The prognostic impact of stress hyperglycemia ratio on mortality in cardiogenic shock: a MIMIC-IV database analysis

Abstract

Background

The stress hyperglycemia ratio (SHR) has been established as a predictor of unfavorable outcomes across various diseases. However, its relationship with prognosis in patients with cardiogenic shock (CS) remains unclear. This study aims to investigate the association between SHR and outcomes in CS patients.

Methods

A total of 904 CS patients with their first ICU admission were included in this study, utilizing data from the American Medical Information Mart for Intensive Care (MIMIC-IV) database. The primary endpoints were all-cause mortality at 30 days and 360 days. Patients were stratified into three groups based on the tertiles of the SHR.

Results

The mean age of the cohort was 67.62 years, with 67.3% of participants being men. During the follow-up period, 221 patients (24.4%) died within 30 days, and 360 patients (39.8%) died within 360 days. The 30-day all-cause mortality rates were 16.9%, 22.3%, and 34.2% in the T1, T2, and T3 groups, respectively (p < 0.001), while the 360-day all-cause mortality rates were 34.9%, 39.0%, and 45.6%, respectively (p = 0.015). Compared with patients in T1, those in T3 exhibited a significantly higher risk of 30-day all-cause mortality (HR = 2.140, 95% CI: 1.522–3.008, p < 0.001) and 360-day all-cause mortality (HR = 1.495, 95% CI: 1.157–1.931, p = 0.002). Restricted cubic spline (RCS) analyses demonstrated an approximately linear relationship between SHR and 360-day all-cause mortality (p for overall = 0.011; p for nonlinearity = 0.099). However, a nonlinear association was observed between SHR and 30-day all-cause mortality (p for overall < 0.001; p for nonlinearity = 0.030), with the risk increasing significantly when SHR exceeded 1.176. Subgroup analyses revealed that the effect of SHR was consistent across most subgroups except in patients with and without acute myocardial infarction (AMI). In patients with AMI, SHR was associated with a significantly elevated risk of mortality, whereas no significant association was observed in patients without AMI. For 30-day all-cause mortality, the HR was 1.059 (95% CI: 1.040–1.078) in patients with AMI and 1.002 (95% CI: 0.966–1.040) in those without AMI (p for interaction = 0.007). For 360-day all-cause mortality, the HR was 1.043 (95% CI: 1.026–1.061) in patients with AMI and 0.984 (95% CI: 0.955–1.014) in those without AMI (p for interaction < 0.001).

Conclusion

Elevated SHR was significantly associated with increased 30-day and 360-day all-cause mortality in patients with CS, particularly in those with CS complicated by AMI. SHR may serve as a valuable marker for risk stratification and guiding subsequent interventions in CS patients. However, further prospective studies are needed to confirm these findings.

Introduction

Cardiogenic shock (CS) is a severe condition characterized by low cardiac output, resulting in inadequate perfusion and hypoxia of end organs. This often progresses to multiple organ failure and is associated with high mortality rates [1]. Despite advancements in mechanical circulatory support devices that have significantly reduced mortality, the in-hospital mortality rate remains alarmingly high, ranging from 27–51% [2]. Therefore, identifying critically ill patients at high risk of death is of paramount importance.

Critical illness often disrupts glucose metabolism, leading to insulin resistance and glucose intolerance, which manifest as elevated blood glucose levels. This condition, known as stress hyperglycemia, is an adaptive evolutionary response that can enhance survival in life-threatening situations [3]. However, stress hyperglycemia is also associated with an increased risk of adverse outcomes. Previous studies have linked stress hyperglycemia to higher mortality rates, prolonged hospital stays, and a greater incidence of complications [3]. Blood glucose concentrations, however, are influenced by various factors beyond acute stress, limiting their ability to accurately reflect acute glucose spikes. In contrast, the stress hyperglycemia ratio (SHR)—calculated using blood glucose and hemoglobin A1c (HbA1c), provides a more reliable measure of glucose metabolic changes in critically ill patients [4]. Studies have identified SHR as an independent risk factor for mortality in conditions such as acute myocardial infarction [5], heart failure [6], diabetes [7], acute ischemic stroke [8].

As the most severe form of cardiovascular disease, CS poses a particularly high risk of poor prognosis due to mechanisms associated with stress hyperglycemia, including glucose metabolism disorders [3], circulatory disorders [6], hypoperfusion [6], and inflammatory responses [9]. Despite this, limited research has explored the relationship between SHR and outcomes in CS patients. The present study aimed to evaluate the impact of SHR on 30-day and 360-day all-cause mortality in CS patients using data from the American Medical Information Mart for Intensive Care (MIMIC)-IV database.

Methods

Study design and population

This was a retrospective observational cohort study with longitudinal follow-up, based on the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 3.0) database [10]. This publicly accessible clinical database includes 94,458 intensive care unit (ICU) stays recorded between 2008 and 2022 at the Beth Israel Deaconess Medical Center. The study variables were extracted by the author L.F.X., who completed the Collaborative Institutional Training Initiative (CITI) examination (Certification Number: 57983166). Individual patient consent was not required, as the database contains anonymized patient health information. The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University and was conducted in compliance with the Declaration of Helsinki.

Study population

The study population included adult patients diagnosed with CS as identified by ICD-9 code 785.51 and ICD-10 code R57.0. Patients under 16 years of age or those with incomplete information on ICU outcomes were excluded. For individuals with multiple ICU admissions due to CS, only data from the first admission were included in the analysis. (Fig. 1).

Fig. 1

Flowchart of study participants

Data extraction

Clinical and demographic data were extracted using PostgreSQL software (version 13.7.2) and Navicat Premium software (version 16) by executing Structured Query Language (SQL) queries. Demographic data included gender, age, and race. Clinical data encompassed vital signs, weight, height, comorbidities, clinical outcomes, and laboratory test results. For laboratory tests, the first results obtained after ICU admission were used. The stress hyperglycemia ratio (SHR) was calculated using the formula: SHR = (admission glucose) (mg/dL) / (28.7 * HbA1c [%]-46.7) [4].

Outcomes

The endpoints were 30-day all-cause mortality and 360-day all-cause mortality.

Statistical analysis

Continuous variables that followed a normal distribution are presented as means with standard deviations (SD), and group comparisons were made using one-way analysis of variance (ANOVA). For variables with non-normally distributed data, the median and interquartile range (25-75%) are reported, and comparisons between groups were performed using the Mann-Whitney U test. Categorical data are expressed as counts and percentages, with comparisons made using the chi-square test.

Patients were divided into three groups based on the tertiles of SHR levels. Cumulative hazards for all-cause mortality were estimated using the Kaplan-Meier method. Univariable and multivariable Cox regression models were applied to assess the relationship between SHR levels and the study endpoints. Based on prior literature and clinical relevance [11], the following covariates were selected for inclusion in the multivariable Cox models: Model 1 (unadjusted), Model 2 (adjusted for age and gender), and Model 3 (adjusted for age, gender, systolic blood pressure, creatinine, SOFA score, chronic pulmonary disease, and acute myocardial infarction [AMI]).

Furthermore, restricted cubic spline (RCS) analysis was used to explore the relationship between SHR and outcomes, with four knots positioned at the 5th, 35th, 65th, and 95th percentiles, as recommended by Harrell. Subgroup analyses were conducted to assess the impact of SHR (per 0.1) in various patient subgroups patients based on age (> 65 and ≤ 65 years old), gender (male and female), race (white and non-white), diabetes status (diabetes and non-diabetes) and AMI status (AMI and non-AMI).

A two-tailed p-value of < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 25.0 (IBM, USA) and R version 4.1.2 (R Foundation).

Results

Baseline characteristics

A total of 904 patients with CS were extracted from the American MIMIC-IV cohort. The mean age of this cohort was 67.62 years (SD 13.51), with 67.3% being male. During the follow-up period, 221 (24.4%) patients died within 30 days, and 360 (39.8%) patients died within 360 days. Patients were divided into three groups based on their stress hyperglycemia ratio (SHR) levels: T1 (SHR < 1.005, n = 301), T2 (1.005 ≤ SHR < 1.395, n = 305), and T3 (SHR ≥ 1.395, n = 298).

The baseline characteristics of the three groups are presented in Table 1. The groups were similar in terms of demographic features, with the exception that patients in the T2 and T3 groups had a higher proportion of white patients (65.6% and 56.7%, respectively, p = 0.026). Although the admission systolic blood pressure was similar across the groups, patients in T2 and T3 tended to have higher diastolic blood pressure (p = 0.001), heart rate (p = 0.001), and respiratory rate (p = 0.004). Additionally, the T3 group had a higher proportion of myocardial infarction (MI) (56.7%), while the T1 group had a higher prevalence of liver disease (15.9%) and diabetes (48.2%).

Table 1 Characteristics and outcomes of participants categorized by SHR

Regarding laboratory results, patients in the T3 group had higher levels of lactate (p < 0.001), white blood cells (p < 0.001), platelets (p < 0.001), hemoglobin (p = 0.001), neutrophils (p < 0.001), potassium (p = 0.005), alanine aminotransferase (p = 0.003), aspartate aminotransferase (p < 0.001), and glucose (p < 0.001). In contrast, they had lower levels of sodium (p = 0.004), international normalized ratio (INR) (p < 0.001), prothrombin time (PT) (p < 0.001), and HbA1c (p < 0.001). Furthermore, the T3 group had higher SAPSII (p < 0.001), LODS (p = 0.002), OASIS (p < 0.001), and SIRS (p < 0.001) scores.

Clinical outcomes

Figure 2 shows the 30-day and 360-day all-cause mortality rates. The 30-day all-cause mortality rates for T1, T2, and T3 were 16.9%, 22.3%, and 34.2%, respectively (p < 0.001), while the 360-day all-cause mortality rates were 34.9%, 39.0%, and 45.6%, respectively (p = 0.015).

Fig. 2

The 30-day and 360-day all-cause mortality in different groups

Figure 3 presents the Kaplan-Meier survival curves for the three groups. For 30-day all-cause mortality, the cumulative incidence in the T1 and T2 groups was similar, while the T3 group had a significantly higher cumulative incidence of both 30-day and 360-day all-cause mortality compared to the T1 and T2 groups (log-rank p < 0.05).

Fig. 3

Kaplan-Meier survival analysis curves for all-cause mortality

Figure 4 displays the Cox regression analysis. In univariate Cox regression, patients in the T3 group had a significantly higher risk of 30-day all-cause mortality compared to those in the T1 group (HR = 2.302, 95% CI: 1.644, 3.222, p < 0.001) (Model 1). In multivariate Cox regression analyses (Model 2 and Model 3), similar trends were observed, with significantly increased risk of 30-day all-cause mortality in the T3 group (HR = 2.294, 95% CI: 1.639, 3.211, p < 0.001 for Model 2; HR = 2.140, 95% CI: 1.522, 3.008, p < 0.001 for Model 3). For 360-day all-cause mortality, patients in the T3 group also had a significantly higher risk compared to the T1 group in all models: Model 1 (HR = 1.532, 95% CI: 1.187, 1.976, p = 0.001), Model 2 (HR = 1.554, 95% CI: 1.204, 2.005, p < 0.001), and Model 3 (HR = 1.495, 95% CI: 1.157, 1.931, p = 0.002).

Fig. 4

The association between SHR groups and mortality. Model 1: Unadjusted; Model 2: Adjusted for age, gender; Model 3: Adjusted for age and gender, and model 3 adjusted for age, gender, systolic blood pressure, creatinine, sofa score, chronic pulmonary disease, and acute myocardial infarction (AMI)

Figure 5 shows the restricted cubic spline (RCS) analyses, which reveal a J-shaped association between SHR and 30-day all-cause mortality (p for overall < 0.001, p for nonlinear = 0.030), with a significant increase in risk when SHR > 1.176. In contrast, an approximate linear relationship was observed between SHR and 360-day all-cause mortality (p for overall = 0.011, p for nonlinear = 0.099).

Fig. 5

Potential nonlinear for the levels of SHR with 30-day and 360-day all-cause mortality measured by restricted cubic spline regression with 4 knots located at the 5th, 35th, 65th and 95th percentiles

Subgroup analyses

Subgroup analyses were conducted to evaluate the association between SHR and outcomes in various patient subgroups (Fig. 6). SHR had a consistent effect across subgroups based on age (> 65 vs. ≤65 years), gender, race, and the presence of diabetes (all p-interaction > 0.05). However, the effect of SHR differed between patients with and without acute myocardial infarction (AMI). In patients with AMI, SHR was associated with a higher risk of both 30-day and 360-day all-cause mortality (HR = 1.059, 95% CI: 1.040, 1.078 for 30-day mortality; HR = 1.043, 95% CI: 1.026, 1.061 for 360-day mortality), whereas in patients without AMI, SHR was not significantly associated with these outcomes (HR = 1.002, 95% CI: 0.966, 1.040 for 30-day mortality; HR = 0.984, 95% CI: 0.955, 1.014 for 360-day mortality). The p-interaction values were 0.007 for 30-day mortality and < 0.001 for 360-day mortality.

Fig. 6

Subgroup analysis for the association of SHR with all-cause mortality

Discussion

To the best of our knowledge, this study is the first to investigate the association between the SHR and outcomes in patients with CS. Our findings indicate that an elevated SHR (≥ 1.395) is independently associated with higher 30-day and 360-day all-cause mortality. This study underscores the prognostic significance of SHR in patients with CS.

Disruption of glucose metabolism is common in critically ill patients and has been shown to be associated with clinical outcomes [12,13,14]. Stress hyperglycemia typically refers to transient hyperglycemia that occurs during periods of stress. A combination of factors acts collectively and synergistically to induce this condition. Stress hyperglycemia is primarily mediated by the hypothalamic-pituitary-adrenal (HPA) axis and the sympatho-adrenal system, which release cortisol and catecholamines. These hormones further elevate blood glucose levels by promoting gluconeogenesis and glycogenolysis while inhibiting glucose uptake in peripheral tissues [3, 9]. Additionally, inflammatory mediators and altered adipokine release from adipose tissue contribute to peripheral insulin resistance and exacerbate stress hyperglycemia [9, 15]. While stress hyperglycemia has protective effects, helping maintain metabolic homeostasis and survival during stress [3], increasing evidence suggests that prolonged or inappropriate hyperglycemia is strongly associated with poor outcomes in various diseases, particularly in cardiovascular conditions. In patients with MI, elevated SHR has been significantly linked to a higher risk of all-cause mortality [16]. A meta-analysis further confirmed the prognostic value of SHR in MI patients[ 5 ]. In patients with acute decompensated heart failure, those in the highest quintile of SHR (compared to those in the second quintile) were found to have a significantly higher risk of all-cause death (HR = 2.76, 95% CI 1.63–4.68), cardiovascular death (HR = 2.81, 95% CI 1.66–4.75), and heart failure rehospitalization (HR = 1.54, 95% CI 1.03–2.32) [ 6 ]. These findings highlight the prognostic significance of SHR in cardiovascular disease.

CS represents the most critical state of cardiovascular disease, and stress hyperglycemia is commonly observed in these patients, often associated with poor prognosis. Thoegersen et al. [17] found that CS patients with elevated glucose levels upon admission had an increased 30-day mortality. Similarly, in the IABP-SHOCK II trial, higher admission glucose concentrations were independently associated with 30-day and 1-year mortality [18], and this association was independent of the patient’s diabetic status What’s more, this association was independent of diabetic state [17, 18]. These studies highlight that glucose metabolism disorders are a significant pathophysiological factor in CS and are linked to worse outcomes. However, blood glucose levels are influenced by various factors, particularly chronic blood glucose conditions. In contrast, the SHR, which adjusts for average glycemic status, more accurately reflects the metabolic changes in critically ill patients under stress. SHR is a better biomarker of critical illness than absolute hyperglycemia [4]. However, the relationship between SHR and prognosis among CS patients are not well understood, and our study extended previous findings, suggesting SHR was an important predictor of poor prognosis in patients with CS.

The underlying mechanisms linking the SHR to outcomes in CS patients remain unclear, but several potential mechanisms may help explain this critical pathophysiological process. First, elevated SHR has been shown to be associated with circulatory disturbances and hypoperfusion. A sharp increase in blood glucose levels can trigger excessive production of reactive oxygen species by endothelial cells and the myocardium, leading to endothelial dysfunction, impaired vasodilation, and circulatory disorders [6]. Additionally, high blood glucose levels can activate platelets, inhibit fibrinolysis, and increase circulating adhesion molecules, which promote capillary leukocyte plugging and activate coagulation, further exacerbating circulatory disturbances and hypoperfusion [6, 19]. Second, inflammatory mediators, such as tumor necrosis factor-α, interleukin-1, interleukin-6, and C-reactive protein, contribute to stress hyperglycemia by inducing peripheral insulin resistance [9]. Therefore, stress hyperglycemia reflects the inflammatory response in CS patients, which, in turn, exacerbates hemodynamic disturbances, organ hypoperfusion, and cardiac dysfunction, ultimately leading to higher mortality [20]. Third, a high SHR is associated with the overactivation of the HPA axis and the sympathoadrenal system [3], which will lead to fluid retention, increased preload, and worsening pump failure [21]. Fourth, MI with left ventricular dysfunction remains the most frequent cause of CS [22]. Previous studies have shown that blood glucose levels are closely related to the degree of myocardial injury, as reflected by increased cardiac markers and reduced left ventricular function [23]. Moreover, several other mechanisms induced by hyperglycemia, such as impaired wound healing, increased infection risk, mitochondrial dysfunction, insulin resistance, electrolyte and fluid shifts, acid/base disturbances, immune dysregulation, lipotoxicity, catabolism of muscle and adipose tissue, and extracellular matrix deposition, collectively contribute to poor clinical outcomes [9]. Thus, the mechanisms by which SHR correlates with outcomes in CS patients are complex and multifactorial. These mechanisms often interact with one another, compounding their effects and worsening the prognosis.

The clinical implication of the present study is that the SHR should be carefully considered in patients with CS, particularly in those with MI complicated by CS, as demonstrated in our subgroup analysis. This analysis revealed that SHR was independently associated with 30-day and 360-day all-cause mortality in patients with AMI, but not in those without AMI. AMI as the underlying cause of CS not only leads to reduced perfusion of peripheral organs but also results in severe cardiac complications, such as heart failure, arrhythmias, and mechanical complications, which significantly increase the risk of mortality. Additionally, while it is not surprising that diabetic patients with elevated SHR have poor outcomes, previous studies, including our subgroup analysis, also showed that SHR affects the prognosis in non-diabetic patients. This finding suggests that SHR is a significant risk factor for prognosis in CS patients, regardless of their diabetic status.

The innovation of this study lies in its identification of the association between the SHR and prognosis in patients with CS. This is the first report to highlight the prognostic value of SHR in such patients, thereby extending previous research findings. Furthermore, since SHR more accurately reflects a patient’s glycemic metabolism, and prior studies have shown that its predictive value surpasses that of admission glucose levels [24], SHR may serve as a valuable prognostic marker for patients with CS.

In terms of managing stress hyperglycemia, intensive glycemic control in both ICU and non-ICU patients has not demonstrated clear clinical benefits [25,26,27]. This may suggest that SHR primarily reflects the severity of the disease rather than serving as a target for treatment. However, excessively elevated blood glucose levels still require management, as the RCS analyses showed that the risk of 30-day all-cause mortality increased significantly when SHR > 1.176. Previous studies have indicated that mild to moderate stress hyperglycemia may offer protective effects during stress and critical illness, while excessively high blood glucose can be harmful through various mechanisms, as previously mentioned [3, 9]. Therefore, SHR not only reflects disease severity but also provides valuable insights for optimizing treatment strategies in patients with CS.

There are several limitations to our study. First, it is a retrospective observational study with a relatively limited sample size. Despite using rigorous statistical methods, potential biases and uncontrolled factors may have influenced the outcomes. Therefore, large-scale prospective studies are needed for further clarification. Second, we focused solely on the impact of SHR on all-cause mortality, without considering other endpoints such as major cardiovascular adverse events, length of hospitalization, or hospitalization expenses, as these measures were not available in the MIMIC-IV dataset. Future prospective studies should include a broader range of clinical outcomes. Third, our study included CS patients from 2008 to 2022, a period during which management guidelines and interventions for CS were evolving. This may have affected our results, highlighting the need for contemporary clinical studies to validate our findings.

Conclusion

Elevated SHR is significantly associated with both 30-day and 360-day all-cause mortality in patients with CS, irrespective of diabetic status. SHR should be considered as a key factor in clinical management, although the optimal approach for managing SHR in patients with CS warrants further investigation.