Assessment of stress hyperglycemia ratio as a prognostic indicator in geriatric intensive care patients experiencing delirium: a retrospective cohort analysis.

BACKGROUND: The stress hyperglycemia ratio (SHR), which is designed to mitigate the influence of baseline glycemic levels, can effectively identify stress-induced hyperglycemia and predict adverse outcomes in seriously ill patients. While SHR possesses considerable clinical significance, research addressing its prognostic implications among elderly ICU patients experiencing delirium remains notably scarce. This study seeks to elucidate how SHR relates to all-cause mortality in this population. METHODS: The research used MIMIC-IV data and categorized participants according to SHR quartiles. Main outcomes were all-cause mortality at 28 and 365 days, with secondary outcomes at 90 and 180 days. Analyses involved quartile stratification, Kaplan-Meier curves, Cox regression, and restricted cubic spline (RCS) modeling to explore non-linear associations between SHR and mortality. RESULTS: This study analyzed 1088 elderly patients with a median age of 75.74 and found that those with higher SHR had an increased all-cause mortality at 28, 90, 180, and 365 days, as shown by KM analysis. Multivariable Cox regression confirmed greater mortality risk in the highest SHR group. RCS analysis revealed non-linear, S-shaped relationships, with inflection points at 0.85 and 1.77 for 28 day and 0.89 and 1.90 for 365 day mortality. Each standard deviation (SD) rise in SHR within the ranges of 0.85-1.77 for 28 day mortality and 0.89-1.90 for 365 day mortality corresponded to a 68% increase in 28 day mortality risk (p = 0.0053, 95% CI 1.17-2.41) and a 43% increase in 365 day mortality risk (p = 0.0114, 95% CI 1.08-1.89). CONCLUSION: In ICU elderly patients with delirium, a higher SHR index is strongly linked to greater all-cause mortality across diverse temporal spans. The SHR index helps evaluate disease severity and guide treatment decisions.