Introduction

Acute myocardial infarction (AMI) is a leading cause of mortality and disability worldwide, with a particularly severe impact on low- and middle-income countries [1]. Standard pharmacological management following AMI is typically initiated within the first 24 hours and includes dual antiplatelet therapy, statins, beta-blockers, and inhibitors of the renin–angiotensin–aldosterone system [2]. Previous studies have demonstrated that sodium–glucose cotransporter-2 inhibitors (SGLT2i) exerted favorable effects on left ventricular ejection fraction and reduced N-terminal pro B-type natriuretic peptide (NT-proBNP) levels – two important prognostic markers in hospitalized patients with AMI. However, their benefits in reducing hard outcomes such as hospital readmission and mortality have not been established [3,4]. Based on these findings, we hypothesized that the use of SGLT2i might be associated with lower in-hospital mortality in patients following AMI. Therefore, the primary objective of this study was to evaluate the association between SGLT2i use and in-hospital prognosis in patients with AMI.

Methods

A retrospective study was conducted on 394 adult patients diagnosed with AMI from January 2022 to October 2024 at Nguyen Trai Hospital, Ho Chi Minh City, Vietnam. The data were accessed for research purposes from July 31 to October 24, 2024. The following cases were excluded: (1) inadequate confirmation of AMI diagnosis, (2) incomplete medical records, (3) secondary AMI related to procedures/surgery or type 2 MI, and (4) estimated glomerular filtration rate (eGFR) < 20 mL/min/1.73 m². Baseline characteristics of the study population are presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics.

https://doi.org/10.1371/journal.pone.0345315.t001

The diagnosis of AMI was based on medical records and verified according to the Fourth Universal Definition of Myocardial Infarction (2018) [5]. Therapeutic management was implemented following the current treatment guidelines of the Viet Nam Ministry of Health [6]. Diagnostic tests and administered medications, including SGLT2i, were recorded from the electronic hospital database, with the time of prescription calculated from the point of hospital admission. Patient prognosis was assessed based on in-hospital survival from the medical records.

Nguyen Trai Hospital is a public tertiary referral center in Ho Chi Minh City, Vietnam. The diagnosis and management of AMI follow the clinical guidelines issued by the Ministry of Health. Laboratory assessments, including troponin T and NT-proBNP, are conducted in the hospital’s ISO 15189–accredited laboratory.

All descriptive data are presented as mean ± SD (standard deviation), median, or frequency (%). Continuous variables between the two groups were compared using either the unpaired t-test or the Mann-Whitney U-test, as appropriate. Frequencies were compared using the Chi-squared test. The predictive potential was evaluated using univariable and multivariable logistic regression analyses. Multivariable logistic regression analysis included confounders that were significant in the univariable model. Both univariable and multivariable logistic regression analyses investigated 1-SD changes in continuous variables. NT-proBNP values were natural log–transformed prior to regression analyses. All analyses were performed using SPSS software, version 27.0.

The investigation complied with the principles outlined in the 1975 Declaration of Helsinki. All data were fully anonymized prior to analysis, with no identifiable personal information accessible to the investigators. The study protocol received approval from the institutional review board of Nguyen Trai Hospital (No. 1162/QĐ-BVNT, July 24, 2024), which waived the requirement for informed consent due to the retrospective nature of the study and the use of de-identified data.

Results

Clinical characteristics associated with in-hospital mortality

Univariable logistic regression analysis of clinical characteristics showed that age, left ventricular ejection fraction, shock, sepsis, serum creatinine, hemoglobin, and NT-proBNP were all significantly associated with in-hospital mortality. However, in the multivariable analysis, only left ventricular ejection fraction (OR 0.91, 95% CI: 0.85–0.97) and sepsis (OR 5.14, 95% CI: 1.04–25.36) remained independently associated with in-hospital mortality (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Univariable and multivariable logistic regression analyses of clinical characteristics associated with in-hospital mortality.

https://doi.org/10.1371/journal.pone.0345315.t002

Treatment factors associated with in-hospital outcomes included the use of BB/ACEi/ARB/MRA (OR 0.12, 95% CI: 0.05–0.25) and SGLT2i (OR 0.27, 95% CI: 0.07–0.96) (Table 3). In subgroup univariable analyses, both SGLT2i and heparin were associated with lower in-hospital mortality only among patients who did not undergo revascularization (S2 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Univariable and multivariable logistic regression analyses of treatment factors associated with in-hospital outcomes.

https://doi.org/10.1371/journal.pone.0345315.t003

After jointly adjusting for clinical characteristics and treatment variables, early SGLT2i use was consistently associated with favorable in-hospital prognosis across the different models evaluated (S3 Table).

Discussion

The analysis of 394 patients with AMI admitted to a tertiary hospital in Ho Chi Minh City, Vietnam, demonstrated that initiation of SGLT2i within 24 hours of admission was independently associated with lower in-hospital mortality. Additionally, reduced left ventricular ejection fraction and the presence of sepsis were associated with adverse prognosis, whereas guideline-directed medical therapy was linked to improved in-hospital survival.

Among comorbidities, diabetes mellitus was present in 30.5% of participants, with a higher proportion in the mortality group, although the difference was not statistically significant. Heart failure occurred more frequently in patients with adverse outcomes (26.4% vs. 14.7%, p = 0.03). Due to the retrospective nature of the study, it was not possible to distinguish whether heart failure was pre-existing or de novo in association with AMI. Large randomized controlled trials, such as DAPA-MI and EMPACT-MI, did not show statistically significant differences in heart failure hospitalization or all-cause mortality after AMI between the SGLT2i and control groups [7,8]. This difference may be explained by the later initiation of SGLT2i (mean 5–10 days) in the two aforementioned trials compared with our study (within 24 hours). Delayed initiation may miss the “window period” after myocardial infarction—a phase during which left ventricular injury and remodeling progress rapidly and are particularly sensitive to the effects of SGLT2 inhibition. Findings from a secondary analysis of the EMMY trial demonstrated that initiating empagliflozin within 24 hours after percutaneous coronary intervention resulted in a reduction in NT-proBNP comparable to that achieved with later initiation (24–72 hours), while also improving left ventricular functional and structural parameters without increasing the incidence of serious adverse events [9]. The exclusion of participants with common high-risk features in the AMI population, such as diabetes mellitus and heart failure, in the DAPA-MI trial may have reduced the ability to detect the true therapeutic benefit of the drug. A recent meta-analysis identified diabetes as a key effect modifier, demonstrating that for every 1% increase in the prevalence of diabetes in the study population, the all-cause mortality rate decreased by 0.49% [10]. This indicates that patients with diabetes mellitus are particularly sensitive to the cardiovascular benefits of SGLT2i, especially when treatment is initiated early.

It is noteworthy that a considerable proportion of patients in this study did not undergo coronary revascularization, a common scenario in developing countries, and represents a notable difference compared with strictly controlled clinical trials. Indeed, our subgroup analyses showed that the association of early SGLT2i use with in-hospital outcome was observed only in patients who did not undergo coronary revascularization. Failure to achieve coronary revascularization leads to a larger ischemic burden, elevated left ventricular filling pressures, and accelerated adverse ventricular remodeling, rapidly predisposing patients to acute heart failure and early mortality. In this high-risk setting, the favorable clinical associations observed with early SGLT2i use are consistent with rapid hemodynamic, metabolic, and cellular effects. SGLT2 inhibition improves ventricular loading conditions through osmotic diuresis and natriuresis, preferentially reducing interstitial congestion, lowering preload, and improving arterial stiffness and afterload without provoking excessive neurohormonal activation. At the same time, restoration of tubuloglomerular feedback lowers intraglomerular pressure and mitigates acute kidney injury, thereby stabilizing the cardiorenal axis during the vulnerable post-infarction phase [11,12]. Beyond hemodynamics, SGLT2 inhibition induces favorable myocardial metabolic remodeling by shifting substrate utilization away from inefficient glucose metabolism toward ketone bodies and fatty acid oxidation, thereby enhancing myocardial ATP production and mechanical efficiency. Experimental models of post-AMI heart failure demonstrate that empagliflozin reduces LV dilatation, improves systolic function, and attenuates adverse remodeling despite similar infarct size at baseline [13]. At the cellular level, SGLT2 inhibition activates a fasting-like transcriptional program via AMPK and SIRT1 while suppressing Akt/mTOR signaling, promoting autophagy, reducing oxidative stress and inflammation, and preserving mitochondrial function in both the heart and kidney [14]. These coordinated effects provide a plausible biological explanation for the early favorable clinical associations observed in patients with AMI, particularly among those who do not undergo coronary revascularization.

This study was conducted at a high-volume tertiary cardiovascular center that admits a substantial number of patients with AMI, thereby partially reflecting the real-world diagnosis and management practices in tertiary hospitals in Ho Chi Minh City. Data were obtained from electronic medical records with high completeness and detail, ensuring reliability in the analysis of both clinical characteristics and in-hospital treatment. However, several limitations should be acknowledged. As the study was conducted at a single center, its generalizability to the overall population of patients with AMI in Vietnam may be limited. Additionally, the retrospective design carries inherent constraints in data collection, particularly regarding the rationale for early SGLT2i initiation (whether related to diabetes mellitus, heart failure, or chronic kidney disease). Thus, the decision to initiate SGLT2i might not be random and influenced by baseline heart failure severity, as reflected by the lower LVEF in the SGLT2i group. These limitations underscore the need for larger, multicenter, prospective cohort studies to further elucidate the prognostic factors and therapeutic strategies, including early SGLT2 inhibition, that may be associated with improved outcomes in AMI within the Vietnamese healthcare setting.

Conclusions

Early initiation of SGLT2i within 24 hours of hospital admission was independently associated with lower in-hospital mortality in patients with AMI. This association was most pronounced in high-risk patients, particularly those who did not undergo coronary revascularization. These findings suggest a potential association between early SGLT2i use and in-hospital outcomes, which warrants confirmation in prospective randomized trials.

Supporting information

Acknowledgments

The authors gratefully acknowledge Minh Van Le and Thien Ngoc Nguyen for their assistance in sample collection.