95% CI: 0.5881-0.6662). It is superior to single admission blood sugar level. In addition, SHR has an incremental impact on evaluating various diseases in predicting severe disturbance of consciousness and all-cause mortality in critically ill patients with cerebral infarction.

Conclusions: SHR is an important predictive indicator for severe disturbance of consciousness and all-cause mortality of patients with cerebral infarction.