Abstract
Background
To evaluate the independent and combined effects of mean blood glucose (MBG) and glycemic variability (GV) on 28-day mortality in older adults with acute myocardial infarction (AMI).
Methods
A retrospective cohort analysis was conducted using the Medical Information Mart for Intensive Care IV (MIMIC-IV) database, focusing on AMI patients aged ≥ 65 years. Participants were stratified into age cohorts: 65–75, 75–85, and ≥ 85 years. MBG and GV were derived from all glucose readings throughout the ICU stay. The primary outcome was 28-day mortality, and hypoglycemia was a secondary outcome. Cox regression with risk-adjusted restricted cubic splines was applied to examine associations.
Results
Among 1,242 older adults (mean age 77.3 ± 7.5 years), the mean MBG was 147.1 ± 43.1 mg/dL and median GV was 33.5 (IQR 20.7–55.0) mg/dL. During follow-up, 295 participants (23.8%) died within 28 days. Multivariable analysis showed that each 10 mg/dL increase in MBG and GV was linked to a 17% (Model 3: HR = 1.17; 95% CI, 1.13–1.20) and 6% (HR = 1.06; 95% CI, 1.03–1.08) higher mortality risk, respectively (both p < 0.001). Further analysis, using the low MBG/low GV group as the reference, revealed distinct mortality risk patterns. The high MBG/low GV subgroup exhibited the strongest association with mortality, showing a 5.77-fold increased risk (HR = 5.77; 95% CI, 3.36–9.92; p < 0.001). These associations remained consistent in both diabetic and non-diabetic subgroups (p < 0.001).
Conclusions
Elevated mean blood glucose and glycemic variability independently predicted 28-day mortality in older adults with acute myocardial infarction. Critically, sustained hyperglycemia demonstrated greater clinical detriment than acute glucose fluctuations, necessitating risk-stratified glycemic protocols prioritizing MBG control in this vulnerable population.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12877-025-06499-z.
Keywords: Mean blood glucose, Glycemic variability, Acute myocardial infarction, Intensive care, Mortality, Older adults
Background
Acute myocardial infarction (AMI) presents a significant clinical challenge, particularly among the older adults, who experience substantially higher mortality rates following such events [1–3]. Effective management of AMI requires strict control of physiological parameters, with blood glucose regulation recognized as a key prognostic factor due to its established association with adverse outcomes [4]. While multiple studies in adults have demonstrated the negative prognostic impact of elevated mean blood glucose (MBG) and increased glycemic variability (GV) following AMI [5–8], research specifically investigating this association in older adults (aged >65 years) with AMI remains limited.
This study aimed to clarify the independent and combined effects of mean blood glucose levels and glycemic variability on 28-day mortality in older AMI patients. We sought to generate clinically relevant insights that could inform targeted interventions and improve patient outcomes, utilizing the Medical Information Mart for Intensive Care IV (MIMIC-IV) database, which contains comprehensive data on critically ill patients in intensive care units (ICUs) [9].
Methods
Data source
This retrospective study utilized MIMIC-IV (version 3.0), which contains clinical data from 94,458 ICU admissions of 65,366 unique patients treated at Beth Israel Deaconess Medical Center (BIDMC) in Boston between 2008 and 2019. This single-center longitudinal database contains medical records of 523,740 inpatients treated at BIDMC during this period. One author (Jun Zhou) obtained access to the MIMIC-IV database (certification no. 65830225) after completing the Collaborative Institutional Training Initiative (CITI) program.
Study population
Patient selection adhered to the following inclusion criteria: (1) AMI diagnosis per International Classification of Diseases, Tenth Revision (ICD-10) codes, and (2) aged ≥ 65 years. Exclusion criteria were as follows: (1) not the first ICU admission, (2) ICU stay ≤ 2 days, and (3) missing survival or glycemic data.
Data extraction and definitions
Demographic characteristics, vital signs, comorbidities, ICU treatments, and laboratory measurements were extracted using Navicat Premium (version 16.2) via SQL queries. The following data were included:
Demographic variables: age, sex, marital status, race, type of ICU admission, ST-segment elevation myocardial infarction (STEMI) or not, and body mass index (BMI). Demographic variables: age, sex, marital status, race, type of ICU admission, ST-segment elevation myocardial infarction (STEMI) or not, and body mass index (BMI).
Comorbidities: congestive heart failure, diabetes, cerebrovascular disease, severe liver disease, hypertension, atrial fibrillation (AF), and cardiogenic shock.
Laboratory indicators: white blood cell count, hemoglobin, platelets, creatinine, HbA1c, cardiac troponin I, N-terminal pro-B-type natriuretic peptide (NT-proBNP), triglycerides, cholesterol, and glucose levels.
Disease severity scores: Charlson Comorbidity Index, Acute Physiology Score III (APS III), and Sequential Organ Failure Assessment (SOFA) score.
Therapies: continuous renal replacement therapy (CRRT), mechanical ventilation (MV), insulin, vasoactive drugs, statins, β-blockers, calcium channel blockers (CCBs), angiotensin-converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), alteplase, antiplatelet agents, anticoagulants, and percutaneous coronary intervention (PCI).
MBG was calculated using all biochemical glucose measurements recorded during the ICU stay. GV was defined as the standard deviation (SD) of these glucose values. Hypoglycemia was defined as at least one blood glucose measurement < 70 mg/dL during ICU admission.
Outcome
The primary outcome was 28-day mortality, and the secondary outcome was the occurrence of hypoglycemia. Mortality data were derived from recorded time of death in the MIMIC-IV database.
Statistical analysis
Continuous variables were presented as mean ± SD or median (interquartile range, IQR), depending on their distribution. Group comparisons were conducted using one-way ANOVA for normally distributed data and the Kruskal–Wallis test for non-normally distributed data. Categorical variables were expressed as frequencies (%) and compared using the chi-square test. Missing data were addressed with imputed datasets, except for laboratory variables such as triglycerides, cholesterol, and NT-proBNP, which were excluded when missingness exceeded 40%.
Participants were stratified into age cohorts: 65–75, 75–85, and ≥ 85 years. MBG was classified into low (< 140 mg/dL), moderate (140–180 mg/dL), or high (≥ 180 mg/dL) based on the American Diabetes Association guidelines for glucose management in critically ill patients. GV was stratified into tertiles: low (T1, < 23.82 mg/dL), moderate (T2, 23.82–45.77 mg/dL), and high (T3, ≥ 45.77 mg/dL), based on the study population distribution. For interaction analyses, MBG was dichotomized into high (≥ 180 mg/dL) and low (< 180 mg/dL), while GV was dichotomized into high (≥ 45.77 mg/dL) and low (< 45.77 mg/dL). This classification resulted in four distinct subgroups: low MBG and low GV, low MBG and high GV, high MBG and low GV, and high MBG and high GV. To ensure comparability, both MBG and GV were standardized and expressed in 10-mg/dL increments.
Univariate analyses were conducted to identify preliminary associations with 28-day mortality. Multivariable Cox regression models were used to assess the relationships between MBG, GV, and their combination with 28-day mortality, adjusting for potential confounders based on clinical relevance and literature, including covariates that changed effect estimates by ≥ 10%. MBG and GV were analyzed as continuous and categorical variables in three models. Generalized variance inflation factors (GVIF^(1/(2Df))) for all covariates were < 2, indicating absence of significant multicollinearity. We conducted three sensitivity analyses: (1) multivariable analysis using non-imputed data, (2) separate assessments of the associations between MBG, GV, and mortality in diabetic and non-diabetic populations, (3) subgroup analysis. Restricted cubic spline models were used to explore potential nonlinear associations between MBG, GV, and mortality. Kaplan–Meier survival curves stratified by MBG, GV, and their combinations were compared using log-rank tests.
Statistical analyses were performed using R 4.2.2 (http://www.R-project.org, The R Foundation) and Free Statistics software version 2.0, with statistical significance set at p < 0.05.
Results
Baseline characteristics
This study enrolled 1,242 patients aged ≥ 65 years with AMI (mean age 77.3 ± 7.5 years), categorized into three cohorts: 65–75 years (n = 540), 75–85 years (n = 473), and ≥ 85 years (n = 229). The flowchart of the patient inclusion process is presented in Fig. 1 and the baseline demographics are summarized in Table 1. The sex distribution was comparable across age groups (64.3% male overall; p = 0.603). BMI decreased with age (29.1 ± 7.0 vs. 25.9 ± 6.1 kg/m²; p < 0.001), while the prevalence of STEMI exhibited a U-shaped pattern. The comorbidity burden increased with age, as reflected by higher Charlson Comorbidity Index scores. Notably, diabetes prevalence paradoxically declined in older cohorts (53.9% vs. 34.5%; p < 0.001), in contrast to the increased prevalence of congestive heart failure and atrial fibrillation.
Fig. 1.
Study flowchart
Table 1.
Baseline characteristics of older patients with acute myocardial Infarction
| Variables | Total | Age, year | p - value | ||
|---|---|---|---|---|---|
| (n = 1242) | 65–75 (n = 540) | 75–85 (n = 473) | ≥85 (n = 229) | ||
| General Characteristics | |||||
| Age, year | 77.3 ± 7.5 | 70.4 ± 2.8 | 79.6 ± 2.9 | 89.1 ± 2.7 | < 0.001 |
| Male, n (%) | 798 (64.3) | 353 (65.4) | 304 (64.3) | 141 (61.6) | 0.603 |
| Marital status, n (%) | < 0.001 | ||||
| Single | 193 (15.5) | 109 (20.2) | 58 (12.3) | 26 (11.4) | |
| Married | 574 (46.2) | 256 (47.4) | 224 (47.4) | 94 (41.0) | |
| Divorced | 91 (7.3) | 47 (8.7) | 38 (8.0) | 6 (2.6) | |
| Widowed | 164 (13.2) | 37 (6.9) | 64 (13.5) | 63 (27.5) | |
| Other | 220 (17.7) | 91 (16.9) | 89 (18.8) | 40 (17.5) | |
| Race, n (%) | 0.361 | ||||
| White | 754 (60.7) | 314 (58.1) | 294 (62.2) | 146 (63.8) | |
| Asian | 31 (2.5) | 13 (2.4) | 10 (2.1) | 8 (3.5) | |
| Black | 97 (7.8) | 49 (9.1) | 32 (6.8) | 16 (7.0) | |
| Hispanic/Latino | 37 (3.0) | 22 (4.1) | 10 (2.1) | 5 (2.2) | |
| Other | 323 (26.0) | 142 (26.3) | 127 (26.8) | 54 (23.6) | |
| BMI, kg/m2 | 27.9 ± 6.4 | 29.1 ± 7.0 | 27.5 ± 5.4 | 25.9 ± 6.1 | < 0.001 |
| STEMI, n (%) | 267 (21.5) | 130 (24.1) | 83 (17.5) | 54 (23.6) | 0.029 |
| Charlson comorbidity index | 7.7 ± 2.5 | 6.9 ± 2.4 | 8.2 ± 2.4 | 8.4 ± 2.2 | < 0.001 |
| APS III | 50.2 ± 19.9 | 49.5 ± 20.3 | 50.1 ± 19.9 | 52.1 ± 19.0 | 0.252 |
| SOFA score | 2.0 (0.0, 4.0) | 2.0 (0.0, 4.0) | 2.0 (0.0, 4.0) | 1.0 (0.0, 3.0) | 0.108 |
| Comorbidities | |||||
| Diabetes, n (%) | 607 (48.9) | 291 (53.9) | 237 (50.1) | 79 (34.5) | < 0.001 |
| Congestive heart failure, n (%) | 809 (65.1) | 331 (61.3) | 320 (67.7) | 158 (69) | 0.042 |
| Cerebrovascular disease, n (%) | 231 (18.6) | 98 (18.1) | 90 (19.0) | 43 (18.8) | 0.935 |
| Severe liver disease, n (%) | 20 (1.6) | 14 (2.6) | 5 (1.1) | 1 (0.4) | 0.052 |
| HBP, n (%) | 476 (38.3) | 213 (39.4) | 172 (36.4) | 91 (39.7) | 0.535 |
| AF, n (%) | 619 (49.8) | 229 (42.4) | 261 (55.2) | 129 (56.3) | < 0.001 |
| Cardiogenic Shock, n (%) | 335 (27.0) | 142 (26.3) | 123 (26) | 70 (30.6) | 0.396 |
| Laboratory Measurements | |||||
| White blood cells, 109/L | 10.8 ± 6.5 | 10.9 ± 5.3 | 10.7 ± 7.2 | 10.9 ± 7.3 | 0.920 |
| Hemoglobin, g/dL | 12.0 ± 2.3 | 12.3 ± 2.4 | 11.8 ± 2.2 | 11.8 ± 1.9 | 0.002 |
| Platelets, 109/L | 224.7 ± 84.4 | 231.4 ± 83.7 | 220.6 ± 81.1 | 217.2 ± 91.9 | 0.042 |
| Creatinine, mg/dL | 1.3 (0.9, 1.9) | 1.2 (0.9, 1.9) | 1.3 (1.0, 1.9) | 1.4 (1.0, 2.1) | < 0.001 |
| HbA1c, % | 6.5 ± 1.6 | 6.8 ± 1.9 | 6.4 ± 1.5 | 6.0 ± 1.0 | < 0.001 |
| MBG, 10 mg/dL | 14.7 ± 4.3 | 15.0 ± 4.5 | 14.6 ± 4.2 | 14.2 ± 3.9 | 0.037 |
| GV, 10 mg/dL | 3.4 (2.1, 5.5) | 3.6 (2.1, 5.9) | 3.2 (2.1, 5.5) | 3.2 (2.0, 4.7) | 0.029 |
| cTnI, μg/L | 1.0 (0.3, 3.0) | 1.1 (0.3, 3.1) | 0.8 (0.2, 2.5) | 1.2 (0.3, 3.2) | 0.024 |
| Treatment | |||||
| CRRT, n (%) | 109 (8.8) | 57 (10.6) | 38 (8.0) | 14 (6.1) | 0.106 |
| Mechanical ventilation, n (%) | 724 (58.3) | 337 (62.4) | 284 (60.0) | 103 (45.0) | < 0.001 |
| Vasoactive drugs, n (%) | 592 (47.7) | 264 (48.9) | 214 (45.2) | 114 (49.8) | 0.397 |
| Alteplase, n (%) | 138 (11.1) | 66 (12.2) | 50 (10.6) | 22 (9.6) | 0.512 |
| PCI, n (%) | 233 (18.8) | 92 (17.0) | 87 (18.4) | 54 (23.6) | 0.101 |
| Antiplatelet Therapy, n (%) | 1161 (93.5) | 508 (94.1) | 448 (94.7) | 205 (89.5) | 0.025 |
| Antithrombotic Therapy, n (%) | 88 (7.1) | 39 (7.2) | 28 (5.9) | 21 (9.2) | 0.286 |
| Calcium Channel Blockers, n (%) | 404 (32.5) | 180 (33.3) | 167 (35.3) | 57 (24.9) | 0.019 |
| Statins, n (%) | 1148 (92.4) | 505 (93.5) | 439 (92.8) | 204 (89.1) | 0.096 |
| ACEI, n (%) | 408 (32.9) | 200 (37) | 152 (32.1) | 56 (24.5) | 0.003 |
| ARB, n (%) | 158 (12.7) | 67 (12.4) | 69 (14.6) | 22 (9.6) | 0.171 |
| Insulin, n (%) | 954 (76.8) | 447 (82.8) | 375 (79.3) | 132 (57.6) | < 0.001 |
Abbreviations STEMI ST-segment elevation myocardial infarction, NSTEMI Non-ST-segment elevation myocardial infarction, APS III Acute Physiology Score III, SOFA Sequential Organ Failure Assessment, AF Atrial Fibrillation, CRRT Continuous Renal Replacement Therapy, cTnI Cardiac Troponin I, PCI Percutaneous Coronary Intervention, ARB Angiotensin II Receptor Blocker, ACEI Angiotensin-Converting Enzyme Inhibitor, MBG Mean Blood Glucose, GV Glucose Variability
Glucose measurements (MBG and GV) are reported in 10 mg/dL units to align with the scaled analysis in regression models and forest plots
Therapeutic strategies varied significantly with age. MV, antiplatelet therapy, and insulin use decreased progressively in older adults. Mortality increased significantly across the age strata (17.6% vs. 34.9% in the oldest cohort; p < 0.001), coinciding with worsening renal function and decreasing hemoglobin levels. Glycemic profiles showed modest reductions in mean blood glucose levels and attenuated variability in the ≥ 85-year subgroup. Although the frequency of hypoglycemia decreased with age, this association was not statistically significant (p = 0.190).
Association between MBG, GV, and 28-d mortality in older adults with AMI
Univariable analysis (Supplementary Table 1, Additional File 1) indicated that patients aged ≥ 85 years had a 2.20-fold higher mortality risk compared to those aged ≤ 65 years (95% confidence interval [CI], 1.63–2.96). The Kaplan–Meier survival curves (Fig. 2A, B) revealed significant differences among the unadjusted glycemic strata.
Fig. 2.
Kaplan–Meier curves of MBG, GV, and their combination for 28-day mortality in older adults with acute myocardial infarction (A) MBG stratification: Low (< 140 mg/dL), Moderate (140–180 mg/dL), High (≥ 180 mg/dL) (B) GV stratification: Low (< 23.82 mg/dL), Moderate (23.82–45.77 mg/dL), High (≥ 45.77 mg/dL) (C) Combined MBG and GV stratification: Low MBG + Low GV: MBG < 180 mg/dL and GV < 45.77 mg/dL Low MBG + High GV: MBG < 180 mg/dL and GV ≥ 45.77 mg/dL High MBG + Low GV: MBG ≥ 180 mg/dL and GV < 45.77 mg/dL High MBG + High GV: MBG ≥ 180 mg/dL and GV ≥ 45.77 mg/dL
Subsequent multivariable Cox regression analysis (Table 2) confirmed that both glycemic markers independently contributed to mortality risk. MBG exhibited a strong dose-response relationship, with patients in the high MBG group experiencing a 5.22-fold higher risk of mortality (95% CI, 3.56–7.66) compared to those in the low MBG group. Additionally, each 10 mg/dL increase in MBG was associated with a 17% increase in mortality risk (hazard ratio [HR] = 1.17; 95% CI, 1.13–1.20). In contrast, GV appeared to demonstrate a threshold effect, as only the highest tertile remained statistically significant after full adjustment (HR = 2.14; 95% CI, 1.46–3.13; p < 0.001), corresponding to a 6% increase in risk per 10 mg/dL increment (HR = 1.06; 95% CI, 1.03–1.08).
Table 2.
Multivariable analysis of MBG, GV, and their combination with 28-Day mortality in older patients with acute myocardial Infarction
| Variable |
n. total |
n. event (%) |
Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|---|---|
| HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | |||
| MBG, per 10 mg/dL | 1242 | 295 (23.8) | 1.12 (1.10–1.15) | < 0.001 | 1.13 (1.11–1.16) | < 0.001 | 1.17 (1.13–1.20) | < 0.001 |
| MBG, Categories | ||||||||
| Low | 683 | 104 (15.2) | 1(Ref) | 1(Ref) | 1(Ref) | |||
| Moderate | 320 | 91 (28.4) | 2.02 (1.53–2.68) | < 0.001 | 2.07 (1.56–2.75) | < 0.001 | 2.16 (1.56–3.00.56.00) | < 0.001 |
| High | 239 | 100 (41.8) | 3.27 (2.49–4.31) | < 0.001 | 3.58 (2.71–4.72) | < 0.001 | 5.22 (3.56–7.66) | < 0.001 |
| P for Trend | < 0.001 | < 0.001 | < 0.001 | |||||
| GV, per 10 mg/dL | 1242 | 295 (23.8) | 1.05 (1.03–1.07) | < 0.001 | 1.05 (1.03–1.07) | < 0.001 | 1.06 (1.03–1.08) | < 0.001 |
| GV, Tertiles | ||||||||
| Low | 414 | 66 (15.9) | 1(Ref) | 1(Ref) | 1(Ref) | |||
| Moderate | 414 | 98 (23.7) | 1.55 (1.13–2.11) | 0.006 | 1.50 (1.10–2.05) | 0.011 | 1.29 (0.92–1.81) | 0.133 |
| High | 414 | 131 (31.6) | 2.15 (1.60–2.90) | < 0.001 | 2.25 (1.67–3.03) | < 0.001 | 2.14 (1.46–3.13) | < 0.001 |
| P for Trend | < 0.001 | < 0.001 | < 0.001 | |||||
| MBG & GV Combination | ||||||||
| Low MBG + low GV | 793 | 143 (18.0) | 1(Ref) | 1(Ref) | 1(Ref) | |||
| Low MBG + high GV | 210 | 52 (24.8) | 1.42 (1.04–1.96) | 0.029 | 1.49 (1.08–2.05) | 0.015 | 1.60 (1.11–2.32) | 0.012 |
| High MBG + low GV | 35 | 21 (60.0) | 5.02 (3.17–7.94) | < 0.001 | 5.66 (3.56–9.00.56.00) | < 0.001 | 5.77 (3.36–9.92) | < 0.001 |
| High MBG + high GV | 204 | 79 (38.7) | 2.42 (1.84–3.18) | < 0.001 | 2.65 (2.01–3.49) | < 0.001 | 3.50 (2.39–5.12) | < 0.001 |
| P for Trend | < 0.001 | < 0.001 | < 0.001 | |||||
Model 1: Unadjusted
Model 2: Adjusted for sex, age, race
Model 3: Model 2 + AMI type, ICU duration, marital status, Charlson Comorbidity Index, APS III, SOFA score, congestive heart failure, cerebrovascular disease, severe liver disease, hypertension, cardiogenic shock, diabetes, creatinine, HbA1c, cTnI, hypoglycemia, CRRT, mechanical ventilation, alteplase, PCI, statins, ACE inhibitors, ARBs, insulin, vasoactive drugs, antiplatelet therapy, anticoagulant therapy, calcium channel blockers, β-blockers, and BMI
Variable Definitions
MBG Categories:
Low: <140 mg/dL
Moderate: 140–180 mg/dL
High: ≥180 mg/dL
GV Tertiles:
Low: <23.82 mg/dL
Moderate: 23.82–45.77 mg/dL
High: ≥45.77 mg/dL
MBG & GV Combinations:
Low MBG + Low GV: MBG < 180 mg/dL & GV < 45.77 mg/dL
Low MBG + High GV: MBG < 180 mg/dL & GV ≥ 45.77 mg/dL
High MBG + Low GV: MBG ≥ 180 mg/dL & GV < 45.77 mg/dL
High MBG + High GV: MBG ≥ 180 mg/dL & GV ≥ 45.77 mg/dL
Notes: MBG groups were defined by preset clinical cutoffs
Restricted cubic spline analyses further characterized the exposure-response relationships (Supplementary Fig. 1, Additional File 1). Both MBG and GV were significantly associated with mortality at the global level (p < 0.001). The nonlinearity test for MBG was not statistically significant (p-nonlinearity = 0.194), whereas GV exhibited a consistent linear association across its range (p-nonlinearity = 0.943).
Association between the combination of MBG and GV and 28-day mortality in older adults with AMI
Kaplan–Meier survival analysis (Fig. 2C) revealed that among the combined glycemic stratifications, the high MBG/low GV subgroup experienced an unexpectedly rapid decline in survival, with only 40.0% of patients surviving at 28 days and a 50% mortality rate reached by day 14—substantially earlier than in the other subgroups. Using the low MBG/low GV group as the reference, multivariable Cox regression analysis demonstrated distinct risk patterns. Specifically, patients in the low MBG/high GV group had a 60% higher mortality risk (hazard ratio [HR] = 1.60; 95% confidence interval [CI], 1.11–2.32; p = 0.012), those in the high MBG/high GV group exhibited a 3.5-fold increased risk (HR = 3.50; 95% CI, 2.39–5.12; p < 0.001), and notably, the high MBG/low GV subgroup showed the strongest association with mortality, with a 5.77-fold increased risk (95% CI, 3.36–9.92; p < 0.001). Significant trend associations were observed across all stratified analyses (p < 0.001).
Other factors associated with 28-day mortality
Multivariable Cox regression analysis (Supplementary Table 2, Additional File 1) identified several independent predictors of increased 28-day mortality, including higher APS III scores (HR = 1.02; 95% CI, 1.01–1.02; p < 0.001), higher Charlson Comorbidity Index scores (HR = 1.15; 95% CI, 1.08–1.22; p < 0.001), cerebrovascular disease (HR = 2.07; 95% CI, 1.53–2.79; p < 0.001), cardiogenic shock (HR = 2.03; 95% CI, 1.53–2.70; p < 0.001), continuous renal replacement therapy (CRRT) (HR = 2.01; 95% CI, 1.39–2.91; p < 0.001), and mechanical ventilation (MV) (HR = 1.91; 95% CI, 1.41–2.59; p < 0.001).
Conversely, β-blockers, CCBs, thrombolysis, PCI, ACEIs, ARBs, insulin, and diabetes were associated with protective effects (p < 0.05). Longer ICU stays, higher SOFA scores, and elevated HbA1c levels were also linked to lower mortality risk. These findings underscore the key prognostic factors influencing short-term survival in older adults with AMI.
Association between MBG and GV with hypoglycemia in older adults with AMI
In the multivariable analysis evaluating the associations of MBG and GV with the secondary outcome of hypoglycemia (Supplementary Table 3, Additional File 1), GV demonstrated a dose-dependent relationship, with patients in the highest tertile exhibiting a markedly increased risk compared to those in the lowest tertile (tertile 3 vs. tertile 1: HR = 4.40; 95% CI, 2.56–7.57).
Conversely, higher MBG levels were paradoxically associated with a protective effect, with each 10 mg/dL increment corresponding to a 7% reduction in hypoglycemia risk (HR = 0.93; 95% CI, 0.88–0.97; p = 0.002). Furthermore, combined analysis revealed that patients with low MBG/high GV had significantly increased odds of hypoglycemia (odds ratio [OR] = 2.42; 95% CI, 1.67–3.52), whereas no hypoglycemic events were observed in the high MBG/low GV subgroup.
Sensitivity analysis
Three complementary sensitivity analyses confirmed the robustness of the associations between glycemic metrics and 28-day mortality in older adults with AMI.
First, a complete-case analysis without imputation (n = 580) demonstrated that each 10 mg/dL increase in MBG and GV was associated with an increased mortality risk (MBG: adjusted HR = 1.27; 95% CI, 1.20–1.35; GV: HR = 1.11; 95% CI, 1.06–1.16; both p < 0.001). Stratification by clinically defined MBG categories (< 140, 140–180, ≥ 180 mg/dL) and GV tertiles revealed pronounced dose-response gradients, with the highest MBG category and GV tertile exhibiting 8.95- and 3.93-fold mortality risks, respectively (p < 0.001). Notably, the high-MBG/low-GV subgroup displayed an exceptionally elevated risk (HR = 9.35 vs. the low-MBG/low-GV reference; p < 0.001), suggesting a synergistic effect between sustained hyperglycemia and reduced glucose stabilization (Supplementary Table 4, Additional File 1).
Second, stratified analyses (Supplementary Table 5, Additional File 1) revealed significant associations between both MBG and GV and mortality across diabetic and non-diabetic cohorts. MBG demonstrated a consistent mortality correlation when analyzed as both a continuous and categorical variable in both populations, whereas GV exhibited significant associations only as a continuous variable. Notably, categorical GV stratification showed a significant mortality association only in non-diabetic patients with high GV. A threshold-driven mortality pattern was observed across glycemic parameters: for MBG categorized as low (< 140 mg/dL), moderate (140–180 mg/dL), and high (≥ 180 mg/dL), mortality rates increased progressively from 10.4% to 37.9% in diabetic patients, whereas non-diabetic patients exhibited a more pronounced escalation from 16.9% to 71.4%. GV tertile analysis revealed incremental mortality in both cohorts, with a 15.3%–26.9% gradient across the low-to-high GV strata, whereas non-diabetic patients showed a steeper increase from 16.1% in the lowest tertile to 51.9% in the highest tertile. Strikingly, the high-MBG/low-GV combination was associated with the highest mortality in both cohorts, contrasting with conventional risk patterns.
Third, subgroup analyses stratified by age, sex, AMI type, and diabetes status demonstrated consistent MBG-mortality associations, with some quantitative variations observed between AMI subtypes. No significant heterogeneity was detected across sex, diabetes status, or age categories (all P-interaction > 0.05), as indicated by overlapping confidence intervals. GV exhibited similar stability across the subgroups (Fig. 3A, B).
Fig. 3.
Forest plot of MBG and GV for 28-day mortality in in older adults with acute myocardial infarction
Discussion
This study systematically examined the independent and combined effects of MBG and GV on 28-day mortality in older adults with AMI. Multivariable Cox regression analysis revealed a dose-dependent relationship between MBG and mortality, with the highest MBG group (MBG ≥ 180 mg/dL) exhibiting a hazard ratio (HR) of 5.22 (95% CI, 3.56–7.66). In contrast, GV demonstrated a threshold effect, with a significantly increased mortality risk observed only in the highest GV tertile (HR = 2.14; 95% CI, 1.46–3.13). Notably, the mortality risk associated with hyperglycemia was substantially greater in non-diabetic patients than in those with diabetes. Among the glycemic subgroups, the highest mortality risk was observed in patients with high MBG and low GV (HR = 5.77; 95% CI, 3.36–9.92), whereas those with both high MBG and high GV had a comparatively lower risk (HR = 3.50; 95% CI, 2.39–5.12), a pattern consistent across both diabetic and non-diabetic populations. Curve-fitting analysis demonstrated significant linear associations between MBG, GV, and 28-day mortality.
Dysglycemia is prevalent among critically ill patients, primarily due to stress-induced disruptions in glucose homeostasis mediated by glucagon, cortisol, thyroid hormones, and growth hormone [10–12]. However, data on dysglycemic patterns in older adults with critical AMI remain limited, with most studies focusing on admission hyperglycemia and the prognostic significance of the stress hyperglycemia ratio (SHR). Chen et al. demonstrated that SHR predicts in-hospital adverse events in 2,485 PCI-treated older adults with AMI, particularly in those with diabetes [13]. In a meta-analysis examining SHR and cardiovascular mortality, six studies demonstrated that hospitalized patients with cardiovascular diseases in the highest SHR quantile had significantly elevated risks of both short- and long-term all-cause mortality compared to those with lower SHR (pooled RR, 1.67; 95% CI, 1.46–1.91; P < 0.001) [14]. Similarly, Kosiborod et al. [15] analyzed data from the CCP registry (n = 141,680; age >65 years) and identified distinct mortality patterns, wherein higher admission glucose levels were associated with increased 30-day mortality. However, in diabetic patients, excess mortality risk was evident only in cases of severe hyperglycemia. Notably, normoglycemic non-diabetic patients exhibited lower adjusted 30-day mortality than those with diabetes but demonstrated the steepest escalation in mortality risk at higher glucose levels. The association between elevated glucose levels and worse outcomes in older adults with AMI is consistent with the findings of the present study. While our findings align with established evidence linking hyperglycemia to adverse outcomes in older AMI patients, this study introduces a novel approach by employing MBG and GV as primary metrics—moving beyond isolated admission glucose measurements. This methodology provides a more comprehensive prognostic assessment and underscores the mortality reduction potential through optimized glycemic control in this population. Kosiborod et al. [5] further demonstrated that glucose metrics spanning the entire hospitalization period are superior to those limited to the first 24–48 h in predicting patient outcomes. Persistent hyperglycemia was found to be a stronger predictor of mortality than admission glucose, with MBG identified as the most practical marker for glycemic control in AMI patients, irrespective of diabetes status. These findings align with those of the present study. Hyperglycemia exacerbates oxidative stress and inflammation [6], impairs endothelium-dependent vasodilation, disrupts endothelial repair, and amplifies immune responses in ischemic myocardial tissues. These mechanisms collectively hinder cardiac remodeling, increase infarct size, and elevate mortality risk [16–18]. Compared with single-point glucose measurements, MBG throughout hospitalization provides a more comprehensive assessment of glycemic control, offering greater prognostic value in older adults with AMI.
GV represents another critical aspect of dysglycemia, contributing to increased coronary plaque vulnerability through mechanisms such as inflammation and oxidative stress [19]. Elevated GV has been associated with cardiac fibrosis [20] and adverse left ventricular remodeling [7], both of which contribute to poor cardiovascular outcomes. Additionally, heightened GV activates the sympathetic-adrenal system, increasing the risk of both hypoglycemic and hyperglycemic episodes [21] and potentially elevating the likelihood of ventricular arrhythmias [10, 13].
Gerbaud et al. [22] found that increased GV (standard deviation >2.7 mmol/L) during hospitalization was the strongest independent predictor of midterm major adverse cardiovascular events (MACE) in patients with diabetes following AMI. Similarly, Yang et al. [23] conducted a study involving 7,510 Chinese patients with STEMI and found that high GV during hospitalization was significantly associated with an increased risk of 30-day all-cause mortality and MACE, regardless of diabetes status. The study by Chai et al. [8] produced consistent results.
To our knowledge, this study is the first to investigate the association between MBG, GV, and mortality in critically ill older adults with AMI, and pioneers the combined assessment of these metrics for mortality risk stratification. While MBG reflects overall glycemic control and GV captures glucose fluctuations, their integration provides a comprehensive evaluation of dysglycemic burden. Intriguingly, although prior studies reported differential hyperglycemic responses between diabetic and non-diabetic subgroups, we observed a consistent mortality pattern across all cohorts: patients with high MBG and low GV exhibited the highest mortality risk, surpassing those with elevated GV alone. This suggests sustained hyperglycemia may exert greater detrimental effects than acute glucose excursions in this population.
The precise mechanisms underlying this phenomenon warrant further investigation. Zheng [24] et al. demonstrated time-dependent endothelial apoptosis and mitochondrial fragmentation in human umbilical vein endothelial cells following prolonged hyperglycemic exposure (≥ 12 h), with mitochondrial disintegration preceding vascular injury—a potential pathway linking sustained hyperglycemia to oxidative stress and cellular demise [25]. Furthermore, elevated glucose levels correlate with multiple pathological processes including: impaired myocardial glucose utilization and microvascular dysfunction [26], prothrombotic states and vascular inflammation [27], free fatty acid-induced insulin resistance and endothelial dysfunction [28], reactive oxygen species generation [29]. Collectively, these mechanisms may exacerbate tissue damage in the ischemic myocardium.
Strengths and limitations
This study has several strengths, including a large sample size, complete mortality data, and the first analysis of both the independent and combined effects of MBG and GV in older adults with AMI. However, several limitations should be acknowledged.
First, the retrospective, single-center design and missing data in the MIMIC-IV database limit the generalizability of our findings, highlighting the need for prospective, multicenter studies. Second, variability in the timing and frequency of blood glucose measurements across patients may have introduced bias. Third, the underlying mechanisms driving the highest mortality risk in patients with high MBG and low GV in both diabetic and non-diabetic populations remain unclear and warrant further investigation. Fourth, although oral hypoglycemic agents (OHAs) were not included as covariates, only 16 diabetic patients not using insulin (2.6% of diabetic cohort; Supplementary Table 6) were identified, suggesting minimal confounding impact on outcome associations. Finally, as with all observational studies, causality cannot be established, and despite adjusting for multiple confounders, residual confounding factors cannot be ruled out.
Conclusions
Elevated mean blood glucose and glycemic variability independently predicted 28-day mortality in older adults with acute myocardial infarction. Critically, sustained hyperglycemia demonstrated greater clinical detriment than acute glucose fluctuations, necessitating risk-stratified glycemic protocols prioritizing MBG control in this vulnerable population.
Supplementary Information
Acknowledgements
We thank Dr. Liu Jie (People’s Liberation Army General Hospital, Beijing, China), Dr. Li Haibo, Dr. Jiang Guosong, and the team of Clinical Scientists for their valuable assistance in revising this manuscript.
Abbreviations
- MBG
Mean blood glucose
- GV
Glycemic variability
- AM I
Acute myocardial infarction
- STEMI
ST-segment elevation myocardial infarction
- NSTEMI
Non-ST-segment elevation myocardial infarction
- APS III
Acute Physiology Score III
- SOFA
Sequential Organ Failure Assessment
- AF
Atrial Fibrillation
- CRRT
Continuous Renal Replacement Therapy
- CTNI
Cardiac Troponin I
- PCI
Percutaneous Coronary Intervention
- ARB
Angiotensin II Receptor Blocker
Authors’ contributions
Jun Zhou, Xiaomei Deng, Peng Zhou, Xinlin Luo, Hao Li, and Xiaoyun Fan contributed to the conception, design, conduct of the study, and the analysis and interpretation of the results. Jun Zhou was responsible for data mining and drafting the initial manuscript. Peng Zhou conducted the statistical analysis. Xiaomei Deng and Xinlin Luo were responsible for figure generation. Hao Li edited, reviewed, and approved the final version of the manuscript. Xiaoyun Fan is the guarantor of this work, with full access to all data in the study and responsibility for the integrity and accuracy of the data analysis. Hao Li is the co-corresponding author.
Funding
This study was supported by the following funding sources:
1. “Epidemiological Survey and Prediction Model Construction of Air Pollution-Induced Exacerbation in COPD Patients,” 2023 Anhui Province Clinical Medical Research Transformation Project, Project No: 202304295107020044;
2. “Biochemistry and Molecular Biology Co-construction Project of the School of Geriatrics and Life Sciences,” Anhui Medical University, 2022 Clinical and Preliminary Discipline Co-construction Project, Project No: 2022lcxk020;
3. The Clinical Research Project of Shenzhen Second People’s Hospital, Project No: 2023yjlcyj003.
4. Shenzhen Clinical Research Center for Cardiovascular Diseases Fund, Project No: 20220819165348002.
Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The study adhered to the Declaration of Helsinki, and the requirement for informed consent was waived as the data were anonymized. The institutional review board of Beth Israel Deaconess Medical Center approved the study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Hao Li, Email: snapzero@163.com.
Xiaoyun Fan, Email: 13956988552@126.com.
References
- 1.Tsao CW, Aday AW, Almarzooq ZI, Anderson CAM, Arora P, Avery CL, et al. Heart disease and stroke Statistics-2023 update: A report from the American heart association. Circulation. 2023;147(8):e93–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Morici N, De Servi S, De Luca L, Crimi G, Montalto C, De Rosa R, et al. Management of acute coronary syndromes in older adults. Eur Heart J. 2022;43(16):1542–53. [DOI] [PubMed] [Google Scholar]
- 3.García-Blas S, Cordero A, Diez-Villanueva P, Martinez-Avial M, Ayesta A, Ariza-Solé A, et al. Acute coronary syndrome in the older patient. J Clin Med. 2021;10(18):4132. 10.3390/jcm10184132. [DOI] [PMC free article] [PubMed]
- 4.Marx N, Müller-Wieland D, Verket M, Schütt K. [Management of cardiovascular diseases in patients with diabetes: ESC guidelines 2023]. Herz. 2024;49(1):15–8. [DOI] [PubMed] [Google Scholar]
- 5.Kosiborod M, Inzucchi SE, Krumholz HM, Xiao L, Jones PG, Fiske S, et al. Glucometrics in patients hospitalized with acute myocardial infarction: defining the optimal Outcomes-Based measure of risk. Circulation. 2008;117(8):1018–27. [DOI] [PubMed] [Google Scholar]
- 6.Esdaile H, Hill N, Mayet J, Oliver N. Glycaemic control in people with diabetes following acute myocardial infarction. Diabetes Res Clin Pract. 2023;199:110644. [DOI] [PubMed] [Google Scholar]
- 7.Gohbara M, Iwahashi N, Kataoka S, Hayakawa Y, Sakamaki K, Akiyama E, et al. Glycemic variability determined by continuous glucose monitoring system predicts left ventricular remodeling in patients with a first ST-Segment elevation myocardial infarction. Circ J. 2015;79(5):1092–9. [DOI] [PubMed] [Google Scholar]
- 8.Chai TY, McLean M, Wong VW, Cheung NW. Glycaemic variability is associated with adverse cardiovascular outcomes in patients hospitalised with an acute myocardial infarction. J Clin Transl Endocrinol. 2019;18:100203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Johnson AEW, Bulgarelli L, Pollard TJ, Gow B, Moody B, Horng S, et al. MIMIC-IV (version 3.0). PhysioNet; 2024. RRID:SCR_007345. Available from: 10.13026/hxp0-hg59.
- 10.Su Y, Fan W, Liu Y, Hong K. Glycemic variability and in-hospital death of critically ill patients and the role of ventricular arrhythmias. Cardiovasc Diabetol. 2023;22(1):134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lu L, Lu Y, Gao C, Zhang N. Age moderates the relationships between obesity, glucose variability, and intensive care unit mortality: a retrospective cohort study. J Intensive Care. 2021;9:68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lee TF, Drake SM, Roberts GW, Bersten A, Stranks SN, Heilbronn LK, et al. Relative hyperglycemia is an independent determinant of In-Hospital mortality in patients with critical illness. Crit Care Med. 2020;48(2):e115–22. [DOI] [PubMed] [Google Scholar]
- 13.Chen G, Li M, Wen X, Wang R, Zhou Y, Xue L, et al. Association between stress hyperglycemia ratio and In-hospital outcomes in older adults with acute myocardial infarction. Front Cardiovasc Med. 2021;8:698725. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Esdaile H, Khan S, Mayet J, Oliver N, Reddy M, Shah ASV. The association between the stress hyperglycaemia ratio and mortality in cardiovascular disease: a meta-analysis and systematic review. Cardiovasc Diabetol. 2024;23(1):412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kosiborod M, Rathore SS, Inzucchi SE, Masoudi FA, Wang Y, Havranek EP, et al. Admission Glucose and Mortality in older adults Hospitalized With Acute Myocardial Infarction: Implications for Patients With and Without Recognized Diabetes. Circulation. 2005;14(23):3078–86. [DOI] [PubMed] [Google Scholar]
- 16.Esposito K, Nappo F, Marfella R, Giugliano G, Giugliano F, Ciotola M, et al. Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: role of oxidative stress. Circulation. 2002;106(16):2067–72. [DOI] [PubMed] [Google Scholar]
- 17.Esposito K, Marfella R, Giugliano D. Stress Hyperglycemia, Inflammation, and cardiovascular events. Diabetes Care. 2003;26(5):1650–1. [DOI] [PubMed] [Google Scholar]
- 18.Monnier L, Mas E, Ginet C, Michel F, Villon L, Cristol JP, et al. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA. 2006;295(14):1681. [DOI] [PubMed] [Google Scholar]
- 19.Kuroda M, Shinke T, Sakaguchi K, Otake H, Takaya T, Hirota Y, et al. Effect of daily glucose fluctuation on coronary plaque vulnerability in patients Pre-Treated with Lipid-Lowering therapy. JACC: Cardiovasc Interventions. 2015;8(6):800–11. [DOI] [PubMed] [Google Scholar]
- 20.Saito S, Teshima Y, Fukui A, Kondo H, Nishio S, Nakagawa M, et al. Glucose fluctuations increase the incidence of atrial fibrillation in diabetic rats. Cardiovasc Res. 2014;104(1):5–14. [DOI] [PubMed] [Google Scholar]
- 21.Shimabukuro M, Tanaka A, Sata M, Dai K, Shibata Y, Inoue Y, et al. α-Glucosidase inhibitor Miglitol attenuates glucose fluctuation, heart rate variability and sympathetic activity in patients with type 2 diabetes and acute coronary syndrome: a multicenter randomized controlled (MACS) study. Cardiovasc Diabetol. 2017;16(1):86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Catargi B, Darier R, Montaudon M, Beauvieux MC, Coffin-Boutreux C, Coste P, et al. Response to Comment on Gerbaud Glycemic Variability Is a Powerful Independent Predictive Factor of Midterm Major Adverse Cardiac Events in Patients With Diabetes With Acute Coronary Syndrome. Diabetes Care. 2019;42:674–81 Diabetes Care. 2019;42(10):e170–1. [DOI] [PubMed] [Google Scholar]
- 23.Yang X, Su G, Zhang T, Yang H, Tao H, Du X, et al. Comparison of admission glycemic variability and glycosylated hemoglobin in predicting major adverse cardiac events among type 2 diabetes patients with heart failure following acute ST-segment elevation myocardial infarction. J Transl Int Med. 2024;12(2):188–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zheng Y, Luo A, Liu X. The imbalance of mitochondrial Fusion/Fission drives High-Glucose-Induced vascular injury. Biomolecules. 2021;11(12):1779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Williams M, Caino MC. Mitochondrial dynamics in type 2 diabetes and cancer. Front Endocrinol. 2018;9:211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Scognamiglio R, Negut C, de Kreutzenberg SV, Tiengo A, Avogaro A. Effects of different insulin regimes on postprandial myocardial perfusion defects in type 2 diabetic patients. Diabetes Care. 2006;29(1):95–100. [PubMed] [Google Scholar]
- 27.Gresele P, Guglielmini G, De Angelis M, Ciferri S, Ciofetta M, Falcinelli E, et al. Acute, short-term hyperglycemia enhances shear stress-induced platelet activation in patients with type II diabetes mellitus. J Am Coll Cardiol. 2003;41(6):1013–20. [DOI] [PubMed] [Google Scholar]
- 28.Oliver MF. Metabolic causes and prevention of ventricular fibrillation during acute coronary syndromes. Am J Med. 2002;112(4):305–11. [DOI] [PubMed] [Google Scholar]
- 29.Guha M, Bai W, Nadler JL, Natarajan R. Molecular mechanisms of tumor necrosis factor alpha gene expression in monocytic cells via hyperglycemia-induced oxidant stress-dependent and -independent pathways. J Biol Chem. 2000;9(23):17728–39. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.