This retrospective cohort study was conducted at the Heart Hospital in Doha, Qatar, which is the main tertiary cardiology center within Hamad Medical Corporation (HMC), the country's principal public healthcare provider. The study was approved by the Medical Research Centre and Institutional Review Board at HMC (approval number: MRC-01-22-529). In the study, we included all patients with type 2 DM who were admitted to the Heart Hospital over 4 years between 1 June 2017 and 1 June 2021, with ACS complicated by new AHF defined as newly diagnosed HF with EF < 40% or clinical HF requiring diuretic therapy during the index admission regardless of the EF. Using a whole-population sampling approach, all patients with ischemia-induced new HF who met the eligibility criteria were included. The study population was then divided into two groups: (1) SGLT2 inhibitor users and (2) SGLT2 inhibitor non-users.

Patients were included in the study if they met the following criteria: (1) adult patients (≥18 years); diagnosed with type 2 DM; (2) admitted with ACS, including ST-elevation myocardial infarction (STEMI), non-ST-elevation myocardial infarction (NSTEMI), or unstable angina (UA) complicated by AHF; and (3) SGLT2 inhibitor naïve prior to the index admission. Patients were excluded if they had one of the following: (1) acute kidney injury (AKI) defined as a rise in serum creatinine by at least two times the baseline value, according to the KDIGO criteria (14), (2) chronic kidney disease (CKD) with creatinine clearance < 25 ml/min/1.73 m², (3) chronic HF for at least 6 months with EF < 40% prior to the current index admission of ACS, or (4) dispensed SGLT2 inhibitor for less than their follow-up period except for discontinuation due to an adverse drug reaction (ADR).

The primary outcomes included the composite of ACS, HF hospitalization, or all-cause mortality, while the secondary outcomes included ACS, HF hospitalization, all-cause mortality, stroke, and atrial fibrillation (AF). All outcomes were evaluated within 1 month and 12 months post-discharge. Patients were followed for 12 months post-discharge after the index admission, until they developed the primary and/or secondary endpoints within 12 months of discharge, or until censoring if they were lost to follow-up during the 12-month follow-up period. Loss to follow-up was defined as failure to attend confirmed outpatient clinic appointments and/or failure to refill active prescriptions on due dates.

The baseline characteristics, medical history, concurrent medications, and the outcomes of interest were collected from the HMC's electronic medical records system (i.e., Cerner). This was achieved by reviewing clinical care documentation during the index admission and subsequent admissions, outpatient clinic visits, and emergency visits to any HMC hospitals. Relevant diagnostic investigations performed during the follow-up period were also reviewed. Data collection was conducted from 1 June 2022 to 31 December 2022, and the relevant data were manually extracted using a pretested data collection form.

Descriptive statistics were reported in the form of frequencies and percentages for categorical variables, mean ± standard deviation (SD) for normally distributed continuous variables, and median with interquartile range (IQR) for skewed continuous variables. Pearson's chi-squared test was used to compare categorical variables between the two study groups (i.e., SGLT2 users vs. non-users), while Student's t-test and Mann–Whitney U-test were, respectively, applied to compare normally distributed continuous variables and skewed continuous variables between the groups. The primary analysis was designed to assess through the Cox proportional hazard regression model if early initiation of SGLT2 inhibitors in ACS complicated by AHF was associated with favorable clinical outcomes. Hazard ratios (HRs) with 95% confidence intervals (CIs) were computed and presented.

A propensity score-matched model (1:1) was used to adjust for differences in baseline characteristics between SGLT2 inhibitor users and non-users. The matching was done with a caliper of 0.1. A multivariate logistic regression model was used to obtain propensity scores, with the following variables included: age, gender, Asian ancestry, smoking history, type of ACS diagnosis upon admission, percutaneous coronary intervention (PCI) performance during the index admission, medical history, ejection fraction (EF), new mitral valve regurgitation during the index admission, HbA1c value, and concurrent medications upon discharge, including dual antiplatelet therapy, statin, beta-blocker, angiotensin-converting enzyme (ACE) inhibitor/angiotensin receptor blocker (ARB), sacubitril/valsartan, aldosterone receptor antagonist, and ivabradine. The absolute mean differences were calculated for all the variables after matching with a difference of <0.1 considered as minimal match imbalance. All p-values were two-sided with a p-value of <0.05 indicating statistical significance. Data analyses were performed using the Statistical Package for Social Sciences program version 28.0 (IBM SPSS Statistics for Windows; IBM Corp., Armonk, NY, USA).

A total of 2,588 subjects with type 2 DM who were admitted with ACS to the Heart Hospital between 1 June 2017 and 1 June 2021 were initially identified (Figure 1). A total of 465 patients with ACS-induced new-onset HF during the index admission were identified as eligible subjects, after excluding subjects with AKI, CKD with CrCl < 25 ml/min/1.73 m², and known HF with reduced EF and those who dispensed SGLT2 inhibitors for less than the follow-up period post-discharge. The 465 eligible subjects were divided into two groups: SGLT2 inhibitor users (n = 353) vs. SGLT2 inhibitor non-users (n = 112). Following a 1 : 1 propensity score matching, 78 subjects were included per arm.

The comparison between the study groups before and after propensity score matching is presented in Table 1. In the unmatched groups, there was a significantly higher proportion of males in the SGLT2 inhibitor user arm [335 (94.9%) vs. 96 (85.7%), p-value = 0.001], whereas the incidence of STEMI was higher in the control arm (i.e., SGLT2 inhibitor non-user arm) [92 (82.1%) vs. 212 (60.1%), p-value < 0.001]. Concurrent medical conditions, including hypertension, coronary artery disease, and AF, were similar between the two groups, except for CKD, which was higher in the SGLT2 inhibitor non-user group [11 (9.8%) vs. 12 (3.4%), p-value = 0.006]. Similarly, EF < 30% was more frequent among the SGLT2 inhibitor non-user group [28 (25%) vs. 45 (12.7%), p-value = 0.002]. SGLT2 inhibitor users had significantly higher baseline HbA1c levels [HbA1c > 9%: 227 (65%) vs. 33 (30%), p-value < 0.001]. All concurrent medications for ACS and HF were balanced between the two groups, except for ACE inhibitors/ARBs and aldosterone receptor antagonists, which were significantly more prescribed in the SGLT2 inhibitor user arm [302 (85.6%) vs. 69 (61.6%), p-value < 0.001, 76 (21.5%) vs. 9 (8%), p-value = 0.001, respectively]. Consequently, a propensity score-matched approach was applied to adjust for these differences in baseline characteristics between the study groups.

The 1 : 1 propensity score matching yielded balanced groups across all baseline variables with an absolute standardized difference (ASD) of <0.1, indicating minimal match imbalances, as shown in Table 1.

Prior to the propensity score matching, the 1-month primary composite outcome of ACS/HF/all-cause mortality was significantly lower among the SGLT2 inhibitor users compared to the non-users as shown in Table 2 [6 (1.7%) vs. 13 (11.6%), HR 0.13 (0.05–0.35), p-value < 0.001]. Similarly, the 12-month primary composite outcome maintained a similar trend with nearly twice the likelihood of the hazard in the SGLT2 inhibitor non-users [35 (9.9%) vs. 26 (23.2%), HR 0.31 (0.19–0.52), p-value < 0.001]. Furthermore, among the secondary outcomes at 1 month, HF-related hospitalization and stroke were significantly lower in the SGLT2 inhibitor users compared to non-users [1 (0.3%) vs. 9 (8%), p-value < 0.001; 1 (0.3%) vs. 3 (2.7%), p-value = 0.045, respectively]. At 12 months, the HF hospitalization and stroke were similarly significantly lower in the SGLT2 inhibitor user group, in addition to lower AF incidence [1 (0.3%) vs. 3 (2.7%), p-value = 0.045].

After the propensity score matching, the primary composite outcomes remained significantly lower with the use of SGLT2 inhibitors (1 month: 2 (2.6%) vs. 9 (11.5%), HR 0.20 (0.04–0.94), p-value = 0.041; 12 months: 11 (14.1%) vs. 18 (23.1%), HR 0.46 (0.22–0.99), p-value = 0.046, respectively], as presented in Table 2 and illustrated by Kaplan–Meier curves (Figures 2A,B). The significantly lower HF hospitalization rate persisted among SGLT2 inhibitor users at 1 month and 12 months following matching, as presented in Table 2.