Prognostic value of stress hyperglycemia ratio in critically Ill patients with acute kidney injury: a machine learning-driven retrospective cohort analysis.
Higher stress hyperglycemia ratios are linked to increased mortality risk in critically ill patients with acute kidney injury, according to a study of 5,555 ICU patients.
Where it sits
this study against the rest of the lypressin corpusSummary and findings
This study investigated the association between the stress hyperglycemia ratio (SHR) and short-term mortality in critically ill patients with acute kidney injury (AKI). A total of 5,555 ICU patients were analyzed, and machine learning models were developed to assess mortality risk. The findings suggest that higher SHR levels are linked to increased mortality risk, but no specific therapeutic claims are made.
Abstract
<h4>Objective</h4>Acute kidney injury (AKI) is a serious complication in critically ill patients, contributing to high morbidity and mortality. The stress hyperglycemia ratio (SHR), defined as the ratio of admission blood glucose to estimated average glucose derived from HbA1c, has been previously linked to outcomes in cardiovascular and cerebrovascular conditions, but its role in predicting short-term mortality in ICU patients with AKI remains unclear. This study explores the association between SHR and short-term mortality, and developed machine learning models to improve prognostic accuracy. <b>Methods:</b> We retrospectively analyzed 5,555 ICU patients with AKI from the MIMIC-IV (v3.1) database. Patients were divided into SHR quartiles. Primary outcomes were 30- and 90-day all-cause mortality. The association between SHR and mortality was assessed using restricted cubic spline (RCS) modeling, Cox regression, and Kaplan-Meier analysis. Feature selection was performed using the Boruta algorithm and LASSO regression. Twelve machine learning models were developed and systematically compared for 30-day mortality prediction. Model performance was evaluated using multiple metrics, and the best-performing model was further assessed using decision curve analysis and calibration analysis. SHAP analysis was applied to interpret the contributions of individual features. <b>Results:</b> A U-shaped relationship was found between SHR and mortality, with higher SHR levels significantly increasing the risk of both 30-day and 90-day death (HR > 1, p < 0.01). Subgroup analyses confirmed SHR's predictive reliability across multiple populations. Among the models, LightGBM demonstrated the best overall predictive performance (AUC = 0.864), outperforming traditional ICU scoring systems. The model showed good calibration and a favorable net clinical benefit. <b>Conclusion:</b> SHR is an independent, nonlinear predictor of short-term mortality in ICU patients with AKI. Machine learning models incorporating SHR, especially LightGBM, significantly improve risk stratification, offering valuable support for early identification and personalized ICU management.<h4>Supplementary information</h4>The online version contains supplementary material available at 10.1007/s13755-026-00443-0.
Background
The study addresses the prognostic value of the stress hyperglycemia ratio (SHR) in critically ill patients with acute kidney injury (AKI). Previous research has linked SHR to outcomes in cardiovascular and cerebrovascular conditions, but its role in predicting mortality in ICU patients with AKI was not well established. Understanding this relationship could enhance risk stratification and management in critical care settings.
Methods
This retrospective cohort study analyzed 5,555 ICU patients with AKI from the MIMIC-IV (v3.1) database. Patients were categorized into quartiles based on their SHR. The primary outcomes measured were 30-day and 90-day all-cause mortality, assessed using Cox regression, Kaplan-Meier analysis, and restricted cubic spline modeling.
Results
The study found a U-shaped relationship between SHR and mortality, with higher SHR levels significantly associated with an increased risk of both 30-day and 90-day mortality (HR > 1, p < 0.01). Subgroup analyses confirmed the predictive reliability of SHR across various populations. The LightGBM model achieved an AUC of 0.864, indicating strong predictive performance.
Interpretation
The findings suggest that SHR is an independent predictor of short-term mortality in ICU patients with AKI, which aligns with previous literature indicating the importance of hyperglycemia in critical illness. However, the clinical significance of the effect size and the reliance on machine learning models may limit direct application in practice. Potential confounds include the retrospective nature of the study and the use of a single database.
Key findings
- HR > 1 for SHR levels, p < 0.01 for both 30-day and 90-day mortality.
- AUC = 0.864 for LightGBM model performance.
- 5,555 ICU patients with AKI analyzed.
Limitations
- Retrospective study design.
- Single database analysis may limit generalizability.
- Machine learning models may not directly translate to clinical practice.