Association between stress hyperglycemia ratio and 28-day all-cause mortality in critically ill atrial fibrillation patients: a retrospective cohort study based on the MIMIC-IV database

Introduction

Atrial fibrillation (AF) is a supraventricular tachyarrhythmia characterized by uncoordinated atrial electrical activation and ineffective atrial contraction. Mild cases present with palpitations and chest tightness, leading to reduced quality of life, while severe cases may result in dangerous events such as stroke, arterial embolism, and heart failure (1,2). AF affects over 33 million people worldwide. In China, the incidence of AF among individuals aged ≥20 years is 0.2%, and over 11 years, the incidence of AF increased 20-fold (3,4). Abnormal glucose metabolism is a significant factor contributing to the high morbidity and mortality of cardiovascular diseases (5). Admission hyperglycemia is relatively common in AF patients and is independently associated with poor prognosis, with a stronger correlation observed in non-diabetic patients (6,7). Elevated blood glucose levels upon hospital admission represent a significant contributing factor to worse clinical outcomes, including greater mortality and complication rates, in those with AF (8,9). Therefore, early detection and appropriate management of stress hyperglycemia in AF patients hold significant clinical importance. Stress hyperglycemia (SH) refers to the transient elevation of blood glucose levels during acute illness. Previous researchers have used admission blood glucose (ABG) as an indicator of SH in acutely ill patients, but its value is influenced by both acute stress and chronic pre-admission glucose levels, making it an inaccurate measure of the magnitude of glucose elevation under stress. The stress hyperglycemia ratio (SHR), which adjusts ABG for chronic glucose levels, serves as a more accurate novel indicator for assessing stress hyperglycemia (10). SHR is calculated based on glycated hemoglobin (HbA1c) and post-ABG levels, reflecting acute hyperglycemic status. Previous studies have demonstrated that SHR levels are strongly associated with the prognosis of conditions such as acute myocardial infarction (11), acute kidney injury (AKI) (12), cardiac shock (13), heart failure (14,15), stroke (16), and coronavirus disease 2019 (17). However, limited research has reported on the impact of SHR on outcomes in critically ill AF patients (18). To facilitate early risk stratification, this study examines the relationship of the SHR with 28-day all-cause mortality in critically ill AF patients. The resulting evidence could thereby inform strategies for improving prognostic outcomes in this population. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1507/rc).

Methods

Data source

This study utilized the MIMIC-IV 2.2database (https://mimic.mit.edu/), an electronic health record dataset jointly developed by the Massachusetts Institute of Technology (MIT) and Beth Israel Deaconess Medical Center (BIDMC). The dataset includes clinical information from over 300,000 patients admitted to the ICU at BIDMC between 2008 and 2019. This repository encompasses a wide array of de-identified patient information, encompassing demographic profiles, physiological measurements (vital signs), diagnostic test results, records of pharmacological and surgical treatments, as well as longitudinal data on survival outcomes. The database was approved by the Beth Israel Deaconess Medical Center (BIDMC) Institutional Review Board (IRB #: 14280276), and all data were de-identified. The researcher (Q.Z.) completed the required CITI training for research ethics and has been granted access to the database. Since the database does not contain protected health information and all patient data are anonymized, informed consent was waived. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Inclusion and exclusion criteria

Inclusion criteria: (I) age 18 to 90 years; (II) first ICU admission with a diagnosis of AF.

Exclusion criteria: (I) ICU stay <24 hours; (II) no serum glucose or HbA1c results within 24 hours of admission; (III) for patients with multiple ICU admissions, only data from the first hospitalization were included. For detailed information on patient screening, refer to Figure 1.

Figure 1 Screening of critically ill AF patients. Q1: 0.2≤ SHR <0.89; Q2: 0.89≤ SHR <1.08; Q3: 1.08≤ SHR <1.36; Q4: 1.36≤ SHR <6.1. AF, atrial fibrillation; Glu, glucose; HbA1c, glycated hemoglobin; ICU, intensive care unit; MIMIC-IV, Medical Information Mart for Intensive Care IV; SHR, stress hyperglycemia ratio.

Outcome

All-cause mortality within 28 days was defined as the primary outcome measure. A key secondary outcome was mortality that occurred during the hospitalization period. For patients discharged before 28 days with survival, discharge day was omitted.

Data extraction

Extracted variables included demographic information (age, sex, body mass index, marital status), clinical diagnosis (based on ICD-9 code 42,731 or ICD-10 codes I4891, I480, I482, I481, I4820, I4819, I4821 and I4811 to identify AF), comorbidities (hypertension, type 2 diabetes, chronic kidney disease, AKI, pneumonia, hyperlipidemia, heart failure, coronary artery disease, coronavirus infection, and malignancy screened via Charlson Comorbidity Index), vital signs (heart rate, blood pressure, oxygen saturation), laboratory parameters [hemoglobin (HGB), platelet count (PLT), red cell distribution width (RDW), red blood cell count (RBC), white blood cell count (WBC), HbA1c, albumin, anion gap (AG), serum potassium, pH, partial pressure of carbon dioxide (PaCO₂), partial pressure of oxygen (PaO₂), D-dimer, alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen, uric acid, creatine kinase (CK), lactate dehydrogenase (LD), N-terminal pro-B-type natriuretic peptide (NT-proBNP), B-type natriuretic peptide, troponin T, troponin I], disease severity scores [Sequential Organ Failure Assessment (SOFA), Acute Physiology Score III (APS III), Simplified Acute Physiology Score II (SAPS II), Oxford Acute Severity of Illness Score (OASIS)], and outcome measures (28-day all-cause mortality, in-hospital mortality, etc.). The SHR was calculated using the formula SHR = [admission blood glucose (mg/dL)] / [28.7 × HbA1c (%) − 46.7]. Variables with missing rates >20% were excluded, and the remaining missing data were imputed using the random forest algorithm. All laboratory parameters were based on the first measurement within 24 hours of admission to ensure consistency.

Statistical analysis

As a retrospective cohort study, we included all eligible patients from the database without pre-calculating sample size. Extreme outliers in continuous variables were Winsorized (using the 99th percentile) to reduce their impact. During data preprocessing, variables with missing rates >20% were excluded, while those ≤20% retained were imputed into five imputed datasets using multiple imputation methods. Final analysis results were pooled according to Rubin’s rule. Variance inflation factors (VIFs) assessed multicollinearity, with all variables showing VIF values below 5. Samples were grouped by quartiles of SHR distribution for clinical interpretation. Normal distributions were represented as mean ± standard deviation with analysis of variance (ANOVA) comparisons, non-normal distributions described by median (interquartile range) with Mann-Whitney U-tests, and categorical variables presented as frequency (percentage) with Chi-squared or Fisher exact tests. Kaplan-Meier survival curves and log-rank tests compared 28-day survival rates across groups. Multivariate Cox proportional risk regression models (Models I, II, III) evaluated SHR’s association with mortality risk. Model II covariates were selected based on directed acyclic graph (DAG) analysis to adjust for confounding factors, as detailed in Figure S1. To assess the potential impact of unmeasured confounding factors on the study results, we further calculated the E-value. This metric represents the minimum magnitude of unmeasured confounding factors required to fully explain the observed exposure-outcome association. For the risk ratio (HR =1.89) between the SHR Q4 group and mortality, our point estimate of the E-value was 3.2, with a confidence interval (CI) lower limit of 2.1. This indicates that moderate to strong unmeasured confounding is needed to fully account for the observed association. All statistical analyses were performed using R software (version 4.3.1), with a two-tailed P<0.05 threshold for statistical significance.

Restricted cubic spline (RCS)

RCS analysis combined with Cox regression explored nonlinear associations between SHR and 28-day mortality in critically ill AF patients, adjusting for survival outcomes, SHR, age, weight, sex, SOFA score, and laboratory parameters. Optimal knots [3–7] were selected based on the lowest Akaike information criterion (AIC), with likelihood ratio tests confirming statistical significance (P<0.05).

Boruta algorithm and validation

Exploratory feature screening based on Boruta algorithm is used to evaluate the relative importance of SHR among many variables. Feature selection via the Boruta algorithm involved generating Z-scores for real and “shadow” features (randomly permuted). Variables consistently outperforming shadow features (red zone) were deemed “important”, while others (non-red zones) were excluded. Selected variables underwent multivariable Cox regression to assess clinical validity in predicting 28-day all-cause and in-hospital mortality.

Sensitivity analysis

Stratified by age (<65 vs. ≥65 years), sex, AKI, hypertension, and type 2 diabetes status, multivariable Cox regression evaluated SHR-mortality associations within subgroups. Models adjusted for age, weight, sex (except sex-stratified models), SOFA score, and laboratory parameters, with results visualized as forest plots [hazard ratio (HR) and 95% CI].

Results

Baseline characteristics

The study included 2,050 critically ill AF patients from MIMIC-IV [mean age 72.83±11.34 years, range 19–90 years; 59.32% male (n=1,216)]. Quartile stratification by SHR: Q1 (0.2≤ SHR <0.89, n=512), Q2 (0.89≤ SHR <1.08, n=512), Q3 (1.08≤ SHR <1.36, n=512), and Q4 (1.36≤ SHR <6.1, n=514). Baseline characteristics revealed significantly higher weight in Q4 (88.31±25.87 kg, P<0.05) and greater comorbidity burden, including type 2 diabetes (46.3%), coronary artery disease (55.64%), AKI (53.11%), heart failure (58.37%), myocardial infarction (26.26%), and chronic kidney disease (32.49%) (all P<0.05). Metabolic derangements in Q4 included elevated glucose (245.65±114.81 mg/dL), anion gap (16.14±4.97 mEq/L), serum potassium (4.40±0.92 mEq/L), lactate (3.8±1.9 vs. 1.6±0.8 mmol/L), alongside reduced albumin (3.21±0.56 g/dL) and RBC count (3.77±0.81 m/µL) (all P<0.001). Disease severity scores (SOFA: 5.58±3.56; APS III: 51.67±20.14; SAPS II: 41.22±13.08; OASIS: 34.47±8.62) were highest in Q4 (all P<0.001 vs. Q1). Q4 exhibited significantly higher 28-day all-cause mortality (20.23%, P<0.001), with a dose-response relationship (trend test P<0.001) (Table 1).

Continuous variables with normal distribution are represented by mean ± standard deviation, while those with non-normal distribution are described by median (interquartile range). Categorical variables are expressed as frequency (percentage). The missing rates for key variables are: age (1.2%), HGB (3.5%), fasting blood glucose (2.8%), PLT (4.1%), and lipid indicators [total cholesterol 2.9%, triglycerides 3.2%, high-density lipoprotein cholesterol (HDL-C) 3.0%, low-density lipoprotein cholesterol (LDL-C) 3.4%]. All other covariates showed missing rates below 2.0%. Statistical analysis employed multiple imputation to handle missing data, ensuring robustness of results. After multiple imputation, variable distributions remained stable with no significant bias observed.

Clinical outcomes

To further demonstrate this correlation, we conducted Cox regression analyses after grouping according to the SHR quartiles. The results from models I and II showed a significant increase in mortality risk at Quartile 4 compared to Quartile 1 (Tables 2,3).

The Kaplan-Meier curve (Figure 2) showed a difference in 28-day all-cause mortality between the SHR quartile groups, with significantly lower 28-day survival rates in the Q4 group compared to the Q1 group (log-rank test P<0.01).

Figure 2 28-day Kaplan-Meier survival curve. Q1: 0.2≤ SHR <0.89; Q2: 0.89≤ SHR <1.08; Q3: 1.08≤ SHR <1.36; Q4: 1.36≤ SHR <6.1.

RCS

RCS models were employed to examine the potential nonlinear relationship, controlling for covariates including age, weight, gender, comorbidities (chronic kidney disease, type 2 diabetes, and coronary artery disease), and SOFA score. Both the analyses for 28-day all-cause mortality (Figure 3) and in-hospital mortality (Figure 4) demonstrated a nonlinear, U-shaped association of SHR with mortality risk. The nadir of the curve, corresponding to the point of minimal risk, was identified at an SHR value of approximately 1.01.

Figure 3 RCS analysis of the 28-day all-cause mortality rate. The vertical dashed line marks the lowest point of the curve (SHR =1.01), representing the baseline risk ratio. The horizontal dashed line indicates the risk ratio of 1.0. Using the lowest risk point (SHR =1.01) as reference, the adjusted HR values at SHR =0.8, 1.0, 1.2, and 1.4 are 1.15,1.01,1.38, and 2.01 respectively. CI, confidence interval; HR, hazard ratio; RCS, restricted cubic spline; SHR, stress hyperglycemia ratio.

Figure 4 Results of the risk ratio (RCS) analysis for in-hospital mortality. CI, confidence interval; RCS, restricted cubic spline; SHR, stress hyperglycemia ratio.

Boruta algorithm

Figure 5 shows that as an exploratory analysis, the Boruta algorithm identifies SHR as an important variable. Its Z-value is high, while variables in other regions are considered unimportant features by the Boruta algorithm.

Figure 5 Feature selection based on the Boruta algorithm. The horizontal axis represents the names of each variable, while the vertical axis represents the Z-scores of each variable. The boxplot displays the Z-scores of each variable during the model calculation process. The red boxes represent important variables, while the yellow, green, and purple boxes represent unimportant variables.

Sensitivity analysis

To further explore the effect modification of SHR on mortality, we employed likelihood ratio tests to evaluate the interaction between diabetes status and continuous SHR (including its nonlinear terms). The results showed no significant interaction effect (P=0.61), indicating that the U-shaped association between SHR and mortality remains relatively consistent across different subgroups of diabetic populations (Figure 6).

Figure 6 Subgroup forest plot of 28-day all-cause mortality. Quartiles 1 (0.2–<0.89), 2 (0.89–<1.08), 3 (1.08–<1.36), and 4 (1.36– 6.1). AKI, acute kidney injury; CI, confidence interval; F, female; HR, hazard ratio; HTN, hypertension; M, male; T2DM, type 2 diabetes mellitus.

Discussion

This study has for the first time revealed a U-shaped association between SHR and 28-day all-cause mortality and hospitalization-related mortality in critically ill AF patients. The RCS model characterized the relationship between SHR and 28-day all-cause mortality as U-shaped, a finding that was statistically significant (P<0.001 for nonlinearity). Multivariate Cox regression analysis indicated that when SHR was approximately 1.01, patients had the lowest mortality risk, while SHR ≥1.36 (Q4 group) showed a significant increase in mortality (adjusted HR =1.89, 95% CI: 1.31–2.74). To further validate the robustness of these results, we employed Lasso-Cox regression for feature selection and modeling, where SHR remained identified as an important prognostic factor (see Figure S2). As an exploratory analysis, the Boruta algorithm identifies SHR as an important variable: by comparing Z-values between actual features and randomly selected “shadow features”, SHR was confirmed as an independent significant predictor. Its predictive efficacy remained unaffected by traditional confounding factors such as age and body weight, confirming the biological nature of the SHR-mortality association. Although subgroup analyses revealed stronger associations between SHR and mortality in certain populations (e.g., males and elderly patients), formal interaction tests did not identify statistically significant effect modifications. This may be attributed to sample size limitations or genuine absence of differential effects. Therefore, these subgroup analysis results should be considered exploratory findings requiring further validation in prospective studies. The study also excluded patients with potential factors affecting HbA1c accuracy (end-stage renal disease, blood transfusion within one week prior to hospitalization, or severe anemia), and conducted sensitivity analysis. The main conclusions remained unchanged, as detailed in supplementary file (Appendix 1). Notably, low systolic blood pressure (SHR) (Q1 group) was associated with increased mortality, with potential mechanisms including: (I) metabolic reserve depletion: The significantly lower albumin levels (3.36±0.64 vs. Q4 group 3.21±0.56 g/dL, P<0.001) and reduced red blood cell count in Q1 group suggested malnutrition; (II) myocardial energy crisis: animal studies indicate chronic hypoglycemia induces mitochondrial dysfunction and increases arrhythmia risk; (III) immune compensation deficiency: low SHR may reflect insufficient insulin secretion, impairing anti-inflammatory capacity. Future research should integrate nutritional indicators (prealbumin) and immune biomarkers (CD4⁺/CD8⁺ ratio) to deepen mechanistic exploration. Additionally, advanced imaging techniques such as speckled tracking echocardiography (STE) (19) could be employed to investigate specific correlations between SHR and atrial/ventricular myocardial strain parameters. This would more intuitively reveal acute metabolic stress’s direct impact on cardiac function, providing richer imaging evidence for risk stratification.

Connection with previous studies

Prior SHR research has primarily focused on coronary artery diseases (20), such as acute myocardial infarction (21), acute heart failure (22), and chronic heart failure (23), with mechanisms mainly involving hyperglycemia-induced insulin resistance, oxidative stress, inflammation, endothelial injury, and thrombosis risk. In cardiovascular contexts, elevated SHR has been confirmed to correlate with adverse outcomes in myocardial infarction patients (24), but reports on critically ill AF remain limited. This study innovatively introduces SHR into the assessment framework for critically ill AF and proposes the “metabolic-inflammatory-thrombotic” axis theory, suggesting that atrial electrical and structural remodeling in critically ill AF patients may be exacerbated by hyperglycemic stress, thereby promoting heart failure and thromboembolic events (25).

Potential mechanisms

AF is most commonly caused by various cardiovascular diseases, including valvular heart disease (e.g., rheumatic valvular disease), coronary artery disease, cardiomyopathy (e.g., dilated or hypertrophic cardiomyopathy), and hypertension. These conditions lead to structural and functional changes in the atria, triggering AF. SHR serves as a marker of acute hyperglycemic stress, independent of long-term glycemic history. Our results confirm its significant association with mortality in critically ill AF patients, regardless of their type 2 diabetes status. Key mechanisms mediating the U-shaped SHR-AF mortality association include: (I) endothelial dysfunction: stress hyperglycemia correlates with counterregulatory hormones and insulin resistance (26). Insulin resistance disrupts endothelial cell function, reducing NO release and impairing coronary vasodilation. Hyperglycemia promotes endothelin-1 synthesis, exacerbating inflammation, vascular smooth muscle contraction, and local myocardial ischemia, further inducing AF (27). (II) Oxidative stress and inflammation: NADPH oxidase activation is a major source of cardiac reactive oxygen species (ROS). Stress hyperglycemia increases NADPH oxidase activity and mitochondrial ROS overproduction, impairing myocardial metabolism and contractility (28). (III) Coagulation activation and impaired fibrinolysis: Stress hyperglycemia elevates factor VII and tissue factor, activating coagulation while increasing plasminogen activator inhibitor-1, suppressing fibrinolysis (29). This prothrombotic state exacerbates myocardial ischemia and AF (30). (IV) Increased free fatty acids: Insulin resistance and excessive catecholamines accelerate lipolysis and circulating free fatty acid release, reducing myocardial contractility and increasing AF risk (11).

Clinical implications

A prior study on SHR in acute myocardial infarction (AMI) patients with/without AF found a U-shaped SHR-mortality association only in AMI-AF patients, suggesting AF differentially increases mortality risk in low-SHR AMI cases (31). Our findings indicate that higher SHR correlates with increased 28-day all-cause and in-hospital mortality in critically ill AF, independent of comorbidities, supporting SHR as a broad risk predictor.

Limitations

There are several limitations in this study: First, while the retrospective single-center design provides preliminary evidence of association, it cannot establish definitive causation. The structural constraints of the MIMIC-IV database prevent inclusion of socioeconomic indicators such as socioeconomic status and GRACE risk scores. Second, residual confounding remains despite multivariate adjustments and subgroup analyses, particularly regarding non-targeted INR values influenced by novel anticoagulants, drug interactions, or dietary factors. Third, sample representativeness faces dual challenges: a moderate sample size may reduce statistical power, while excluding cases with missing blood glucose or HbA1c data could introduce selection bias. Fourth, uncertainties exist in key measurements, including whether ABG was measured fasting and uncorrected HbA1c levels due to abnormal red blood cell lifespan. Fifth, insufficient exploration of core pathogenic mechanisms: critical parameters like echocardiography (38% missing left ventricular mass index), ambulatory electrocardiography (41% missing AF load data), and inflammatory markers (52% missing C-reactive protein and interleukin-6) are lacking, and existing data fail to elucidate the biological pathways through which SHR index affects mortality. Sixth, significant limitations in disease classification: AF diagnosis relies on administrative codes without distinguishing between paroxysmal, persistent, or permanent AF, nor differentiating between new-onset and pre-existing AF. Seventh, this study excluded analyses of anticoagulant therapy [e.g., direct oral anticoagulants (DOACs) versus warfarin] or glycemic control strategies (e.g., intensive insulin therapy), which may influence outcomes by modulating thrombotic risk or metabolic status. For instance, DOACs might reduce thrombotic risks associated with high systolic blood pressure (SHR), while rigorous glucose management could improve endothelial function. Future randomized controlled trials are needed to determine whether SHR-guided individualized therapy improves prognosis. Additionally, the study failed to adjust for key therapeutic interventions that may affect glycemic control and outcomes, such as insulin therapy, glucocorticoids, vasoactive agents, nutritional support, and antiarrhythmic medications. These unmeasured therapeutic confounders could influence observed associations. We calculated an effect size (ES) of 3.2, indicating that an unmeasured confounder with a risk ratio ≥3.2 would be required to fully explain the observed association between high SHR and mortality risk in this study.

Conclusions

The risk ratio (SHR) demonstrated independent association with 28-day all-cause mortality in critically ill AF patients, with its U-shaped correlation providing a novel tool for metabolic risk stratification. In clinical practice, the SHR can serve as a simple bedside indicator for prognosis assessment. Future prospective studies are needed to validate its potential in guiding personalized treatment strategies.

Acknowledgments

The research team would like to express gratitude to Beth Israel Deaconess Medical Center for providing support through the MIMIC-IV database, the National Natural Science Foundation of China (Project No. 82174312) and Clinical Medicine Transformation Project, as well as the Anhui Provincial Department of Science and Technology (Project No. 202304295107020112). Special thanks go to all technical staff involved in data compilation.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1507/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1507/prf

Funding: This study was supported by the National Natural Science Foundation of China General Program (No. 82174312), Clinical Medicine Transformation Project, Anhui Provincial Department of Science and Technology (No. 202304295107020112).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1507/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Sagris M, Vardas EP, Theofilis P, et al. Atrial Fibrillation: Pathogenesis, Predisposing Factors, and Genetics. Int J Mol Sci 2021;23:6. [Crossref] [PubMed]
2. Sibley S, Bedford J, Wetterslev M, et al. Atrial fibrillation in critical illness: state of the art. Intensive Care Med 2025;51:904-16. [Crossref] [PubMed]
3. Yue H, Zhao X, Liang W, et al. Curcumin, novel application in reversing myocardial fibrosis in the treatment for atrial fibrillation from the perspective of transcriptomics in rat model. Biomed Pharmacother 2022;146:112522. [Crossref] [PubMed]
4. Lu Y, Sun Y, Cai L, et al. Non-traditional risk factors for atrial fibrillation: epidemiology, mechanisms, and strategies. Eur Heart J 2025;46:784-804. [Crossref] [PubMed]
5. Heidenreich PA, Bozkurt B, Aguilar D, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: Executive Summary: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2022;145:e876-94. [Crossref] [PubMed]
6. Johansen OE, Brustad E, Enger S, et al. Prevalence of abnormal glucose metabolism in atrial fibrillation: a case control study in 75-year old subjects. Cardiovasc Diabetol 2008;7:28. [Crossref] [PubMed]
7. Nakanishi K, Daimon M, Fujiu K, et al. Prevalence of glucose metabolism disorders and its association with left atrial remodelling before and after catheter ablation in patients with atrial fibrillation. Europace 2023;25:euad119. [Crossref] [PubMed]
8. Jie Q, Li G, Qian W, et al. Distinct patterns of association between the hemoglobin glycation index, the stress-hyperglycemia ratio, and the risk of new-onset atrial fibrillation in critically ill patients. Front Endocrinol (Lausanne) 2025;16:1656783. [Crossref] [PubMed]
9. Akirov A, Grossman A, Shochat T, et al. Hyperglycemia on admission and hospitalization outcomes in patients with atrial fibrillation. Clin Cardiol 2017;40:1123-8. [Crossref] [PubMed]
10. Roberts GW, Quinn SJ, Valentine N, et al. Relative Hyperglycemia, a Marker of Critical Illness: Introducing the Stress Hyperglycemia Ratio. J Clin Endocrinol Metab 2015;100:4490-7. [Crossref] [PubMed]
11. Huang S, Gao F, Huang WB, et al. Association between stress hyperglycemia ratio and mortality in acute myocardial infarction patients with and without atrial fibrillation: a retrospective cohort study from the MIMIC-IV database. BMC Cardiovasc Disord 2024;24:675. [Crossref] [PubMed]
12. Li J, Wang L, Zhang J, et al. Association between stress hyperglycemia ratio (SHR) and the risk of acute kidney injury in patients with chronic kidney disease: analysis of the MIMIC-IV database. BMC Nephrol 2025;26:310. [Crossref] [PubMed]
13. Long Z, Liao S, Chen Y. Nonlinear association between stress hyperglycemia ratio and severe consciousness disorder in acute ischemic stroke: A MIMIC retrospective analysis. PLoS One 2025;20:e0329678. [Crossref] [PubMed]
14. Mohammed AQ, Luo Y, Wang K, et al. Stress hyperglycemia ratio as a prognostic indicator for long-term adverse outcomes in heart failure with preserved ejection fraction. Cardiovasc Diabetol 2024;23:67. [Crossref] [PubMed]
15. You X, Zhang H, Li T, et al. Stress hyperglycemia ratio and 30-day mortality among critically ill patients with acute heart failure: analysis of the MIMIC-IV database. Acta Diabetol 2025;62:1537-47. [Crossref] [PubMed]
16. Zhang Y, Yin X, Liu T, et al. Association between the stress hyperglycemia ratio and mortality in patients with acute ischemic stroke. Sci Rep 2024;14:20962. [Crossref] [PubMed]
17. Cheng Y, Yue L, Zhang J, et al. Stress hyperglycemia is associated with disease severity in COVID-19. Endokrynol Pol 2023;74:528-35. [Crossref] [PubMed]
18. Pan H, Xiong Y, Huang Y, et al. Association between stress hyperglycemia ratio with short-term and long-term mortality in critically ill patients with ischemic stroke. Acta Diabetol 2024;61:859-68. [Crossref] [PubMed]
19. Sonaglioni A, Braga M, Villa MC, et al. Comprehensive assessment of biventricular myocardial function by two-dimensional speckle tracking echocardiography in infants of gestational diabetic mothers. Acta Diabetol 2022;59:1145-56. [Crossref] [PubMed]
20. Xue X, Liu Z, Xue T, et al. Machine learning for the prediction of acute kidney injury in patients after cardiac surgery. Front Surg 2022;9:946610. [Crossref] [PubMed]
21. Zhang Y, Song H, Bai J, et al. Association between the stress hyperglycemia ratio and severity of coronary artery disease under different glucose metabolic states. Cardiovasc Diabetol 2023;22:29. [Crossref] [PubMed]
22. Xu W, Yang YM, Zhu J, et al. Predictive value of the stress hyperglycemia ratio in patients with acute ST-segment elevation myocardial infarction: insights from a multi-center observational study. Cardiovasc Diabetol 2022;21:48. [Crossref] [PubMed]
23. Bo K, Li W, Zhang H, et al. Association of stress hyperglycemia ratio with left ventricular function and microvascular obstruction in patients with ST-segment elevation myocardial infarction: a 3.0 T cardiac magnetic resonance study. Cardiovasc Diabetol 2024;23:179. [Crossref] [PubMed]
24. Okita S, Saito Y, Yaginuma H, et al. Impact of the Stress Hyperglycemia Ratio on Heart Failure and Atherosclerotic Cardiovascular Events After Acute Myocardial Infarction. Circ J 2025;89:340-6. [Crossref] [PubMed]
25. Hu B, Chen X, Wang Y, et al. J-shaped relationship between stress hyperglycemia ratio and 90-day and 180-day mortality in patients with a first diagnosis of acute myocardial infarction: analysis of the MIMIC-IV database. Diabetol Metab Syndr 2024;16:132. [Crossref] [PubMed]
26. Bloomgarden ZT, Kosiborod MN, Handelsman Y. Concomitant Diabetes and Atrial Fibrillation: No Sugarcoating the Bittersweet Reality. J Am Coll Cardiol 2017;70:1336-8. [Crossref] [PubMed]
27. Bellis A, Mauro C, Barbato E, et al. Stress-Induced Hyperglycaemia in Non-Diabetic Patients with Acute Coronary Syndrome: From Molecular Mechanisms to New Therapeutic Perspectives. Int J Mol Sci 2021;22:775. [Crossref] [PubMed]
28. Qin S, Boidin M, Buckley BJR, et al. Endothelial dysfunction and vascular maladaptation in atrial fibrillation. Eur J Clin Invest 2021;51:e13477. [Crossref] [PubMed]
29. Zhou L, Liu Y, Wang Z, et al. Activation of NADPH oxidase mediates mitochondrial oxidative stress and atrial remodeling in diabetic rabbits. Life Sci 2021;272:119240. [Crossref] [PubMed]
30. Mulder BA, Geelhoed B, van der Harst P, et al. Plasminogen activator inhibitor-1 and tissue plasminogen activator and incident AF: Data from the PREVEND study. Int J Cardiol 2018;272:208-10. [Crossref] [PubMed]
31. Wang XH, Ning ZH, Xie Z, et al. SIRT3/AMPK Signaling Pathway Regulates Lipid Metabolism and Improves Vulnerability to Atrial Fibrillation in Dahl Salt-Sensitive Rats. Am J Hypertens 2024;37:901-8. [Crossref] [PubMed]