SGLT2 Inhibitors in Acute Heart Failure: A Meta-Analysis of Randomized Controlled Trials

Acute heart failure (AHF) is a major public health concern, affecting 26 million worldwide. Sodium-glucose cotransporter 2 (SGLT2) inhibitors are a class of glucose-lowering drugs, comprising canagliflozin, dapagliflozin, and empagliflozin that are being explored for AHF. We aim to meta-analyze the effectiveness of SGLT2 inhibitors compared to placebo for primary outcomes including all-cause and cardiovascular mortality, heart failure events, symptomatic improvement, and readmissions. Our secondary outcome is the risk of serious adverse events. This meta-analysis has been designed in accordance with the PRISMA Statement 2020. A systematic search across PubMed, Scopus, and Cochrane Library was conducted through August 13, 2022. The following keywords were utilized: sglt2, sodium-glucose transporter 2 inhibitors, sglt2 inhibitors, decompensated heart failure, de-novo heart failure, and/or acute heart failure. Only randomized controlled trials (RCTs) with adult patients (>18 years), hospitalized with de-novo AHF, acutely decompensated chronic heart failure with reduced, borderline, or preserved ejection, and receiving SGLT2 inhibitors were included. A quantitative analytical methodology was applied where the standardized mean difference (SMD) applying 95% confidence intervals (CI) for continuous outcomes and risk ratio (RR) with 95% CI was yielded. All tests were carried out on Review Manager 5.4 (Cochrane). In total, three RCTs were included pooling in a total of 1831 patients where 49.9% received SGLT2 inhibitors. The mean age was 72.9 years in the interventional group compared to 70.6 years in the placebo. Only 33.7% of the sample was female. The follow-up spanned 2–9 months. Heart failure events were reduced by 62% in the interventional group (RR = 0.66, p < 0.0001). readmissions had a reduced risk of 24% with SGLT2 inhibitors (RR = 0.76, p = 0.03). We assessed the efficacy and safety of SGLT2 inhibitors in preventing complications post-AHF. The odds of all-cause mortality, cardiovascular mortality, heart failure events, and re-admissions rates were substantially reduced within the first 1–9 months of hospitalization.


Introduction
Acute heart failure (AHF) is a major public health problem, affecting nearly 6 million people in the United States and 26 million people globally [1]. AHF is defined as a rapid onset or worsening of signs and symptoms of heart failure (HF) that require urgent medical care [2]. With over 1 million annual hospitalizations in the US, AHF is the leading cause of admissions in the elderly [3]. Approximately 1 in 5 patients with AHF are readmitted within 30 days of being discharged and over 3 in 5 patients are readmitted during the This study has been designed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement 2020 guidelines [23]. This meta-analysis was registered with PROSPERO 2022 CRD42022365431. A comprehensive literature search of digital databases, including PubMed, Scopus, and Cochrane Library was performed from inception to August 13, 2022. Keywords and medical subject headings (MESH) were combined and run on the PubMed database with the following terms: "sglt2," "sodium-glucose transporter 2 inhibitors," "sglt2 inhibitors," "decompensated heart failure," "de-novo heart failure," "acute heart failure". The Boolean (and/or) logic was applied. Search filters were not applied to ensure maximal, relevant publications were obtained. This strategy was tailored to other databases including Scopus, and Cochrane Library. The reference lists were also reviewed for review articles to identify all relevant studies (umbrella methodology). Potential unpublished studies on ClinicalTrials.gov were also reviewed to locate any ongoing trials and include any relevant data.

Eligibility Criteria
All RCTs, including patients aged > 18 years, hospitalized with de-novo acute heart failure (AHF) or acutely decompensated chronic heart failure with reduced, borderline, or preserved ejection fraction, and intervention group receiving SGLT2 inhibitors that have been approved or are currently investigational were included. All observational studies, letters to the editor, single-arm studies, and two-arm studies that did not report any outcomes of interest were excluded. There was no restriction placed with respect to the diagnosis of type 2 diabetes or lack thereof. Outcome measures were decided a priori based on already-published data on heart failure (HF) patients who received SGLT2 inhibitors for non-AHF conditions. Efficacy outcomes included all-cause mortality, patients with heart failure events, total heart failure events, cardiovascular mortality, readmissions, and change in Kansas City Cardiomyopathy Questionnaire (KCCQ) Total Symptom Score. Safety outcomes included serious adverse events, acute kidney injury, hepatic injury, hypotension, hypoglycemia, urinary tract infections (UTIs), and diabetic ketoacidosis (DKA).

Study Selection and Data Extraction
Two investigators (A.S. and Z.S.) independently conducted the screening and finalization of the articles. The titles and abstracts were screened in the first stage by each investigator independently. Any discrepancies were resolved by a third investigator (I.C.O.) in case of any disagreements. Shortlisted articles were screened for full-text eligibility by the two investigators independently. In the final stage, the discrepancies were solved by the third investigator, and the studies were included in the review. Data were systematically collected for efficacy and safety outcomes. For all identified studies, the findings were tabulated as (i) clinical trial identifier, (ii) author-year, (iii) country, (iv) dose of SGLT2, (v) frequency of intervention, (vi) total sample size, (vii) intervention arm sample size, (viii) control arm sample size, (ix) follow-up duration, (x) primary outcome, and (xi) secondary outcome.

Statistical Analysis
All studies identified from the databases were stored in Endnote X9 (Clarivate Analytics, London, UK). The duplicates were removed using the Endnote X9 deduplication tool. No duplicates were found using the online resources (i.e., ClinicalTrials.Gov). The methodology was both quantitative and analytical to ascertain the benefit of SGLT2 in reducing the risk of adverse outcomes following hospitalization for AHF. The KCCQ scale scores were continuous, and the difference in means along with standard deviation was computed. On noting these values, the standardized mean difference (SMD) was computed, reported as Cohen's d, applying 95% confidence intervals (CI). A random-effects model was applied as it was assumed that the observed estimates of treatment effect can vary across studies because of real differences in the treatment effect in each study as well as sampling variability (chance). Forest plots were generated for every outcome documenting Risk Ratio (RR), SMD, 95% CI, heterogeneity, and overall results. The formula for RR and SMD is as follows: RR = risk o f event in the SGLT2 group risk o f event in the control group SMD = di f f erence in mean outcome between groups standard deviation o f outcome among participants The minimum requirement to meta-analyze the findings was two or more studies reporting the same outcome measure. While a funnel plot was not generated to test for publication bias due to the limited number of studies (<10), the heterogeneity between the included studies was tested using the χ 2 -based Q test and the I 2 index. The statistical analysis was conducted using Review Manager 5.4 (RevMan, Cochrane).

Risk of Bias Assessment
The risk of bias was assessed for the included studies individually by two investigators (A.S. and Z.S.). The risk-of-bias tool for randomized trials (RoB 2) version 2 available in the Cochrane Handbook for Systematic Reviews of Interventions was utilized [24]. Five fixed sets of domains of bias were graded as "low," "high," or "some concerns." The five domains comprised: (i) bias from the randomization process, (ii) bias due to deviations from intended interventions, (iii) bias due to missing outcome data, (iv) bias in the measurement of the outcome, and (v) bias in the selection of reported results. In case of discrepancies between the two investigators, the third investigator (I.C.O.) resolved it through consensus.

PRISMA
We identified 2627 articles, and 1985 titles and abstracts were screened after duplicates were removed. There were 529 articles that met the criteria for full-text review and 3 articles were included. The trial selection process is presented in a PRISMA flow diagram ( Figure 1). There was a strong agreement among the reviewers for the inclusion of the articles; the title and abstract screening stage yielded a score of: κ = 0.86; and for full-text screening (κ = 0.88).

Risk of Bias
Each included study was critically appraised using the revised tool for Risk of Bias (ROB 2) for randomized trials at the level of five outcomes: all-cause mortality, heart-failure events, re-admissions, symptom score, and adverse events ( Figure 2). All studies were considered to have a "low" risk of bias based on the five pre-specified domains in the ROB-2 tool. One trial was identified as having "some concerns" due to deviation in reporting pre-specified primary outcomes that may introduce bias in the selection of the reported result (Domain 5).

Baseline Characteristics
A total of 1831 patients were included of whom 913 received intervention (49.9%). The drug of choice was Empagliflozin in EMPULSE [25] and EMPA-RESPONSE-WHF [26] and Sotagiflozin in SOLOIST-WHF [27]. Baseline characteristics were mostly similar between interventional and control groups. The overall mean age in the interventional arm was 72.9 (SD: 5.3) years. In the control arm, the overall mean age was 70.6 (SD: 2.1) years. Only 33.7% (n = 617) of the total sample was female. Follow-up duration ranged from 60 days to 9 months. A total of 770 (84.3%) of the patients in the intervention arm and 758 (82.6%) of the patients in the control arm had diabetes as a comorbid. Key characteristics are summarized in Table 1.

All-Cause Mortality
All three trials reported the data on all-cause mortality at different time points. Compared with the placebo group, the risk of mortality was reduced by 27% in the intervention group (RR: 0.73, 95% CI: 0.49-1.09, p = 0.12, I 2 = 18%) ( Figure 3). For the sensitivity test, SOLOIST-WHF [27] had the highest weight and was thus removed to determine whether there will be changes in the risk. The post-sensitivity analysis was conducted to further explore the risk after the SOLOIST-WHF trial was removed which had the highest weight [27]. It demonstrated an even larger reduction in risk of death such that the group receiving SGLT2 inhibitors still had a lowered risk of all-cause mortality at 51% compared to the placebo group, which was significant (RR: 0.49, 95% CI: 0.25-0.95, p = 0.04, I 2 = 0%). Both EMPA-RESPONSE-AHF [26] and SOLOIST-WHF [27] were analyzed separately as patients with acute decompensated chronic HF (ADCHF) were included in both these trials; the mortality risk reduction was 15% in patients with ADCHF who took SGLT2 inhibitors compared to placebo (RR:0.85, 95% CI: 0.62-1.15, p = 0.39, I 2 = 0%).

Heart Failure Events
Heart failure events (HFEs) were defined as a hospitalization/ER visit, an urgent care visit, or an outpatient visit requiring intensification of management (intravenous therapy, mechanical support, renal, or circulatory support) for worsening signs and/or symptoms of heart failure. All three trials reported heart failure events (HFEs) among the participants. Compared to the placebo group, the intervention group had a significant risk reduction of 62% in HFEs (RR: 0.66, 95% CI: 0.58-0.75, p < 0.0001, I 2 = 0%) ( Figure 4A). After removing the SOLOIST-WHF trial [27], the sensitivity analysis still indicated a 45% reduction in HFEs in favor of the intervention group which did not reach statistical significance (RR: 0.55, 95% CI: 0.24-1.26, p = 0.16, I 2 = 48%). The total number of HFEs presented was identified during the entire duration of the study. Two of three trials reported the total number of HFEs across participants. There was a significant risk reduction of 35% in the number of HFEs across the trials (RR: 0.65, 95% CI: 0.46-0.91, p = 0.01, I 2 = 0) ( Figure 4B). The total number of patients who had worsening heart failure or had died is attributable to cardiovascular causes till the end-of-trial visit was included. Two trials reported a combined endpoint attributable to the total number of patients who died of cardiovascular causes or had HFEs. There was a significant reduction of 28% in the intervention group for risk of either cardiovascular death or HFEs (RR: 0.82, 95% CI: 0.73-0.93, p = 0.002, I 2 = 0%) ( Figure 4C). Clinical benefit is defined as a hierarchical composite of all-cause mortality, number of heart failure events, and time to a first heart failure event, or a 5-point or more difference in baseline score of KCCQ-TSS Change in visual analog scale (VAS) dyspnea score, diuretic response, the percentage change in N-terminal pro-brain natriuretic peptide (NT-proBNP), and length of hospital stay  Forest plot for all-cause mortality of patients with acute heart failure receiving sodiumglucose cotransporter 2 inhibitor or placebo. The total number of events is presented in the given study duration [25][26][27]. CI: confidence interval; M-H: Mantel-Haenszel; SGLT2: sodium-glucose cotransporter 2.

Readmissions
All three trials reported readmissions due to HFEs. The endpoints for readmissions in EMPULSE [25] and EMPA-RESPONSE-AHF [26] were 30 and 60 days, respectively. The SOLOIST-WHF [27] reported readmission rates during the entire duration of the study. There was a 24% risk reduction of first-time readmission among the intervention group who had been discharged following acute heart failure, compared to the placebo group, which was significant (RR: 0.76, 95% CI: 0.60-0.98, p = 0.03, I 2 = 4%) ( Figure 5). A sensitivity analysis was conducted and SOLOIST-WHF [27] was removed to identify the readmission risk within 60 days as reported in EMPULSE [25] and EMPA-RESPONSE-AHF [26]. The readmission risk was reduced by 15% which did not receive statistical significance due to the smaller sample size of the remaining two trials (RR: 0.85, 95% CI: 0.31-2.31, p = 0.75, I 2 = 36%).

Kansas City Cardiomyopathy Questionnaire Total Symptom Score (KCCQ TSS)
Two of three trials noted the adjusted mean reduction in the Kansas City Cardiomyopathy Questionnaire Total Symptom Score (KCCQ TSS). An analysis was conducted to note the effect size of intervention vs. placebo for KCCQ TSS. While a small effect size in favor of intervention was found, the KCCQ-TSS reduction was significant which indicates improvement in the group receiving SGLT2 inhibitors (Cohen's d = −0.27, 95% CI = −0.40, −0.14, p < 0.0001, I 2 = 41%) ( Figure 6).

Ongoing Clinical Trials of SGLT2 inhibitors and Heart Failure
As of 17th November 2022, 15 ongoing clinical trials are being conducted testing the efficacy and safety of SGLT2 inhibitors on heart failure, diabetes mellites type 2, acute myocardial infarction, and chronic kidney disease ( Table 2). The interventions consist of Empagliflozin (10 mg, 25 mg), Dapagliflozin (10 mg), and Canagliflozin. Control and comparator group standard care approaches consist of either placebo or loop diuretics, vasodilators, inotropic agents, digoxin, and/or vasopressors.   Table 3.

Discussion
The primary objective of our systematic review and meta-analysis was to examine the efficacy of SGLT2 inhibitors compared to placebo for reduction of all-cause and cardiovascular mortality, heart failure events (HFEs), readmissions, and symptomatic improvement in patients admitted for acute heart failure. Our secondary objective was to assess the risk of serious adverse events in patients admitted for acute heart failure, receiving either SGLT2 inhibitors or a placebo. Our meta-analysis compiled the findings of three studies that utilized SGLT2 inhibitors in either acutely decompensated chronic HF (ADCHF) or de novo acute HF (AHF). We found that SGLT2 inhibitors reduced the risk of all-cause mortality by 27% and were significant at 51% when a post-sensitivity analysis was conducted. Furthermore, we also found a significant reduction in cardiovascular mortality and HFEs by 28% among patients receiving SGLT2 inhibitors. Specifically, the absolute number of HFEs and total patients with HFEs were significantly reduced by 62% and 35% among those who used SGLT2 inhibitors, respectively. Moreover, we found a significant reduction in first-time re-admission rates by 24% among patients with SGLT2 inhibitors. These findings were supported by significant symptomatic improvement and no additional safety concerns [28].
The initial evidence of the efficacy of SGLT2 inhibitors in the clinical trial setting was noted in EMPA-REG-OUTCOME [19]-where it was determined that individuals hospitalized for heart failure, and who were randomized into treatment groups had a two-fold reduced risk of being re-hospitalized or dying in the first 1-3 months of the first heart failure event [19]. The EMPA-RESPONSE-AHF trial was the first prospective clinical trial to have evaluated the clinical benefits of SGLT2 inhibitors among acute heart failure patients [26]. This trial was a double-blind, randomized, placebo-based, parallel-group study based in multiple centers; 80 individuals with acute heart failure with or without type 2 diabetes mellitus were randomized to either 10 mg/day of empagliflozin or to the control group; all randomized and medicinal administration was pursued within 24 h of admission [26]. The EMPULSE (EMPagliflozin in patients hospitalized with acUte heart faiLure who have been StabilizEd) trial was based in multiple countries/centers, randomized with a double-blind setting where the effects of the SGLT2 inhibitor (empagliflozin) was assessed for safety, clinical benefit and tolerability for acute heart failure [25,29]. In this trial, patients underwent an initial stabilization period which had a median time span of 3 days. The patients also received 10 mg per day of the SGLT2 inhibitor (empagliflozin) or no intervention with standard care for 90 days. The EMPULSE trial met its endpoint, where patients showed more clinical benefits as compared to placebo; the stratified win ratio was 1.36 (95% CI = 1.09 to 1.68, p = 0.0054) [25]. Notably, the benefits were unanimous throughout the different subgroups including those that had decompensated chronic heart failure and ventricular ejection fractions upwards or downwards of 40%. The intervention was considered to be both well-tolerated and safe for the patients.
Other representatives of the class including dapagliflozin are currently being investigated. The DISTATE-AHF, which is a multicenter, prospective, open-label, randomized trial is enrolling 240 patients in the US [30]. The patient population consists of patients with type 2 diabetes mellitus hospitalized with hypervolemic AHF and glomerular filtration rate above 30 mL/min/1.73m 2 [30]. With endpoints consisting of diuretic response, inpatient AHF, 30-day readmission rate, and safety metrics, it is yet to be quantified whether dapagliflozin will be a candidate therapeutic for AHF among patients with diabetes [30]. In the CHIEF-HF trial, 467 participants with heart failure regardless of diabetes or ejection fraction status were randomized to canagliflozin (100 mg) or placebo intervention [31]. While the enrollment was stopped early because of sponsor priorities, Spertus and colleagues (2022) report that the primary outcome of KCCQ TSS was changed at 12 weeks by 4.3 points (p = 0.016)-favoring canagliflozin [31]. The CHIEF-HF met its primary endpoint; with a shift in the paradigm of care for heart failure imminent [31], it is important to further test and review the findings of SGLT2 inhibitors in randomized, double-blind settings.
In our meta-analysis, we ascertain the pooled findings of all three RCTs published in this arena so far. The EMPULSE trial works as an add-on/complement to the results of the previous two trials that administered SGLT2 inhibitors; these trials depict the patient journey from undergoing acute heart failure to in-hospital admission and use of the intervention. Our findings are noteworthy as we collate proof to showcase that SGLT2 inhibitors can be considered an element of usual care, translating to clinical benefit [32]. Our study findings support that SGLT2 inhibitors help manage AHF patients following hospitalization.

Clinical Practice Recommendations
With abundant data on the use of SGLT2 inhibitors for heart failure, it is essential to review the feasibility in clinical practice. So far, literature reports that SGLT2 inhibition interacts with key pathways at the cellular level, thereby providing cardioprotective effects to patient populations. The mechanisms include cardiac remodeling, myocardial calcium handling, lipolysis, and modification of the epicardial thickness [33].
SGLT2 inhibitors reduce the risk of composite and specific heart failure outcomes. The benefits of SGLT2 inhibitor use are large among those with a history of heart failure and with diabetes. Moreover, benefits are also noted in weight control, blood pressure regulation, and hemoglobin levels. It may be worthwhile to consider SGLT2 inhibitor treatment initiation if heart failure predominates since the intervention is an excellent target for blood glucose and control. No notable adverse events are reported among patients with heart failure being treated with SGLT2 inhibitors as compared to the standard of care. The benefits of treatment on heart failure outcomes are insofar high based on current evidence when compared to other therapies used for heart failure management including ACE inhibitors.

Limitations
Our meta-analysis has certain limitations that must be reported. First, the duration of the intervention was variable across all trials. Second, there was limited uniformity in how the trials defined HFEs. Third, the follow-up duration was inconsistent. Certain aspects of care were administered for longer periods across the trials. Fourth, the outcome measures were conducted at different time points e.g., readmission rates. Fifth, the sample size was fairly small, with SGLT2 inhibitor group enrolled participants not reaching generalizability. Lastly, a majority of the sample was diabetic, which may lead to biases when making broadly characterizable decisions for patient care.

Conclusions
The current study looked into the efficacy and safety of SGLT2 inhibitors in preventing complications following AHF. By compiling the findings of three trials, we analyzed a total sample of 1831 patients irrespective of their diabetes status. Moreover, we also qualitatively critiqued 15 ongoing clinical trials administering SGLT2 inhibitors across 27 countries for a variety of cardiovascular conditions including acute heart failure. We found that SGLT2 inhibitors decrease the odds of all-cause mortality, cardiovascular mortality, heart failure events, and re-admission rates within the first 1-9 months of hospitalization. Additional studies are necessary to fully understand the beneficial impact of SGLT2 inhibitors in managing AHF patients without diabetes.