Abstract
Introduction & Objectives
Heart failure is one of the major public health concerns and a leading cause of hospitalization and mortality worldwide, including in Thailand. Sodium-glucose cotransporter 2 inhibitors (SGLT2i) have demonstrated significant cardiovascular benefits in clinical trials. This study aimed to evaluate the real-world effectiveness of SGLT2i in reducing all-cause mortality and heart failure hospitalization among patients with heart failure at Naresuan University Hospital.
Materials & Methods
This retrospective cohort study included patients newly diagnosed with heart failure at Naresuan University Hospital between January 1, 2019, and December 31, 2023. Patients were divided into those who received SGLT2i and those who did not. The primary outcome was a composite of all-cause mortality and hospitalization for worsening heart failure. Rate ratios were calculated using multilevel mixed-effects Poisson regression, adjusting for baseline characteristics.
Results
A total of 1,378 patients were included (1,080 SGLT2i visits; 5,243 non-SGLT2i visits). The proportion of patients with reduced left ventricular ejection fraction (LVEF ≤ 40%) was higher in the SGLT2i group than the non-SGLT2i group (39.9% vs. 29.4%, p < 0.001). The SGLT2i group also had a higher prevalence of coronary artery disease, myopathy, and chronic kidney disease. Rate of composite outcome (death or heart failure hospitalization) was lower in the SGLT2i group (rate 10.93 per 100 vs. 17.58 per 100). The incidence of the composite outcome was significantly lower in the SGLT2i group compared to the non-SGLT2i group (rate ratio 0.60, 95% CI: 0.42–0.87; p = 0.006). The all-cause mortality rate in the SGLT2i group was markedly lower (rate ratio 0.02, 95% CI: 0.00–0.18; p = 0.001), while heart failure hospitalization showed a favorable trend without reaching statistical significance.
Conclusion
SGLT2i were associated with a significantly lower risk of death and heart failure hospitalization in this real-world cohort of Thai heart failure patients. These findings reinforce the clinical benefits of SGLT2i and support their broader implementation in heart failure management in Thailand.
Keywords: Heart failure, SGLT2 inhibitors, Mortality, Hospitalization, Real-world study
1. Introduction
Heart failure (HF) represents a major global public health burden, affecting over 64 million people worldwide. Being a leading cause of hospital admission, morbidity, and mortality, HF imposes a substantial impact on healthcare systems. In Thailand, the Thai ADHERE Registry reported an in-hospital mortality rate of 5.5% and an average length of hospital stay of 7.5 days among patients with acute decompensated HF [1]. Long-term prognosis also remains poor, with approximately 50% of patients dying within five years of diagnosis [2].
Sodium-glucose cotransporter-2 inhibitors (SGLT2i), originally developed for glycemic control in type 2 diabetes, have emerged as a mainstay therapy in HF management. Landmark randomized controlled trials, including DAPA-HF, EMPEROR-Reduced, EMPEROR-Preserved, and DELIVER, have demonstrated that SGLT2i significantly lower the risk of HF hospitalization, cardiovascular mortality, and all-cause mortality across both preserved and reduced left ventricular ejection fraction (LVEF), and in both diabetic and non-diabetic populations [3], [4], [5], [6], [7]. A comprehensive meta-analysis by Vaduganathan et al. confirmed the consistent benefits of SGLT2i observed in these major trials [8].
As a result, major international guidelines endorse SGLT2i as a cornerstone treatment for HF. Both the 2021 and 2023 ESC guidelines, as well as the 2022 ACC/AHA/HFSA guideline, support a class I recommendation for SGLT2i in patients with HF regardless of LVEF [2], [9], [10]. Reflecting these developments, the Heart Failure Council of Thailand (HFCT) released focused updates to its 2019 national guideline: the 2022 HFCT update for HFrEF management and the 2023 update for HFmrEF and HFpEF, both of which strongly recommend SGLT2i for reducing HF hospitalization and cardiovascular death [11], [12].
Despite these endorsements, real-world data on the use and effectiveness of SGLT2i in Thai HF patients remain limited. Additionally, access to these agents is restricted by cost, particularly under the Universal Coverage Scheme. Nonetheless, several Thai health economic evaluations have suggested that SGLT2i are cost-effective in the treatment of HF, especially in patients with reduced or mildly reduced ejection fraction [13], [14], [15].
This study aimed to evaluate the clinical effectiveness of SGLT2i in reducing all-cause mortality and hospitalization for worsening HF in a real-world population of patients treated at Naresuan University Hospital over a five-year period. The findings may provide additional evidence to inform clinical practice and support expanded access to SGLT2i under Thailand’s national health policies.
2. Methods
2.1. Study design and setting
Our retrospective cohort study was conducted at Naresuan University Hospital, a tertiary-care academic center in Phitsanulok, Thailand. Adult patients newly diagnosed with HF between January 2019 and December 2023 were included. The study protocol was approved by the Institutional Review Board and the Ethics Committee of Naresuan University (IRB No. P3-00772567).
2.2. Sample size calculation
The sample size was estimated using Stata version 18.0 based on a survival analysis framework. The calculation assumed the following parameters: two-sided alpha level of 0.05, 80% power, hazard ratio of 0.75, and 1:1 ratio of SGLT2i to non-SGLT2i exposure. The expected event incidence in the control group was 33% at 27 months, with an average follow-up duration of 30 months. Based on these assumptions, the minimum required sample size was estimated to be approximately 1,150 patients, corresponding to an expected total of 380 outcome events.
2.3. Study population
Eligible patients were adults aged 18 years or older with a documented diagnosis of HF established through clinical evaluation by board-certified internists. Patients with incomplete medical records or follow-up data were excluded. Each patient could contribute multiple follow-up visits, and the prescription of SGLT2i was evaluated at each visit.
2.4. Exposure
The exposure of interestwas the prescription of any SGLT2i (dapagliflozin or empagliflozin) during the follow-up period. Each outpatient visit was classified as either SGLT2i or non-SGLT2i group, based on whether the patient was receiving an SGLT2i at that time. Patients who switched groups during the study contributed person-time to both the SGLT2i and non-SGLT2i exposure periods accordingly. SGLT2i exposure was treated as a time-varying covariate. A patient’s exposure status could change over time, depending on prescription status at each visit. Each follow-up interval was classified according to the exposure at that visit, allowing for within-patient transitions from unexposed to exposed or vice versa.
2.5. Outcomes
The primary outcome was a composite of all-cause mortality and hospitalization for worsening HF. These events were identified through hospital admission records and death certificates, and confirmed by independent cardiologists who were blinded to SGLT2i exposure. HF hospitalization was defined as an admission with primary discharge diagnosis of HF (ICD-10 code I50.x).
The secondary outcomes included the individual components of the primary composite outcome: all-cause mortality and HF hospitalization, each evaluated independently. Time-to-event data and incidence rates for each secondary outcome were analyzed separately to assess the consistency and relative contribution of each endpoint to the composite primary outcome.
2.6. Statistical analysis
Descriptive statistics were used to summarize baseline characteristics, stratified by SGLT2i use. Categorical variables were presented as counts and percentages, and continuous variables as means with standard deviations or medians with interquartile ranges, as appropriate.
A multilevel mixed-effects Poisson regression model was employed to estimate rate ratios (RRs) and 95% confidence intervals (CIs) of outcomes associated with SGLT2i use, accounting for repeated measures within patients. The model was adjusted for baseline covariates including age, sex, payment type (civil service medical benefits scheme (CSMBS), social security scheme (SSS), universal coverage scheme (UCS) and self-payment), left ventricular ejection fraction (LVEF), comorbidities (diabetes, chronic kidney disease, atrial fibrillation), and concomitant use of ARNI.
An inverse probability of treatment weighting (IPTW) analysis was conducted to account for baseline imbalances, with each patient weighted according to the inverse probability of receiving their assigned treatment. We considered performing a sensitivity analysis by excluding patients with follow-up less than 3 months. However, this was not feasible within the scope of the current dataset. Approximately 10% of patients had short follow-up, which may have introduced bias in the observed outcomes. This limitation is acknowledged in the Discussion.
All analyses were performed using Stata version 18.0 (StataCorp LLC, College Station, TX, USA).
3. Results
3.1. Baseline characteristics
Patients who were newly diagnosed with HF and registry at Naresuan University Hospital from January 2019 through December 2023 were recruited. The baseline characteristics of patients are summarized in Table 1. Among a total of 6,323 outpatient visits (from 1,378 patients), 1,080 visits (17.1%) occurred while patients were receiving an SGLT2i (SGLT2i group), while 5,243 visits (82.9%) occurred while patients were not receiving an SGLT2i (non-SGLT2i group).Table 2..
Table 1.
Baseline characteristics comparing the SGLT2i visits and non-SGLT2i visits.
| Baseline characteristics | Non-SGLT2i (n = 5,243) |
SGLT2i (n = 1,080) |
p-value |
|---|---|---|---|
| Age – no (%) | <0.001 | ||
|
202 (3.9) | 16 (1.5) | |
|
991 (18.9) | 182 (16.9) | |
|
2800 (53.4) | 701 (64.9) | |
|
1250 (23.8) | 181 (16.8) | |
| Male – no (%) | 2496 (47.6) | 621 (57.5) | <0.001 |
| BMI (kg/m2) − Mean ± SD | 24.73 ± 6.1 | 25.29 ± 5.5 | 0.0698 |
| Payment type – no (%) | <0.001 | ||
|
2167 (41.3) | 872 (80.7) | |
|
2936 (56.0) | 185 (17.1) | |
|
102 (2.0) | 18 (1.7) | |
|
37 (0.7) | 5 (0.5) | |
|
1 (0.0) | 0 (0.0) | |
| NT-proBNP – no. (%) | <0.001 | ||
|
27 (8.4) | 13 (26.5) | |
|
295 (91.6) | 49 (73.5) | |
| LVEF – no. (%) | <0.001 | ||
|
819 (15.6) | 80 (7.4) | |
|
1540 (29.4) | 431 (39.9) | |
|
559 (10.7) | 153 (14.2) | |
|
2325 (44.3) | 416 (38.5) | |
| eGFR (ml/min/1.73 mm2) – Mean ± SD | 58 ± 24 | 59 ± 22 | 0.3625 |
| HbA1C (%) − Median (IQR) | 6.3 (1.7) | 6.2 (1.4) | 0.5955 |
| Hb (g/dL) − Mean ± SD | 11.15 ± 2.1 | 12.28 ± 2.1 | <0.001 |
| Hct (%) − Mean ± SD | 34.07 ± 6.2 | 37.52 ± 6.1 | <0.001 |
| Underlying disease − no (%) | |||
|
1544 (29.5) | 319 (29.5) | 0.954 |
|
2782 (53.1) | 589 (54.5) | 0.376 |
|
2269 (43.3) | 483 (44.7) | 0.383 |
|
1023 (19.5) | 203 (18.8) | 0.588 |
|
1040 (19.8) | 161 (14.9) | <0.001 |
|
619 (11.8) | 172 (15.9) | <0.001 |
|
1837 (35.0) | 544 (50.4) | <0.001 |
|
663 (12.7) | 165 (15.3) | 0.020 |
|
259 (4.9) | 34 (3.2) | 0.011 |
| Number of heart failure event before − Median (IQR) | 2 (3) | 4(5) | <0.001 |
| Medications − no (%) | |||
|
1350 (25.8) | 144 (13.3) | <0.001 |
|
951 (18.1) | 205 (19.0) | 0.514 |
|
276 (5.3) | 414 (38.3) | <0.001 |
|
1146 (21.9) | 189 (17.5) | 0.001 |
|
2229 (42.51) | 537 (49.7) | <0.001 |
| ICD or CRT − no (%) | 792 (15.1) | 152 (14.7) | 0.386 |
Table 2.
Rate of outcomes (per 100 person-year) in all visits, SGLT2 visits, and non-SGLT2 visits.
| Outcome | All | Non-SGLT2i | SGLT2i |
|---|---|---|---|
| Primary outcome | |||
| Composite of all-cause mortality, and hospitalization for worsening heart failure | 16.73 (15.47–18.08) |
17.58 (16.20–19.07) |
10.93 (8.35–14.31) |
| Secondary outcome | |||
| Single component: Death from any cause | 3.30 (2.77–3.94) |
3.76 (3.15–4.48) |
0.21 (0.03–1.47) |
| Single component: Hospitalization for worsening heart failure | 13.42 (12.31–14.64) |
13.82 (12.61–15.15) |
10.73 (8.17–14.08) |
Patients in the SGLT2i group were more likely to be male (57.5% vs. 47.6%, p < 0.001) and had higher representation in the 60–79-year age group (64.9% vs. 53.4%). The SGLT2i group had a significantly greater proportion of patients enrolled under the civil servant medical benefit scheme (CSMBS) (80.7% vs. 41.3%), whereas the majority of the non-SGLT2i group were covered by the universal coverage scheme (56.0%, p < 0.001).
The proportion of patients with reduced left ventricular ejection fraction (LVEF ≤ 40%) was higher in the SGLT2i group than the non-SGLT2i group (39.9% vs. 29.4%, p < 0.001). The SGLT2i group also had a higher prevalence of coronary artery disease (50.4% vs. 35.0%, p < 0.001), cardiomyopathy (15.9% vs. 11.8%, p < 0.001), and chronic kidney disease (15.3% vs. 12.7%, p = 0.020). Hemoglobin (12.28 vs. 11.15 g/dL, p < 0.001) and hematocrit levels (37.52% vs. 34.07%, p < 0.001) were significantly higher in the SGLT2i group.
With respect to HF treatments, the SGLT2i group more frequently received sacubitril/valsartan (38.3% vs. 5.3%, p < 0.001) and mineralocorticoid receptor antagonists (49.7% vs. 42.5%, p < 0.001), while the non-SGLT2i group more often received ACE inhibitors (25.8% vs. 13.3%, p < 0.001) and beta-blockers (21.9% vs. 17.5%, p = 0.001). The median number of prior HF events was also higher in the SGLT2i group (4 vs. 2 events, p < 0.001).
3.2. Rate
The rate of primary composite outcome (all-cause mortality and hospitalization for worsening HF) was lower in the SGLT2i group than non-SGLT2i group (rate 10.93 per 100 vs. 17.58 per 100). Consistent with the primary endpoint, the SGLT2i group demonstrated lower rates of secondary outcomes, including all-cause mortality (rate 0.21 per 100 vs. 3.76 per 100) and HF hospitalization (rate 10.73 per 100 vs. 13.82 per 100), compared with the non-SGLT2i group.
3.3. Primary outcome
During the study period, the composite outcome of all-cause mortality and hospitalization for worsening HF occurred in 53 patients (1 death, 52 HF admissions) in the SGLT2i group and 580 patients (124 deaths, 456 HF admissions) in the non-SGLT2i group.
In the multivariate analysis adjusted for baseline characteristics, SGLT2i use was associated with a significantly lower risk of the composite outcome (adjusted RR 0.60, 95% CI: 0.42–0.87; p = 0.006) (Table 3A). Meanwhile, IPTW analysis showed a trend towards lower risk of the composite outcome in the SGLT2i group compared with non-SGLT2i group, but the difference was not statistically significant (adjusted RR 0.90, 95% CI: 0.55–1.49; p = 0.688) (Table 3B).Table 4..
Table 3A.
Rate ratios of outcomes in the SGLT2 visits and non-SGLT2 visits.
| Variable | Crude ratio |
Adjusted rate ratio* |
||
|---|---|---|---|---|
| SGLT2i | Non-SGLT2i | SGLT2i | Non-SGLT2i | |
| Primary outcome | ||||
| Composite of all-cause mortality, and hospitalization for worsening heart failure | ||||
| Total event (no.) | 53 | 580 | ||
| Rate ratio (95% CI) | 0.58 (0.41–0.81) | Ref | 0.60 (0.42–0.87) | Ref |
| P value | 0.001 | 0.006 | ||
| Secondary outcome | ||||
| Single component: Death from any cause | ||||
| Total event (no.) | 1 | 124 | ||
| Rate ratio (95% CI) | 0.02 (0.00–0.18) | Ref | 0.02 (0.00–0.18) | Ref |
| P value | 0.001 | 0.001 | ||
| Single component: Hospitalization for worsening heart failure | ||||
| Total event (no.) | 52 | 456 | ||
| Rate ratio (95% CI) | 0.78 (0.55–1.09) | Ref | 0.83 (0.58–1.20) | Ref |
| P value | 0.138 | 0.327 | ||
*Adjusted by multivariable analysis, including sex, age, payment type, underlying diseases (valvular disease, myopathy, coronary artery disease, chronic kidney disease, stroke or TIA), heart failure treatment (ACE inhibitor, ARB, BB, MRA, Diuretic), LVEF.
Table 3B.
Rate ratio of outcomes between the SGLT2 visits and non-SGLT2 visits.
| Variable | Crude ratio |
Adjusted rate ratio* |
||
|---|---|---|---|---|
| SGLT2i | Non-SGLT2i | SGLT2i | Non-SGLT2i | |
| Crude rate | Adjusted rate | |||
| Primary outcome | ||||
| Composite of all-cause mortality, and hospitalization for worsening heart failure | ||||
| Total event/rate | 53/0.109 | 580/0.176 | ||
| Rate ratio (95% CI) | 0.62 (0.46–0.82) | Ref | 0.90(0.55–1.49) | Ref |
| P value | 0.001 | 0.688 | ||
| Secondary outcome | ||||
| Single component: Death from any cause | ||||
| Total event (no.) | 1/0.002 | 124/0.037 | ||
| Rate ratio (95% CI) | 0.055 (0.001–0.311) | Ref | 0.003 (0.000–0.035) | Ref |
| P value | <0.001 | <0.001 | ||
| Single component: Hospitalization for worsening heart failure | ||||
| Total event/rate | 52/0.107 | 456/0.138 | ||
| Rate ratio (95% CI) | 0.78 (0.57–1.04) | Ref | 1.10 (0.67–1.83) | Ref |
| P value | 0.077 | 0.693 | ||
Data was analyzed using mixed effect Poisson regression to estimate crude rates and rates adjusted by inverse probability of treatment weighting. *Adjusted rate ratio with inverse probability of treatment weighting.
Table 4.
Subgroup analysis according to LVEF, comparing outcomes between the SGLT2i and non-SGLT2i groups.
| Subgroup | All | Non-SGLT2i | SGLT2i |
|---|---|---|---|
| Primary outcome | |||
| Composite of all-cause mortality, and hospitalization for worsening heart failure | |||
| Overall | 16.73 (15.47–18.08) |
17.58 (16.20–19.07) |
10.93 (8.35–14.31) |
| HFrEF (LVEF < 40%) | 28.99 (25.54–32.90) |
29.14 (25.43–33.39) |
28.09 (19.97–39.51) |
| HFmrEF (LVEF 40–49%) | 14.72 (11.48–18.88) |
16.76 (13.02–21.59) |
3.16 (0.79–12.65) |
| HFpEF (LVEF ≥ 50%) | 15.06 (13.38–16.95) |
16.45 (14.57–18.56) |
5.20 (1.68–16.12) |
| Secondary outcome | |||
| Single component: Death from any cause | |||
| Overall | 3.30 (2.77–3.94) |
3.76 (3.15–4.48) |
0.21 (0.03–1.47) |
| HFrEF (LVEF < 40%) | 4.44 (2.69–4.40) |
4.65 (3.30–6.53) |
0.00 |
| HFmrEF (LVEF 40–49%) | 4.75 (3.06–7.36) |
5.59 (3.61–8.66) |
0.00 |
| HFpEF (LVEF ≥ 50%) | 3.99 (2.83–5.61) |
3.91 (3.05–5.01) |
0.41 (0.06–2.88) |
| Single component: Hospitalization for worsening heart failure | |||
| Overall | 13.42 (12.31–14.64) |
13.82 (12.61–15.15) |
10.73 (8.17–14.08) |
| HFrEF (LVEF < 40%) | 25.00 (21.82–28.65) |
24.49 (21.11–28.42) |
28.09 (19.97–39.51) |
| HFmrEF (LVEF 40–49%) | 9.97 (7.37–13.49) |
11.18 (8.20–15.24) |
3.16 (0.79–12.65) |
| HFpEF (LVEF ≥ 50%) | 11.62 (10.16–13.29) |
12.54 (10.92–14.41) |
5.68 (3.37–9.60) |
3.4. Secondary outcomes
When each component of the composite outcome was analyzed separately:
All-cause mortality occurred in 1 patient in the SGLT2i group versus 124 patients in the non-SGLT2i group. Multivariate analysis showed a marked reduction in mortality in the SGLT2i group (adjusted RR 0.02, 95% CI: 0.00–0.18; p = 0.001) (Fig. 1).
Fig. 1.
Kaplan–Meier survival curve for all-cause mortality.
Hospitalization for worsening HF occurred in 52 patients in the SGLT2i group and 456 patients in the non-SGLT2i group. Although there was a favorable trend toward reduced HF hospitalization in the SGLT2i group, the difference did not reach statistical significance (adjusted RR 0.83, 95% CI: 0.58–1.20; p = 0.327) (Fig. 2).
Fig. 2.
Kaplan–Meier survival curve for hospitalization due to worsening heart failure.
These findings suggest that the benefit of SGLT2i was primarily driven by a significant reduction in mortality rather than hospitalization for worsening HF.
4. Discussion
Our five-year retrospective cohort study provides real-world evidence on the effectiveness of sodium-glucose cotransporter-2 inhibitors (SGLT2i) in Thai patients with heart failure (HF). SGLT2i use was associated with a significantly lower risk of the composite outcome of all-cause mortality and HF hospitalization (adjusted RR 0.60, 95% CI: 0.42–0.87; p = 0.006). The effect was primarily driven by a substantial reduction in mortality (adjusted RR 0.02, 95% CI: 0.00–0.18), while the trend toward lower HF hospitalization did not reach statistical significance. The very low number of deaths observed in the SGLT2i group may be underreported, as the data were obtained only from hospital records without linkage to national registration.
Our results are in agreement with key randomized controlled trials (RCTs) such as DAPA-HF (3) and EMPEROR-Reduced (5) for HFrEF, as well as EMPEROR-Preserved (6) and DELIVER (7) for HFmrEF and HFpEF. These RCTs collectively established SGLT2i as one of the few therapies demonstrating consistent benefit across the full spectrum of LVEF. The meta-analysis by Vaduganathan et al. further emphasized this, showing that SGLT2i reduced HF hospitalizations and cardiovascular death in both diabetic and non-diabetic patients(8).
The lack of significant reduction in HF hospitalization may be influenced by several factors. Some possible explanations include variation in care practice, delayed treatment initiation, or underreporting of hospitalization events. Moreover, baseline imbalances, such as more frequent use of sacubitril/valsartan and mineralocorticoid receptor antagonists in the SGLT2i group, may confound the treatment effect, despite adjustment in the multivariable analysis [16]. Lastly, non-standardized criteria for HF hospitalization in electronic health records may limit internal validity.
Mechanistically, SGLT2i improve HF outcomes through osmotic diuresis, plasma volume contraction, increased erythropoietin production, and ketone-based myocardial energetics. These mechanisms reduce left ventricular filling pressure, improve cardiac efficiency, and limit fibrosis and oxidative stress—benefits relevant in both HFrEF and HFpEF [17], [18], [19], [20], [21].
To further understand the clinical utility of SGLT-2i in different HF phenotypes, we conducted a subgroup analysis stratified by LVEF. In patients with HFpEF and HFmrEF, SGLT2i use was associated with a lower rate of the composite outcome compared to non-SGLT-2 group. These findings are in line with the EMPEROR-Preserved(6) and DELIVER trials(7), which demonstrated benefit of SGLT2i in reducing HF hospitalization among patients with HFpEF and HFmrEF. However, the result of subgroup analysis was imprecise with a wide confidence interval and zero mortality in some groups. These results should be further explored rather than definitive.
In contrast, patients with HFrEF in our cohort had similar event rates between the two groups. This contrasts with the DAPA-HF(3) and EMPEROR-Reduced trials (5), where SGLT2i significantly reduced both mortality and hospitalization in HFrEF patients. One possible explanation is that real-world patients with HFrEF in our study had a higher burden of comorbidities and prior HF events, which may have contributed to persistently high event rates despite SGLT2i therapy. Additionally, small subgroup sizes may have limited the precision of the estimates, as reflected by the wide confidence intervals.
For the secondary outcomes, mortality in the SGLT2i group was extremely low across all LVEF subgroups, with zero events in both HFrEF and HFmrEF, and a single death in a patient with HFpEF. Although this suggests a possible survival benefit across LVEF phenotypes, the low number of deaths limits definite conclusions. Similarly, HF hospitalization rates were numerically lower in the SGLT2i group for HFmrEF and HFpEF, but not for HFrEF, where the rate was paradoxically higher than in the control group. This may reflect confounding by indication, as patients with more severe disease may have been preferentially initiated on SGLT2i in clinical practice.
Importantly, our findings support current international and national treatment guidelines. The 2021 and 2023 ESC guidelines[2], [10] and the 2022 ACC/AHA/HFSA guideline[8] grant SGLT2i a class I recommendation across LVEF ranges. In Thailand, the Heart Failure Council of Thailand (HFCT) issued focused updates in 2022 and 2023 that similarly endorse SGLT2i as part of standard therapy[11], [12]. Our results provide local real-world data supporting these guidelines, particularly for underrepresented patient groups in prior trials.
Comparisons with the Kongmalai et al. real-world Thai registry study also reaffirm the clinical benefit of SGLT2i in routine care settings[22]. Notably, their study included only patients with diabetes, while our study expands this benefit to a broader HF population regardless of diabetic status.
The applicability of SGLT2i in patients with multiple comorbidities further strengthens their role in real-world practice. In our study, many patients in the SGLT2i group had concurrent conditions such as chronic kidney disease (CKD) and atrial fibrillation (AF)-comorbidities frequently seen in advanced HF populations. Prior trials, such as DAPA-CKD[23] and EMPA-KIDNEY[24], have shown that SGLT2i reduce renal function decline and cardiovascular events even in non-diabetic individuals.
In addition, several real-world registries have demonstrated that SGLT2i retain effectiveness across subgroups traditionally underrepresented in RCTs, including elderly patients, those with frailty, or advanced NYHA class. These findings, combined with our data, encourage broader clinical use in routine outpatient practice.
Another crucial consideration is the observed benefit among non-diabetic HF patients, confirming the glucose-independent mechanisms of SGLT2i. This reinforces their therapeutic value even in patients with normoglycemia, a population where other glucose-lowering agents provide no benefit or may even increase risk[25].
However, treatment adherence in real-world settings is often suboptimal. Barriers such as polypharmacy, concern over genital infections, lack of provider familiarity, and cost may lead to early discontinuation or underutilization. Implementation science approaches are necessary to address these gaps, especially in resource-limited settings like Thailand.
The divergence between outcomes seen in RCTs and real-world cohorts, including ours, highlights important nuances. RCTs involve closely monitored patients with protocol-defined follow-up, while real-world data capture clinical complexity, socioeconomic disparities, and systemic healthcare limitations. As such, real-world studies complement RCTs, providing external validity and insight into long-term utility and tolerability.
Lastly, the emergence of SGLT2i represents a broader shift toward precision medicine in HF. Their benefit across LVEF, glycemic status, renal function, and even body composition suggests a class of drugs that transcends traditional HF classifications. As more biomarkers and phenotypic profiling tools become available, clinicians will be better able to match patients with therapies like SGLT2i based on individual response predictors.
4.1. Strengths and limitations
A major strength of this study is the use of a large, real-world dataset from a tertiary university hospital in Thailand, which reflects clinical practice under varied reimbursement schemes. The longitudinal nature and multilevel modeling approach allowed for adjustment of repeated measures over time and incorporation of multiple covariates. Importantly, SGLT2i exposure was modeled as a time-varying covariate, allowing each patient’s exposure status to change over time depending on prescription at each visit, which better reflects real-world treatment dynamics.
However, several limitations should be acknowledged. First, the retrospective design limits causal inference, and unmeasured confounders may remain despite multivariate adjustment. Residual confounding may persist, particularly due to differences in background therapies such as ARNI and MRA, which were more frequently used among SGLT2i users. Second, mortality data may have been underestimated due to loss to follow-up or incomplete death reporting in the electronic medical record. Furthermore, patients in the non-SGLT2i group were more often covered by the universal coverage health scheme, resided predominantly within the catchment area of the hospital, and were less likely to afford and seek care at alternative healthcare facilities. Therefore, deaths in the non-SGLT2i group were more frequently documented in the study hospital, which may explain why mortality appeared higher than the SGLT2i group. Nevertheless, the difference in hospital admission rates between the SGLT2i and non-SGLT2i groups was modest (10.7 versus 13.8, respectively). However, we acknowledge that this unconventional approach for mortality outcomes may introduce misclassification bias, particularly if effects persist beyond a single visit. Due to limitations of our database, standardized diagnostic criteria were not available. Mortality was identified only from our hospital records, without linkage to the national death registration. The very low number of deaths among SGLT2i users may also reflect under-ascertainment, particularly for out-of-hospital deaths. Future studies incorporating national death registry linkage would help ensure more complete mortality ascertainment.
Third, approximately 10% of patients had less than three months of follow-up, which may have reduced the opportunity to capture mortality events or recurrent hospitalizations. Early censoring in these patients could bias the results toward underestimating adverse outcomes. We were unable to exclude them in a sensitivity analysis due to analytic constraints, and this limitation should be considered when interpreting the findings. Lastly, data were collected from a single center, which may limit the generalizability of our findings to other regions or healthcare systems in Thailand.
4.2. Future Directions
Future research should include multicenter, prospective registries to confirm these findings and to explore subgroup effects by sex and comorbidities. Linkage with national death registries and standardized definitions of heart failure hospitalization would enhance data completeness and comparability across studies. Cost-effectiveness analyses and implementation studies are also needed to support broader access to SGLT2i under national reimbursement programs.
5. Conclusion
This real-world, retrospective cohort study demonstrated that the use of SGLT2i in Thai patients with HF was associated with a significantly lower risk of all-cause mortality and a trend toward reduced HF hospitalization. These findings are consistent with results from major clinical trials and support current guideline recommendations that advocate SGLT2i as foundational therapy across the spectrum of HF phenotypes.
Given the robust mortality benefit observed in this real-world setting, the incorporation of SGLT2i into routine clinical practice should be encouraged. Policymakers and healthcare providers in Thailand should consider expanding access to SGLT2i through national reimbursement schemes, particularly given their robust cost-effectiveness in Thai populations. Further multicenter studies are warranted to confirm these findings and support equitable implementation nationwide.
Authors’ contributions:
TC and PK designed the study, recruited the patients, acquired, analyzed, and interpreted the data, and drafted the article; CK, TS, and SP recruited the patients, acquired, analyzed, and interpreted the data; NS, PP, CP, AP, PJ revised the manuscript. NT and SC interpreted the data and revised the manuscript for important intellectual content; TC, SC, NT and PK critically revised the manuscript for important intellectual content. All authors have read and approved the final manuscript.
Institutional review board statement:
This study was approved by the institutional review board and the ethics committee of Naresuan University (IRB No. P3-00772567).
CRediT authorship contribution statement
Thananan Chanchanayothin: Writing – original draft, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Chuttikan Klomwong: Writing – original draft, Methodology, Investigation, Data curation. Tanawit Saisri: Writing – original draft, Methodology, Investigation, Data curation. Suphasin Panudom: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Sakchai Chaiyamahapurk: Writing – review & editing, Validation, Supervision, Methodology, Formal analysis, Data curation, Conceptualization. Nonthikorn Theerasuwipakorn: Writing – review & editing, Supervision, Methodology, Conceptualization. Noppachai Siranart: Writing – review & editing. Patavee Pajareya: Writing – review & editing. Nattakorn Songsirisuk: Writing – review & editing. Chalinee Pravarnpat: Writing – review & editing. Akenarong Pipatputthapong: Writing – review & editing. Pongpun Jittham: Writing – review & editing. Paisit Kosum: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- 1.Laothavorn P., Hengrussamee K., Kanjanavanit R., Moleerergpoom W., Laorakpongse D., Pachirat O., et al. Thai Acute Decompensated Heart failure Registry (Thai ADHERE) Glob. Heart. 2010;5(3) [Google Scholar]
- 2.McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Bohm M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2021/08/28 ed. 2021 Sep 21;42(36):3599–726. [DOI] [PubMed]
- 3.McMurray JJV, Solomon SD, Inzucchi SE, Kober L, Kosiborod MN, Martinez FA, et al. Dapagliflozin in Patients with Heart Failure and Reduced Ejection Fraction. N Engl J Med. 2019/09/20 ed. 2019 Nov 21;381(21):1995–2008. [DOI] [PubMed]
- 4.Docherty KF, Jhund PS, Anand I, Bengtsson O, Bohm M, de Boer RA, et al. Effect of Dapagliflozin on Outpatient Worsening of Patients With Heart Failure and Reduced Ejection Fraction: A Prespecified Analysis of DAPA-HF. Circulation. 2020/09/05 ed. 2020 Oct 27;142(17):1623–32. [DOI] [PMC free article] [PubMed]
- 5.Packer M, Anker SD, Butler J, Filippatos G, Pocock SJ, Carson P, et al. Cardiovascular and Renal Outcomes with Empagliflozin in Heart Failure. N Engl J Med. 2020/09/01 ed. 2020 Oct 8;383(15):1413–24. [DOI] [PubMed]
- 6.Anker SD, Butler J, Filippatos G, Ferreira JP, Bocchi E, Bohm M, et al. Empagliflozin in Heart Failure with a Preserved Ejection Fraction. N Engl J Med. 2021/08/28 ed. 2021 Oct 14;385(16):1451–61. [DOI] [PubMed]
- 7.Solomon SD, McMurray JJV, Claggett B, de Boer RA, DeMets D, Hernandez AF, et al. Dapagliflozin in Heart Failure with Mildly Reduced or Preserved Ejection Fraction. N Engl J Med. 2022/08/27 ed. 2022 Sep 22;387(12):1089–98. [DOI] [PubMed]
- 8.Vaduganathan M, Docherty KF, Claggett BL, Jhund PS, de Boer RA, Hernandez AF, et al. SGLT-2 inhibitors in patients with heart failure: a comprehensive meta-analysis of five randomised controlled trials. Lancet. 2022/08/31 ed. 2022 Sep 3;400(10354):757–67. [DOI] [PubMed]
- 9.Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022/04/02 ed. 2022 May 3;145(18):e895–1032. [DOI] [PubMed]
- 10.McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Bohm M, et al. 2023 Focused Update of the 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2023/08/25 ed. 2023 Oct 1;44(37):3627–39. [DOI] [PubMed]
- 11.Ariyachaipanich A., Buakhamsri A., Arintaya Phrommintikul S.C., Krittayaphong R., Senthong V., Naratreekoon B., et al. 2022 HFCT Focused Update of the 2019 HFCT Heart failure guidelines: Part 1 - Heart failure Classification and Pharmacological Treatment for Heart failure with Reduced Ejection Fraction (HFrEF) J. Med. Assoc. Thai. 2022;105(11):1153–1159. [Google Scholar]
- 12.Chirakarnjanakorn S., Senthong V., Ariyachaipanic A., Phrommintikul A., Buakhamsri A., Krittayaphong R., et al. 2023 HFCT Focused Update of the 2019 HFCT Heart failure guidelines Part 2: Diagnosis and Management of HFmrEF and HFpEF. J. Med. Assoc. Thai. 2023;106(10):1005–1014. [Google Scholar]
- 13.Deerochanawong C., Vareesangthip K., Piyayotai D., Thongsuk D., Pojchaijongdee N., Permsuwan U. Cost-Utility Analysis of Dapagliflozin as an Add-on to Standard Treatment for patients with Type 2 Diabetes and High risk of Cardiovascular Disease in Thailand. Diabetes Ther. 2021 Jul;12(7):1947–1963. doi: 10.1007/s13300-021-01088-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Krittayaphong R., Permsuwan U. Cost-utility analysis of add-on dapagliflozin treatment in heart failure with reduced ejection fraction. Int. J. Cardiol. 2021 Jan;1(322):183–190. doi: 10.1016/j.ijcard.2020.08.017. [DOI] [PubMed] [Google Scholar]
- 15.Krittayaphong R., Permsuwan U. Cost-Utility Analysis of Combination Empagliflozin and Standard Treatment Versus Standard Treatment Alone in Thai Heart failure patients with Reduced or Preserved Ejection Fraction. Am. J. Cardiovasc. Drugs. 2022 Sep;22(5):577–590. doi: 10.1007/s40256-022-00542-9. [DOI] [PubMed] [Google Scholar]
- 16.McMurray J.J.V., Packer M., Desai A.S., Gong J., Lefkowitz M.P., Rizkala A.R., et al. Angiotensin-neprilysin inhibition versus enalapril in heart failure. N. Engl. J. Med. 2014 Sep 11;371(11):993–1004. doi: 10.1056/NEJMoa1409077. [DOI] [PubMed] [Google Scholar]
- 17.Cowie M.R., Fisher M. SGLT2 inhibitors: mechanisms of cardiovascular benefit beyond glycaemic control. Nat. Rev. Cardiol. 2020 Dec;17(12):761–772. doi: 10.1038/s41569-020-0406-8. [DOI] [PubMed] [Google Scholar]
- 18.Verma S., McMurray J.J.V. SGLT2 inhibitors and mechanisms of cardiovascular benefit: a state-of-the-art review. Diabetologia. 2018 Oct;61(10):2108–2117. doi: 10.1007/s00125-018-4670-7. [DOI] [PubMed] [Google Scholar]
- 19.Girardi A.C.C., Polidoro J.Z., Castro P.C., Pio-Abreu A., Noronha I.L., Drager L.F. Mechanisms of heart failure and chronic kidney disease protection by SGLT2 inhibitors in nondiabetic conditions. Am. J. Physiol. Cell Physiol. 2024 Sep 1;327(3):C525–C544. doi: 10.1152/ajpcell.00143.2024. [DOI] [PubMed] [Google Scholar]
- 20.Saucedo-Orozco H., Voorrips S.N., Yurista S.R., de Boer R.A., Westenbrink B.D. SGLT2 Inhibitors and Ketone Metabolism in Heart failure. J Lipid Atheroscler. 2022 Jan;11(1):1–19. doi: 10.12997/jla.2022.11.1.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Vallon V., Verma S. Effects of SGLT2 Inhibitors on Kidney and Cardiovascular Function. Annu. Rev. Physiol. 2021 Feb;10(83):503–528. doi: 10.1146/annurev-physiol-031620-095920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kongmalai T., Tansawet A., Pattanaprateep O., Ratanatharathorn C., Amornritvanich P., Looareesuwan P., et al. Can SGLT-2 inhibitors improve cardiovascular outcomes and ensure safety for patients with type 2 diabetes and heart failure in Thailand? a real-world multicentre retrospective cohort study. BMJ Open. 2024 Dec 12;14(12) doi: 10.1136/bmjopen-2024-090226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Heerspink HJL, Stefansson BV, Correa-Rotter R, Chertow GM, Greene T, Hou FF, et al. Dapagliflozin in Patients with Chronic Kidney Disease. N Engl J Med. 2020/09/25 ed. 2020 Oct 8;383(15):1436–46. [DOI] [PubMed]
- 24.The EKCG, Herrington WG, Staplin N, Wanner C, Green JB, Hauske SJ, et al. Empagliflozin in Patients with Chronic Kidney Disease. N Engl J Med. 2022/11/05 ed. 2023 Jan 12;388(2):117–27. [DOI] [PMC free article] [PubMed]
- 25.Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, Hantel S, et al. Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes. N Engl J Med. 2015/09/18 ed. 2015 Nov 26;373(22):2117–28. [DOI] [PubMed]