Influence of age on the development of stress hyperglycemia during first‐onset ST‐elevation myocardial infarction Authors: Hon Jen Wong, Joy Xin Yi Au, Yao Hao Teo, Yao Neng Teo, Chen Ee Low, Chun En Yau, Norman HY Lin, Pei Chia Eng, Li Feng Tan, Reshma A Merchant, Ping Chai, Anand Adinath Ambhore, James W L Yip, Andrew Fu‐Wah Ho, David Foo, Pow‐Li Chia, Patrick Zhan‐Yun Lim, Khung Keong Yeo, Weien Chow, Daniel Thuan Tee Chong, Derek J Hausenloy, Mark Y Y Chan, Ching‐Hui Sia Published: 2026 ABSTRACT ABSTRACTIntroductionStress hyperglycemia is associated with adverse outcomes in ST‐elevation myocardial infarction (STEMI), but whether age influences its development or prognostic significance remains unclear. We examined the association between age and stress hyperglycemia and whether age modifies its relationship with mortality.MethodologyWe studied non‐diabetics with first‐onset STEMI from the Singapore Myocardial Infarction Registry (SMIR) (2008–2020). The stress hyperglycemia ratio (SHR) was calculated as random plasma glucose (mmol/L)/[1.59 × HbA1c (%) – 2.59], with stress hyperglycemia defined as SHR ≥1.51. Logistic regression identified factors independently associated with stress hyperglycemia. Restricted cubic splines assessed non‐linearity between age and odds of stress hyperglycemia. Multivariate Cox regression evaluated associations with 30‐day, 1‐year, and 10‐year all‐cause mortality. Analyses were stratified by age and sex.ResultsAmong 8,779 patients, 1,718 (19.6%) developed stress hyperglycemia. Patients with stress hyperglycemia were older and more likely to present with advanced heart failure. Increasing age was independently associated with higher odds of stress hyperglycemia, with no evidence of non‐linearity. Stress hyperglycemia conferred higher short‐ and long‐term mortality across all age groups, although the relative association attenuated with advancing age (P for interaction <0.001). Similar trends were observed in both sexes.ConclusionIn non‐diabetic patients with first‐onset STEMI, the likelihood of stress hyperglycemia increased with age. However, stress hyperglycemia was associated with the greatest risk of mortality in younger patients.Keywords: myocardial infarction, stress hyperglycemia, survival Abbreviations diabetes mellitus percutaneous coronary intervention stress hyperglycemia ratio ST‐elevation myocardial infarction INTRODUCTION ST‐elevation myocardial infarction (STEMI) is a major global health burden 1 . It contributes significantly to the 15% of global deaths attributed to ischemic heart disease 2 . Its high morbidity and mortality highlight the need for early risk stratification 1 . In some patients, STEMI is accompanied by a transient elevation in plasma glucose caused by an acute neurohormonal stress response 3 . This phenomenon is referred to as stress hyperglycemia, which can occur even in patients without diabetes mellitus (DM). The presence of stress hyperglycemia during STEMI is associated with an increased risk of mortality, heart failure, and other adverse outcomes 4 , 5 , 6 . It is therefore recognized as an important prognostic marker 1 . The stress hyperglycemia ratio (SHR), which takes into account baseline glycemic status via glycated hemoglobin levels, has emerged as a potentially more accurate reflection of acute glycemic disturbance than absolute plasma glucose levels 7 . Hence, SHR may improve risk stratification and has demonstrated prognostic value in patients with STEMI, particularly among non‐diabetic patients. 8 The magnitude of the stress hyperglycemic response may vary with age. 3 An advancing age is associated with impaired glucose tolerance, which could arise due to alterations in β‐cell function, changes in the pulsatility of insulin secretion and/or reductions in insulin sensitivity 9 . Furthermore, aging may be accompanied by a decline in excess metabolic capacity, an innate adaptive mechanism that enables each cell to withstand various stressors 10 . In glucose‐regulating cells and organs, this decline may further compromise glycemic control when challenged by a severe stressor such as a STEMI. This idea of metabolic reserve 11 parallels the broader concept of frailty 12 , where a cumulative decline in multiple physiological systems leads to a significant decline in homeostatic reserve, such that even minor stressors can precipitate a disproportionate deterioration in overall health. Hence, we hypothesize that older patients with diminished metabolic reserve are more likely to manifest an exaggerated stress hyperglycemia response 3 , 11 . Although the association between stress hyperglycemia and adverse outcomes after myocardial infarction is previously shown, whether the risk of development and its prognostic value vary by age remains inadequately characterized. In this study, we aimed to address this gap in the literature by investigating age‐related differences in the occurrence of stress hyperglycemia and its impact on outcomes in non‐diabetic patients with STEMI. METHODS We conducted a retrospective observational study on non‐diabetic patients with first‐onset STEMI using data from the Singapore Myocardial Infarction Registry (SMIR). SMIR is maintained by the National Registry of Diseases Office (NRDO) and captures data on all acute myocardial infarction (AMI) cases in Singapore. It includes cases diagnosed in both public (contributing approximately 98%) and private hospitals, alongside a limited number of physician‐certified out‐of‐hospital AMI deaths. Mandatory reporting to SMIR was instituted following the enactment of the National Registry of Diseases Act in 2012. AMI was identified using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) code 410 prior to 2012, and the ICD‐10 Australian Modification (ICD‐10‐AM) code I21 from 2012 onwards. Relevant data were collated from multiple sources, including hospital inpatient discharge records, laboratory cardiac biomarker databases, patient medical claims, and the national death registry. Electrocardiograms were interpreted and diagnoses centrally adjudicated by NRDO to ensure consistency. Annual audits were conducted to assess data accuracy and inter‐rater reliability, with logic checks implemented to identify and resolve discrepancies. The definition of STEMI including (1) typical chest pain lasting at least 20 min, (2) significant ST‐segment elevation (≥0.1 mV in limb leads or ≥0.2 mV in precordial leads), and (3) confirmation by elevated cardiac troponins. Regardless, all diagnoses of STEMI made were ultimately based on clinical judgment and guidelines applicable at that point in time. SHR was calculated as Random plasma glucosemmol/L1.59×HbA1c%−2.59 7 , 13 , 14 . We included all first‐onset STEMI cases reported to SMIR from January 2008 to December 2020 occurring in patients without DM. Non‐DM was defined as the absence of a prior diagnosis of DM and a presenting HbA1c of <6.5% 15 . We restricted the analysis to non‐DM patients to isolate acute stress‐induced hyperglycemia and minimize confounding from chronic glucose tolerance or insulin deficiency, which could obscure interpretation of stress hyperglycemia. This approach also sought to eliminate potential confounding from the effects of anti‐diabetic medications, including those related to poor glycemic control or medication non‐compliance. Patients with missing data that prevented the calculation of SHR were excluded. The primary outcome was the development of stress hyperglycemia, defined as SHR ≥1.51. This cutoff was derived and validated using the same dataset in a previous study 16 . Statistical analysis was performed using R statistical software Version 4.3.1 (R Core Team, 2021), with statistical significance defined as P < 0.05. Continuous variables were summarized via means and standard deviations (SDs) while categorical variables were presented as frequencies and percentages. Comparisons between patients who developed stress hyperglycemia and those who did not were conducted using the Chi‐square test for categorical variables and the Wilcoxon rank‐sum test for continuous variables. The median follow‐up duration was estimated using the reverse Kaplan–Meier method 17 . To assess the influence of age on the development of stress hyperglycemia, we first stratified patients based on age of onset (<40 years, 40 to <60 years, 60 to <80 years and ≥80 years). Univariable logistic regression was then utilized to identify baseline and clinical characteristics associated with stress hyperglycemia. Significant variables were subsequently included in a multivariable model. To examine potential non‐linear associations between age of onset and the development of stress hyperglycemia, multivariate logistic regression analyses with restricted cubic splines with 3 knots were performed and represented graphically. The number of knots used was based on the model with the lowest Akaike information criterion. These analyses were conducted separately for males and females—with the intention of identifying any potential sex‐specific effects. Time‐to‐mortality data were graphically represented using the Kaplan–Meier method and compared using the log‐rank test. Cox proportional hazards regression was utilized to evaluate the association of stress hyperglycemia with 30‐day, 1‐year and 10‐year all‐cause mortality. We directly performed subgroup analysis based on age of STEMI onset (<40 years, 40 to <60 years, 60 to <80 years and ≥80 years) as an interaction term. Multivariate analyses adjusted for sex, race, hemoglobin level, hypertension, hyperlipidemia, presence of chest pain at presentation, heart failure status at admission and reperfusion at admission. This analysis was also performed separately for males and females. Missing values within categorical variables were encoded as a distinct category ‘Missing’ to optimize sample size within the multivariable model 18 , 19 , 20 . These variables were included only if they had low missingness (<10%) 21 . Missing numerical data were handled via complete case analysis. The local institutional review board granted exemption from written consent for participants in this study (National Healthcare Group Domain Specific Review Board: 2023/00263), as it utilized de‐identified data. Only the statistician had access to anonymized patient‐level data. The co‐authors interpreted the analyzed aggregate data. This study was carried out in accordance with the Declaration of Helsinki 22 . RESULTS A total of 8,779 non‐DM patients with first‐onset AMI were included—without stress hyperglycemia (n = 7,061, 80.4%) and with stress hyperglycemia (n = 1,718, 19.6%). The median follow‐up duration was 5.26 years (interquartile range 3.00 –8.02). Baseline and clinical characteristics are presented in Table 1. The mean age of onset was 60 years (SD 12) and most patients in the study cohort were male (n = 7,585, 86%). The mean SHR was 1.30 (SD 0.45). Patients presenting with stress hyperglycemia were more likely to be older and be non‐smokers vs those presenting without stress hyperglycemia. The prevalence of hyperlipidemia was lower in the stress hyperglycemia group, with correspondingly lower total cholesterol, low‐density lipoprotein, and triglyceride levels, and higher high‐density lipoprotein levels. The proportion of patients with hypertension was similar between groups. Those with stress hyperglycemia also had marginally lower BMI (24.4 vs 24.9 kg/m2) and hemoglobin levels (14.13 vs 14.53 g/dL). Table 1Baseline and clinical characteristicsOverall (n = 8,779)Without stress hyperglycemia (n = 7,061, 80.4%)With stress hyperglycemia (n = 1,718, 19.6%) P Age, years, median (IQR)58 (51–67)58 (51–66)61 (53–71)<0.001Sex, n (%)Male7,585 (86%)6,197 (88%)1,388 (81%)<0.001Race, n (%)0.11Chinese6,010 (68%)4,794 (68%)1,216 (71%)Malay1,632 (19%)1,331 (19%)301 (18%)Indian1,002 (11%)822 (12%)180 (10%)Others135 (1.5%)114 (1.6%)21 (1.2%)Medical history, n (%)Hypertension3,878 (44%)3,094 (44%)784 (46%)0.3Hyperlipidemia4,772 (54%)3,910 (55%)862 (50%)<0.001Previous PCI141 (1.6%)108 (1.5%)33 (1.9%)0.4Smoking status, n (%)<0.001Current4,220 (48%)3,493 (49%)727 (42%)Ex‐smoker1,130 (13%)920 (13%)210 (12%)Never3,379 (38%)2,631 (37%)748 (44%)Missing50 (0.6%)17 (0.2%)33 (1.9%)BMI, kg/m2, median (IQR) [n = 8,289)24.4 (22.2–27.1)24.5 (22.3–27.2)24.1 (22.0–26.7)<0.001Random plasma glucose, mmol/L, median (IQR)7.60 (6.60–9.10)7.20 (6.40–8.20)11.20 (9.90–13.30)<0.001HbA1c, %, median (IQR)5.70 (5.40–5.90)5.70 (5.50–5.90)5.60 (5.30–5.90)<0.001Total cholesterol*, mmol/L, median (IQR) [n = 8,491]5.12 (4.36–5.96)5.17 (4.40–6.00)4.93 (4.10–5.79)<0.001LDL cholesterol*, mmol/L, median (IQR) [n = 8,444]3.42 (2.71–4.20)3.48 (2.78–4.24)3.25 (2.50–4.05)<0.001HDL cholesterol*, mmol/L, median (IQR) [n = 8,479]1.10 (0.94–1.30)1.10 (0.94–1.30)1.13 (0.95–1.37)<0.001Triglyceride † , mmol/L, median (IQR) [n = 8,498]1.29 (0.93–1.80)1.30 (0.95–1.83)1.19 (0.87–1.68)<0.001Hemoglobin, g/dL, median (IQR)14.60 (13.50–15.60)14.70 (13.60–15.60)14.40 (13.00–15.40)<0.001Presenting symptoms, n (%)Dyspnea4,601 (52%)3,716 (53%)885 (52%)0.4Chest pain7,792 (89%)6,472 (92%)1,320 (77%)<0.001Diaphoresis5,430 (62%)4,456 (63%)974 (57%)<0.001Back pain805 (9.2%)691 (9.8%)114 (6.6%)<0.001Shoulder pain642 (7.3%)549 (7.8%)93 (5.4%)<0.001Jaw pain573 (6.5%)515 (7.3%)58 (3.4%)<0.001Syncope224 (2.6%)151 (2.1%)73 (4.2%)<0.001Heart failure status at admission, Killip class, n (%) [n = 8,771]<0.001I7,435 (85%)6,215 (88%)1,220 (71%)II427 (4.9%)363 (5.1%)64 (3.7%)III342 (3.9%)204 (2.9%)138 (8.0%)IV567 (6.5%)278 (3.9%)289 (17%)BMI, body mass index; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; PCI, percutaneous coronary intervention; SD, standard deviation; STEMI, ST‐elevation myocardial infarction.*To convert total cholesterol, HDL and LDL from mmol/L to mg/dL, multiply by 38.67. † To convert triglycerides from mmol/L to mg/dL, multiply by 88.57. Baseline and clinical characteristics BMI, body mass index; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; PCI, percutaneous coronary intervention; SD, standard deviation; STEMI, ST‐elevation myocardial infarction. To convert total cholesterol, HDL and LDL from mmol/L to mg/dL, multiply by 38.67. To convert triglycerides from mmol/L to mg/dL, multiply by 88.57. Patients with stress hyperglycemia were less likely to report typical symptoms of STEMI such as chest pain and diaphoresis, but were more likely to experience syncope. They were also more likely to have advanced heart failure at presentation, as evidenced by a higher Killip classification (Table 1). Clinical management The rates of percutaneous coronary intervention (PCI) and coronary artery bypass grafting were similar between groups, despite statistically significant differences (Table 2). Moreover, although statistical significance was observed, there was no clinically meaningful difference in the receipt of single or dual antiplatelet therapy at admission and during hospitalization between the stress hyperglycemia and non‐stress hyperglycemia groups (Table 2). However, patients with stress hyperglycemia appeared less likely to be prescribed dual antiplatelet therapy or more likely to not be prescribed antiplatelet therapy at all at discharge. In addition, those presenting with stress hyperglycemia appeared less likely to receive beta‐blockers, angiotensin‐converting enzyme inhibitors/angiotensin II receptor blockers, and lipid‐lowering drugs during hospitalization. However, there was no difference in the prescription of these drugs between patient groups at discharge (Table 2). Table 2Clinical management of patients with and without stress hyperglycemia during ST‐elevation myocardial infarctionWithout stress hyperglycemia (n = 7,061, 80.4%)With stress hyperglycemia (n = 1,718, 19.6%) P Percutaneous coronary intervention, n (%)6,380 (90%)1,508 (88%)0.002Coronary artery bypass graft, n (%)141 (2.0%)39 (2.3%)0.4Medications prescribed on admission for outpatient‐onset, or within 24 h of onset for inpatient‐onset, n (%)Antiplatelet<0.001SAPT229 (3.2%)64 (3.7%)DAPT6,789 (96%)1,621 (94%)Medications prescribed during hospitalization, n (%)Antiplatelet<0.001SAPT241 (3.4%)66 (3.8%)DAPT6,771 (96%)1,580 (92%)Beta‐blocker6,123 (87%)1,389 (81%)<0.001ACEI/ARB5,207 (74%)1,066 (62%)<0.001Lipid‐lowering drugs6,954 (98%)1,606 (93%)<0.001Medications given at discharge*, n/total (%)Antiplatelet<0.001SAPT327/6,865 (4.8%)102/1,435 (7.1%)DAPT6,485/6,865 (94%)1,306/1,435 (91%)Beta‐blocker5,858/6,865 (85%)1,208/1,435 (84%)0.3ACEI/ARB4,805/6,865 (70%)918/1,435 (64%)<0.001Lipid‐lowering drugs6,772/6,865 (99%)1,384/1,435 (96%)<0.001ACEI/ARB, angiotensin‐converting enzyme inhibitor/angiotensin receptor blocker; DAPT, dual antiplatelet therapy; SAPT, single antiplatelet therapy; STEMI, ST‐elevation myocardial infarction.*For those surviving to discharge. Clinical management of patients with and without stress hyperglycemia during ST‐elevation myocardial infarction ACEI/ARB, angiotensin‐converting enzyme inhibitor/angiotensin receptor blocker; DAPT, dual antiplatelet therapy; SAPT, single antiplatelet therapy; STEMI, ST‐elevation myocardial infarction. For those surviving to discharge. Complications during hospitalization Patients who developed stress hyperglycemia more frequently developed complications during hospitalization, including severe heart failure (Killip class III and IV), cardiogenic shock, left ventricular systolic dysfunction, arrhythmias, heart block, and stroke (Table S1). Influence of age on the development of stress hyperglycemia The mean SHR was 1.22 (SD 0.35) in patients aged <40 years, 1.27 (0.45) in those aged 40 to <60 years, 1.33 (0.46) in those aged 60 to <80 years, and 1.38 (0.48) in those aged ≥80 years. Logistic regression was used to identify variables associated with the development of stress hyperglycemia in the overall cohort (Table 3). After adjustment, older age at onset and more advanced heart failure on admission (Killip class III and IV) were associated with higher odds of stress hyperglycemia. In contrast, male sex, current smoker, and the presence of chest pain at presentation were associated with lower odds. In sex‐stratified analyses, older age of onset and more advanced heart failure at admission (Killip class III and IV) remained independently associated with higher odds of stress hyperglycemia in both males and females (Tables S2 and S3, respectively). In males, chest pain at presentation was independently associated with lower odds. Among females, current smoker, and chest pain at presentation were both independently associated with lower odds of stress hyperglycemia. Table 3Univariable and multivariable models for the odds of developing stress hyperglycemia during ST‐elevation myocardial infarction in the overall cohortCharacteristicUnadjusted OR (95% CI) P Adjusted OR^ (95% CI) P Age of onset<0.0010.008<401.00 (ref)1.00 (ref)40–<601.94 (1.36 to 2.84)1.65 (1.14 to 2.46)60–<802.81 (1.98 to 4.14)1.93 (1.33 to 2.89)≥803.72 (2.52 to 5.62)1.77 (1.14 to 2.80)Male sex0.59 (0.51 to 0.67)<0.0010.77 (0.64 to 0.92)0.005Race0.11Chinese1.00 (ref)–Malay0.89 (0.77 to 1.02)–Indian0.86 (0.72 to 1.02)–Others0.73 (0.44 to 1.14)–Smoking status<0.0010.03Never1.00 (ref)1.00 (ref)Current0.73 (0.65 to 0.82)0.84 (0.73 to 0.96)Ex‐smoker0.80 (0.68 to 0.95)0.89 (0.73 to 1.07)Hypertension1.08 (0.97 to 1.20)0.3–Hyperlipidemia0.81 (0.73 to 0.90)<0.0010.90 (0.80 to 1.01)0.08Body mass index0.97 (0.95 to 0.98)<0.0010.99 (0.98 to 1.01)0.5Hemoglobin0.89 (0.87 to 0.92)<0.0010.98 (0.94 to 1.01)0.2Chest pain at presentation0.30 (0.26 to 0.35)<0.0010.41 (0.34 to 0.48)<0.001Dyspnea at presentation0.96 (0.86 to 1.06)0.4–Diaphoresis at presentation0.77 (0.69 to 0.85)<0.0011.07 (0.95 to 1.22)0.3Heart failure status at admission, Killip class<0.0010.009I1.00 (ref)1.00 (ref)II0.90 (0.68 to 1.17)0.70 (0.52 to 0.94)III3.45 (2.75 to 4.31)2.98 (2.33 to 3.79)IV5.30 (4.44 to 6.31)4.53 (3.76 to 5.46)CI, confidence interval; OR, odds ratio; ref reference.^Only variables significant in univariable analyses were included in the multivariable model. Univariable and multivariable models for the odds of developing stress hyperglycemia during ST‐elevation myocardial infarction in the overall cohort CI, confidence interval; OR, odds ratio; ref reference. ^Only variables significant in univariable analyses were included in the multivariable model. To examine potential non‐linear effects of age and possible sex‐dependent differences, we fitted a multivariable logistic regression model using restricted cubic splines (three knots) with an age‐sex interaction term. Increasing age at onset was independently associated with higher odds of stress hyperglycemia (P = 0.009). However, there was no evidence of a non‐linear effect of age (P = 0.14), nor was there any evidence that the association of age differed by sex (P = 0.25). The fitted spline curve is represented in Figure 1. Restricted cubic spline of odds ratio of developing stress hyperglycemia against age of onset, stratified by sex. (The curves were derived on a univariate model. The reference line indicates an OR of one. Among females, the odds of stress hyperglycemia rose steadily with increasing age. Among males, OR also increased with age and peaked at approximately 60–70 years. Thereafter, OR plateaued, with a slight decline). All‐cause mortality In the overall cohort of non‐diabetic patients with STEMI, the development of stress hyperglycemia was independently associated with a significantly higher risk of all‐cause mortality at 30‐days, 1–2 years and 10‐years compared to no stress hyperglycemia across all age groups (<40 years, 40 to <60 years, 60 to <80 years and ≥80 years). The interaction between age and stress hyperglycemia was significant (P < 0.001), suggesting that the effect of stress hyperglycemia on the risk of short‐ and long‐term all‐cause mortality differed between age groups. Hazard ratios (HRs) were greatest in younger patients, and progressively decreased with increasing age (Table 4). Of note, absolute mortality increased with age, from 1.14% in patients aged <40 years to 23.33% in those aged ≥80 years at 30 days, and from 3.43% to 58.40% at 10 years. However, there was no significant difference in risk at 1‐year among patients aged <40 years. A Kaplan–Meier curve for the overall cohort is presented in Figure 2. Table 4Risks of all‐cause mortality at 30 days, 1, 5, and 10 years in the overall cohort, and in males and females stratified by ageTime pointAgeHR (95% CI) P interactionOverall30 day<0.001<407.50 (1.06–53.3)40–<606.68 (4.57–9.77)60–<802.65 (2.05–3.42)≥802.07 (1.48–2.89)1 year<0.001<405.20 (0.87–31.1)40–<605.26 (3.81–7.26)60–<802.30 (1.87–2.84)≥801.78 (1.35–2.33)10 year<0.001<404.38 (1.32–14.6)40–<603.13 (2.50–3.91)60–<801.85 (1.58–2.17)≥801.42 (1.13–1.78)Male‐only30 day<0.001<4012.86 (1.16–141.9)40–<606.01 (4.07–8.86)60–<802.62 (1.95–3.52)≥801.99 (1.19–3.33)1 year<0.001<406.77 (0.95–48.1)40–<604.89 (3.51–6.81)60–<802.46 (1.94–3.13)≥801.96 (1.30–2.97)10‐year<0.001<404.58 (1.34–15.65)40–<603.00 (2.38–3.80)60–<801.91 (1.60–2.30)≥801.47 (1.04–2.08)Female‐only30‐dayNA* <40NA* 40–<60NA* 60–<80NA* ≥80NA* 1‐year0.028<40NA* 40–<6024.1 (2.96–196.95)60–<801.87 (1.22–2.85)≥801.88 (1.29–2.75)10‐year0.144<40NA* 40–<604.60 (1.95–10.9)60–<801.70 (1.20–2.40)≥801.55 (1.14–2.12)CI, confidence interval; HR, hazard ratio.*Estimates were not reported for subgroups with zero of very few outcome events within one or both arms, which resulted in non‐convergence of the Cox regression model. Risks of all‐cause mortality at 30 days, 1, 5, and 10 years in the overall cohort, and in males and females stratified by age CI, confidence interval; HR, hazard ratio. Estimates were not reported for subgroups with zero of very few outcome events within one or both arms, which resulted in non‐convergence of the Cox regression model. Kaplan–Meier curve comparing patients with and without stress hyperglycemia in the overall cohort. When stratified by sex, similar trends were observed in men, with significant interactions between age and stress hyperglycemia across all time points (P < 0.001). HRs likewise reduced with increasing age. In women, HR estimates for 30‐day mortality across age groups, and for 1‐year and 10‐year mortality in those aged <40 years could not be derived due to absent outcome events. Regardless, a similar pattern was observed for 1‐year mortality in females, although the interaction between age and stress hyperglycemia for 10‐year mortality was not statistically significant (P = 0.144). The Kaplan–Meier curves for males and females are presented in Figure 3a,b, respectively. (a) Kaplan–Meier curve comparing male patients with and without stress hyperglycemia. (b) Kaplan–Meier curve comparing female patients with and without stress hyperglycemia. DISCUSSION In this large retrospective cohort of 8,779 non‐diabetic patients with first‐onset STEMI, we found that patients with stress hyperglycemia were more frequently older, non‐smokers and hyperlipidemic than patients without. At presentation, compared to patients without, patients with stress hyperglycemia were more likely to be asymptomatic and suffer from more advanced heart failure. They also experienced higher rates of in‐hospital complications, including cardiogenic shock, left ventricular systolic dysfunction, arrhythmias, and stroke, compared with patients without stress hyperglycemia. In general, patients aged ≥40 years had approximately twice the odds of developing stress hyperglycemia compared with those aged <40 years. Sex did not appear to significantly modify the effect of age on the odds of stress hyperglycemia. In multivariate logistic regression, male sex, smoking and chest pain at presentation were associated with a lower likelihood of stress hyperglycemia. Stress hyperglycemia was consistently associated with reduced short‐ and long‐term survival across age groups, although its impact on mortality risk appeared to attenuate with increasing age. Our findings are consistent with prior studies showing that stress hyperglycemia in STEMI is associated with greater risks of complications and mortality. Large registry analyses and clinical cohorts have shown that elevated admission glucose is associated with higher short‐term and long‐term mortality after STEMI 4 , 23 , 24 , and meta‐analyses have confirmed a two‐ to three‐fold increase in death following myocardial infarction with admission hyperglycemia 5 , 25 . More evidence that is recent extends the prognostic relevance of SHR beyond mortality, linking elevated SHR to impaired coronary microcirculation, periprocedural type 4a myocardial infarction after PCI, and adverse angiographic features including high thrombus burden, distal embolization, and poor pre‐procedural TIMI flow, suggesting an association with both microvascular dysfunction and procedural ischemic risk 26 , 27 , 28 . However, despite the well‐established prognostic significance of stress hyperglycemia, to the authors' best knowledge, no previous studies have specifically examined the effect of age on its development. By quantifying age‐stratified odds in a large non‐diabetic STEMI cohort, our study provides new evidence that aging itself may be a potential determinant of stress hyperglycemia. Our findings are consistent with the hypothesis that aging is associated with greater stress glycemic responses to acute illness. This likely reflects cumulative metabolic vulnerability in older adults, arising from a multifactorial interplay of diminished insulin secretion, increased insulin resistance, and impaired counter‐regulatory responses involving catecholamines, cortisol, growth hormone, and inflammatory cytokines 10 , 29 . Sarcopenia further reduces glucose disposal capacity 30 , and vascular aging limits endothelial‐mediated metabolic regulation 31 . During acute stress in STEMI, these mechanisms may contribute to an exaggerated and less compensable hyperglycemic response 32 . The resulting metabolic and inflammatory cascade may contribute to myocardial injury through endothelial dysfunction, oxidative stress, thrombosis, and impaired reperfusion 33 , 34 . While modest glycemic elevations may reflect an adaptive stress response 33 , excessive hyperglycemia may reflect a maladaptive metabolic response associated with myocardial injury and adverse outcomes 4 , 5 , which may explain the observed association with mortality observed in our study. Elevated SHR was independently associated with reduced survival across all age and sex groups. The relative impact of stress hyperglycemia on mortality was greatest in younger patients, with HRs highest at younger ages and progressively lower with advancing age. Despite attenuation in relative risk, absolute mortality remained substantially higher in older patients, underscoring the importance of interpreting relative hazards alongside baseline risk. Similar patterns have been reported in previous studies, where admission hyperglycemia predicted mortality more strongly in younger than in older individuals 35 . The greater relative impact observed in younger patients may partly relate to their lower baseline mortality risk, such that a comparable absolute increase in risk translates into a larger relative effect. In contrast, among older patients with greater comorbidity burden, frailty, and competing risks of death, hyperglycemia may increasingly reflect underlying frailty and comorbidity, functioning more as a marker of overall vulnerability 10 than a specific contributor to adverse outcomes. Women in our cohort were more likely than men to develop stress hyperglycemia, consistent with prior observations that older women tend to present with higher admission glucose levels 36 . This susceptibility may reflect postmenopausal hormonal changes 37 , higher comorbidity burden, and delayed presentation 38 . These findings have several implications for clinical practice and research. First, stress hyperglycemia should be recognized as an important, age‐dependent marker of risk in STEMI. Prior studies confirm that SHR predicts adverse cardiovascular outcomes more reliably than admission glucose alone 39 , 40 , supporting its use for early risk stratification. Routine assessment of the SHR may help identify patients—particularly younger or middle‐aged individuals—at heightened risk of adverse outcomes who could benefit from closer hemodynamic and metabolic monitoring. Second, the management of stress hyperglycemia in AMI remains uncertain. Current guidelines do not distinguish stress hyperglycemia from pre‐existing diabetes, offering only broad inpatient glucose targets without clear differentiation by clinical context or illness severity 41 , 42 . Early trials such as the DIGAMI‐1 suggested a mortality benefit from insulin‐glucose infusion 43 , but subsequent studies including DIGAMI‐2 and NICE‐SUGAR failed to reproduce this finding and instead reported a higher incidence of hypoglycemia 44 , 45 . Consequently, current recommendations favor moderate rather than intensive glucose control to avoid iatrogenic hypoglycemia 1 , 46 . Third, management strategies should be individualized according to physiological context. The American Diabetes Association (ADA) and American Association of Clinical Endocrinologists (AACE) highlight that glucose targets in hospitalized patients should balance the risks of hyperglycemia and hypoglycemia rather than pursue strict normalization 47 , 48 . Younger patients, in whom hyperglycemia may act as a direct driver of myocardial injury, might benefit from closer metabolic surveillance and earlier intervention, whereas older or frail patients may require more permissive goals. This functional rather than chronological risk‐based approach aligns with the precision‐medicine paradigm increasingly advocated in acute cardiac care 49 . Beyond the acute phase, stress hyperglycemia may predict long‐term structural and functional consequences. It has been identified as an independent predictor of left ventricular remodeling after first anterior myocardial infarction in non‐diabetic patients 50 , underscoring the need for longitudinal follow‐up and targeted post‐discharge assessment for latent glucose intolerance. This study has several strengths. To our knowledge, it is the largest to examine SHR exclusively in non‐diabetic patients with first‐onset STEMI, incorporating age‐ and sex‐stratified analysis with long‐term survival outcomes. However, various limitations should be acknowledged. The retrospective design may introduce residual confounding, and the study's findings reflect associations rather than causation. A limited number of patients or events within certain subgroups may also reduce the reliability of our estimates. Hence, findings, particularly in such subgroups, should therefore be viewed as hypothesis‐generating rather than confirmatory. In addition, the SHR cutoff (≥1.51) used to define stress hyperglycemia in this study was validated using the same dataset in a previous analysis 16 . Although this approach was appropriate given that our study sought to characterize the utility of stress hyperglycemia as a prognostic entity rather than to re‐validate the cutoff, its generalizability to other AMI populations remains uncertain. Further validation of this threshold, and stress hyperglycemia's prognostic utility across distinct populations in future studies would strengthen its applicability in clinical practice. Of note, restriction to non‐diabetic patients improves internal validity but limits generalizability to the broader STEMI population. In diabetics, admission hyperglycemia reflects chronic dysglycemia and treatment effects, obscuring acute stress responses. Findings should therefore be interpreted in non‐diabetic STEMI, with future studies needed in diabetic cohorts. We also lacked information regarding frailty status among the older patients, which limited our ability to assess the influence of frailty on our findings. HbA1c, which is used in the calculation of SHR, allows admission glucose to be interpreted in relation to the apparent baseline glycemic status. However, HbA1c may be influenced by anemia 51 or other chronic conditions 52 . The use of SHR may also depend on the assumption of stable baseline glycemia. Hence, variations in baseline glycemic status or other factors affecting HbA1c could influence SHR estimates and the corresponding effect estimates. Furthermore, because SHR requires both admission glucose and HbA1c values, exclusion of patients with missing data for either component may have introduced selection bias if those excluded patients differed systematically from those included in the analytic cohort. Relatedly, as we lacked information on pre‐AMI fasting/random plasma glucose levels and HbA1c levels, defining non‐diabetic status as HbA1c <6.5% and the absence of a documented history or diagnosis of DM at presentation may not have fully excluded patients with undiagnosed DM or prediabetes, particularly given the potential influence of acute illness on presenting HbA1c values. Data on infarct size, inflammatory markers, and timing of glucose measurement for SHR calculation that is, in relation to symptom‐onset, arrival at the ED, or at PCI were also unavailable, limiting mechanistic inferences and potentially introducing residual confounding. CONCLUSION In non‐diabetic patients with first‐onset STEMI, stress hyperglycemia increased with age and was associated with reduced survival across age and sex groups. The relative association between elevated SHR and mortality was strongest in younger patients. These findings support SHR as a prognostic marker in STEMI and highlight the need for further studies to better understand age‐related differences in stress hyperglycemia and their clinical implications. FUNDING CHS was supported by the Singapore Ministry of Health National Medical Research Council's Transition Award (MOH‐001368‐00). This study is supported by the Cardiovascular Disease National Collaborative Enterprise (CADENCE) National Clinical Translational Program (MOH‐001277‐01). DISCLOSURE The authors declare no conflict of interest. Approval of the research protocol: N/A. Informed consent: The local institutional review board granted an exemption for written consent (2023/00263) from SMIR participants because our study utilized de‐identified data. Registry and the registration no. of the study/trial: N/A. Animal studies: N/A. AUTHOR CONTRIBUTIONS Conceptualization, Hon Jen Wong, Joy Xin Yi Au, Yao Hao Teo, Ching‐Hui Sia; Methodology, Hon Jen Wong, Joy Xin Yi Au, Yao Hao Teo, Yao Neng Teo, Chen Ee Low, Chun En Yau, Norman HY Lin; Investigation, Hon Jen Wong, Chen Ee Low, Chun En Yau, Norman HY Lin; Data curation, Hon Jen Wong, Chen Ee Low, Chun En Yau, Norman HY Lin; Formal analysis, Hon Jen Wong, Yao Hao Teo, Yao Neng Teo; Validation, Hon Jen Wong, Joy Xin Yi Au, Yao Hao Teo, Ching‐Hui Sia; Writing—original draft, Hon Jen Wong, Joy Xin Yi Au, Yao Hao Teo; Writing—review and editing, Hon Jen Wong, Joy Xin Yi Au, Yao Hao Teo, Yao Neng Teo, Chen Ee Low, Chun En Yau, Norman HY Lin, Pei Chia Eng, Li Feng Tan, Reshma A Merchant, Ping Chai, Anand Adinath Ambhore, James WL Yip, Andrew Fu‐Wah Ho, David Foo, Pow‐Li Chia, Patrick Zhan‐Yun Lim, Khung Keong Yeo, Weien Chow, Daniel Thuan Tee Chong, Derek J Hausenloy, Mark YY Chan, Ching‐Hui Sia; Project administration, Hon Jen Wong, Joy Xin Yi Au, Yao Hao Teo and Ching‐Hui Sia; Supervision, Ching‐Hui Sia. Supporting information Table S1. Complications during hospitalization. Table S2. Univariable and multivariable models for the odds of developing stress hyperglycemia during ST‐elevation myocardial infarction in males. Table S3. Univariable and multivariable models for the odds of developing stress hyperglycemia during ST‐elevation myocardial infarction in females. ACKNOWLEDGMENTS The authors thank the National Registry of Diseases Office, Health Promotion Board, Singapore for providing the data from the SMIR. DATA AVAILABILITY STATEMENT The data used in this study are property of NRDO intended for internal usage. The de‐identified data can be accessed for public health research purposes after approval from the Ministry of Health and Institutional Review Board. References Associated Data Supplementary Materials Table S1. Complications during hospitalization. Table S2. Univariable and multivariable models for the odds of developing stress hyperglycemia during ST‐elevation myocardial infarction in males. Table S3. Univariable and multivariable models for the odds of developing stress hyperglycemia during ST‐elevation myocardial infarction in females. Data Availability Statement The data used in this study are property of NRDO intended for internal usage. The de‐identified data can be accessed for public health research purposes after approval from the Ministry of Health and Institutional Review Board.