Abstract
AIMS: This study aimed to investigate the association between the Stress Hyperglycemia Ratio (SHR) and mortality risk in critically ill patients with cardiometabolic multimorbidity (CMM), and to explore the potential mediating roles of white blood cell count (WBC), albumin, and creatinine. METHODS: We conducted a multicenter retrospective cohort study using data from the MIMIC-IV database (n = 1,214) for derivation and the NWICU database (n = 686) for external validation. The relationship between SHR and all-cause/28-day mortality was analyzed using multivariate logistic regression, restricted cubic splines, and machine learning models. Mediation analysis was performed to assess the roles of WBC, albumin, and creatinine. RESULTS: A nonlinear relationship with threshold effects was observed between SHR and mortality. In the fully adjusted model, the highest SHR tertile (Q3) was independently associated with increased risks of all-cause mortality (OR = 1.58, 95% CI: 1.13–2.20) and 28-day mortality (OR = 1.98, 95% CI: 1.24–3.18). Random Forest and XGBoost models demonstrated the best predictive performance for 28-day mortality (AUC = 0.765). Calibration curves confirmed the reliability of the risk estimates. Mediation analysis indicated that WBC and creatinine may partially mediated the adverse effects of SHR, while albumin exerted a protective mediating effect. CONCLUSIONS: SHR is a significant and independent predictor of mortality in critically ill patients with CMM, exhibiting a nonlinear association. WBC, creatinine, and albumin may partially mediate this relationship, offering insights for future mechanistic studies. However, these findings are specific to the ICU population and warrant further validation in other cohorts. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12872-026-05580-8.