Sodium Glucose Cotransporter-2 Inhibition in Heart Failure: Potential Mechanisms, Clinical Applications and Summary of Clinical Trials

Abstract

Despite current established therapy, heart failure (HF) remains a leading cause of hospitalization and mortality worldwide. Novel therapeutic targets are therefore needed to improve the prognosis of patients with HF. The EMPA-REG OUTCOME trial demonstrated significant reductions in mortality and HF hospitalization risk in patients with type 2 diabetes (T2D) and cardiovascular disease with the antihyperglycemic agent, empagliflozin – a sodium glucose co-transporter 2 (SGLT2) inhibitor. The CANVAS trial subsequently reported a reduction in 3-point MACE (major adverse cardiovascular events) and HF hospitalization risk. While SGLT2 inhibition may have potential application beyond T2D, including HF, the mechanisms responsible for the cardioprotective effects of SGLT2 inhibitors remain incompletely understood.

SGLT2 inhibition promotes natriuresis and osmotic diuresis, leading to plasma volume contraction and reduced preload, as well as decreases in blood pressure, arterial stiffness and afterload, thereby improving subendocardial blood flow in patients with HF. SGLT2 inhibition is also associated with preservation of renal function. Based on data from mechanistic studies and clinical trials, large clinical trials with SGLT2 inhibitors are now investigating the potential use of SGLT2 inhibition in patients with HF with and without T2D. Accordingly, in this review, we summarize key pharmacodynamic effects of SGLT2 inhibitors and the clinical evidence which support the rationale for the use of SGLT2 inhibitors in HF patients with T2D. Since presumably these favorable effects occur independent of blood-glucose lowering, we also explore the potential use of SGLT2 inhibition in patients without T2D with HF or at risk of HF, such as in patients with coronary artery disease or hypertension. Finally, we provide a detailed overview and summary of ongoing cardiovascular outcome trials with SGLT2 inhibitors.

Introduction

Worldwide, 1-2% of the general adult population have heart failure (HF), which is accompanied by reduced quality of life, high morbidity, mortality and significant financial costs¹. Existing therapies such as renin angiotensin aldosterone system (RAAS) inhibition, beta-blockade and angiotensin receptor blockers/neprilysin inhibitors (ARNI) reduce hospitalization and mortality risk in patients with HF and reduced ejection fraction (“HFrEF”)². Despite important cardiovascular benefits with the use of these agents, patients still have an increased risk for morbidity and mortality. These therapies also carry the potential for serious adverse effects including hypotension, kidney dysfunction and electrolyte abnormalities². Identification of novel therapeutic strategies to improve symptoms, reduce mortality, recurrent hospitalization, and acute decompensation is therefore critical to advance outcomes in HF patients.

Type 2 diabetes (T2D) is among the many comorbidities associated with cardiovascular disease (CVD) that contributes to end organ damage, and T2D also intensifies the risk for developing HF³ and HF-related complications, including death⁴. These risks are further compounded in the presence of diabetic nephropathy, highlighting an important interaction between T2D, chronic kidney disease (CKD) and HF⁵. The aims of this review are to summarize experimental and clinical evidence, which support the rationale for the use of antihyperglycemic sodium-glucose cotransport-2 (SGLT2) inhibitor agents in patients with T2D and HF, and to critically appraise whether SGLT2 inhibition may also be applicable in patients with HF without T2D.

HF Current Therapeutic Strategies and Unmet Needs

The two major types of HF, broadly categorized based on systolic function are: (1) HFrEF, left ventricular ejection fraction (LVEF) <40%, and (2) HF with preserved ejection fraction (HFpEF), LVEF ≥40%. Mechanistically, HF is accompanied by activation of several key neurohormonal regulatory systems, including the sympathetic nervous system (SNS), the renin-angiotensin-aldosterone system (RAAS), and by dysfunction of endogenous natriuretic mechanisms⁶. Activation of these compensatory pathways initially maintains blood pressure and preserves renal function. Over time, however, chronic neurohormonal activation increases left ventricular afterload and promotes vascular and cardiac remodeling and HF disease progression.

Current standards of care for HF patients with HFrEF include β-blockers, RAAS inhibitors, angiotensin receptor-neprilysin inhibitor (ARNI), diuretics and digoxin to suppress neurohormones, reduce volume overload and improve cardiac contractility^{6, 7}. Conventional diuretics only provide symptomatic relief for HF patients, but do not impact mortality (Figure 1). In contrast to HFrEF, current HF therapies fail to improve outcomes in patients with HFpEF⁶. Accordingly, HFpEF management focuses on the treatment of co-morbidities such as T2D, hypertension, coronary artery disease, and obesity⁸. Despite overall clinical benefits, currently available HF therapies, such as RAAS inhibitors and diuretics, increase the risk of adverse effects due to hypotension, volume depletion and SNS activation, highlighting the urgent need for safe, novel therapies^{9,10, 11}

Figure 1. Diuretic agents, their mechanisms of action and potential impact in patients with HF.

HF = heart failure; SGLT2i = sodium glucose co-transporter 2 inhibitor; PV = pressure volume; CV = cardiovascular; HHF = hospitalization for heart failure; LV = left ventricular.

Anti-hyperglycemic Agents and Cardiovascular Safety Trials

New insights into HF management have emerged somewhat unexpectedly from trials examining antihyperglycemic agents used in the treatment of T2D. While metformin and insulin may not impact HF progression, thiazolidinediones are associated with an increased risk of edema and HF¹². These findings led to the requirement by regulatory agencies for cardiovascular safety studies for all new antihyperglycemic agents. The dipeptidyl-peptidase 4 inhibitors (DPP-4i), saxagliptin and alogliptin but not sitagliptin, have also been associated with increased risk for hospitalization of HF by the Food and Drug Administration (FDA) and European Medicines Agency (EMA); responsible mechanisms are, however, not currently known^{13, 14}.

In contrast, liraglutide, a GLP-1 receptor agonist (GLP-1RA), was shown to reduce cardiovascular events by 13%, and also reduced all-cause mortality and nephropathy events in the Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) trial, a cardiovascular safety study in patients with T2D at heightened CV (cardiovascular) risk¹⁵ (Figure 2). Importantly, once-daily liraglutide administration did not increase the risk of hospitalization for HF compared to placebo including in the subgroup of patients with pre-existing HF after 3.8 years follow-up¹⁵. The cardioprotective effects associated with liraglutide emerged at 12-18 months in the context of only modest blood pressure and glucose lowering effects, suggesting that clinical benefits were derived through non-hemodynamic, non-glycemic anti-atherosclerotic mechanisms. Favorable effects, including a reduction in 3-point MACE (26%), were also reported in the SUSTAIN-6 trial with the investigational GLP-1RA agent, semaglutide – effects driven primarily by a reduction in the risk for stroke, also over approximately 12-18 months^{16, 17}. Consistent with observations from the LEADER trial, semaglutide did not significantly modify the risk of hospitalization for HF^{16, 17}. Importantly, the results of LEADER and SUSTAIN-6, which were obtained predominantly in T2D patients without HF, have been mirrored in real-world studies and meta-analyses examining GLP-1RA^{18, 19}. Unfortunately, in recently hospitalized patients with advanced HF and reduced LVEF (median 25%), dedicated studies with liraglutide were associated with overall neutral effects on cardiovascular death (at 180 days) and re-hospitalization rates compared to placebo. These data suggest that GLP1-RA agents are safe in HF^20–24 and reduce risk factors for CVD (blood pressure, body weight), particularly among T2D patients, but are unlikely to have direct beneficial effects on myocardial function in HF patients with advanced disease despite the expression of GLP-1 receptors in the heart.

Figure 2.

Hazard ratio and 95% confidence intervals for outcomes examined in EMPA-REG Outcome, CANVAS, LEADER and SUSTAIN 6. CV = cardiovascular; CI = confidence interval¹¹²

In contrast to LEADER and SUSTAIN-6, SGLT2 inhibition with empagliflozin reduced the risk of hospitalization for HF in the EMPA-REG OUTCOME trial^{25, 26}. This CV safety trial randomly assigned patients with T2D and established CVD (approximately 10% with established HF) to 10 or 25 mg of empagliflozin once daily versus placebo in addition to standard of care. Compared to placebo, participants treated with empagliflozin, pooled from both doses, had a 14% reduction in 3-point MACE (CV death, non-fatal myocardial infarction, non-fatal stroke). The 35% reduction in the risk of hospitalization for HF emerged within several months, suggesting a role for acute effects on renal and/or systemic hemodynamic function or direct effects on the cardiovascular system (Figure 3, “Mechanism of Action” section below). Benefits on 3-point MACE were similar across eGFR subgroups down to an eGFR of 30 ml/min/1.73m², despite the fact that SGLT2 inhibitors are associated with less glucosuria as renal function declines. EMPA-REG OUTCOME also reported a 39% reduction in the composite renal endpoint, new onset or worsening nephropathy (defined as incident macroalbuminuria, doubling of serum creatinine, eGFR <45 ml/min/1.73m², initiation of renal replacement therapy or a renal disease related death)²⁶.

Figure 3. Integrated Physiological Basis for Proposed Mechanisms Leading to Cardiorenal Benefits and Risks Associated with Sodium Glucose Cotransport-2 Inhibition.

* = only reported in the CANVAS Program; RAGE = receptor for advanced glycation end products; PTH = parathyroid hormone; FGF-23 = fibroblast growth factor 23; Na⁺ = sodium; AKI = acute kidney injury.

The results of two additional SGLT2 inhibitor CV safety trials – CANagliflozin cardioVascular Assessment Study (CANVAS) and CANagliflozin cardioVascular Assessment Study-Renal (CANVAS-R)²⁷ have also been reported. In CANVAS, 4,330 patients with T2D at high CV risk were randomly assigned 1:1:1 to 100, 300 mg/day canagliflozin or to placebo for 2 years to evalute 3-point MACE (Clinicaltrials.gov NCT01032629)²⁸. In CANVAS-R, the effect of canagliflozin for 2.5 years was evaluated in 5,700 patients with T2D at elevated CV risk, with the primary outcome of kidney disease progresion (Clinicaltrials.gov NCT01989754)²⁹. The secondary outcomes of CANVAS and CANVAS-R have been modified to prioritize vascular death, total mortality and HF and to refine the outcomes addressing myocardial infarction and stroke³⁰. Rather than evaluating separate effects of each dose of canagliflozin, the results of both studies and doses of canagliflozin were combined and reported as the CANVAS Program, with an overall HF prevalence of 14.4%. In the CANVAS Program with canagliflozin (n=10,142 patients with T2D), the primary endpoint, 3-point MACE, was significantly reduced, and the risk of hospitalization for HF was reduced by 33%³¹. From a renal perspective, progression of albuminuria was reduced by 27%, and the composite outcome (sustained 40% reduction in eGFR, the need for renal-replacement therapy, or death from renal causes) was reduced by 40%³¹.

The factors responsible for the acute beneficial effects on hospitalization for HF in EMPA-REG OUTCOME and the CANVAS Program are incompletely understood. The majority of patients (approximately 90%) in the EMPA-REG OUTCOME and CANVAS Program trials did not have HF at baseline, suggesting that the HF beneficial effects observed were due to a primary prevention effect. Importantly, SGLT2 inhibitors are associated with favorable effects on many of the comorbidities associated with HFpEF, including diabetes-related hyperglycemia, obesity and hypertension. Together with other potential protective mechanisms discussed in the next section,³² these effects may also reduce HF risk and/or CV death in HF patients without T2D, a demographic currently being evaluated in ongoing CV outcome trials (Table 1).

Table 1.

Clinical Trials with SGLT2i

Name of Clinical Trial	Type	Treatment Arms	Duration	Sample size	Diabetes Status	Primary Outcomes/Results	Secondary Outcomes	NCT	Estimated Reporting
Cardiovascular Outcome Trials
Empagliflozin Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-REG OUTCOME)	Double blind, placebo controlled RCT (Phase 3)	Empagliflozin 10 mg or 25 mg daily vs. placebo	Up to 4.6 years	7,020 patients with established CV complications (≥18 years)	All T2D	14% reduction in 3-point MACE (CV death, non-fatal MI, non-fatal stroke) pooled from 10mg and 25mg empagliflozin doses	35% reduction in hospitalization for HF, 39% reduction in the composite renal endpoint (new macroalbuminuria, doubling of serum creatinine and GFR ≤45, renal replacement therapy, renal death)	NCT01131676	Reported in 2015
Canagliflozin Cardiovascular Assessment Study (CANVAS Program)	Double-blind, placebo controlled RCT (Phase 3)	Canagliflozin 100 mg or 300 mg daily vs. placebo	3.6 years	10,142 patients with established vascular complications or ≥2 CV risk factors (>30 years)	All T2D	14% reduction in 3-point MACE (CV death, non-fatal MI, non-fatal stroke)	27% reduction in progression of albuminuria, 70$ increase in regression of albuminuria, 40% reduction in the composite renal endpoint (40% reduction in eGFR, renal replacement therapy, renal death)	NCT01032629	Reported in 2017
Evaluation of the Effects of Canagliflozin on Renal and Cardiovascular Outcomes in Participants with Diabetic Nephropathy (CREDENCE)	Double-blind, placebo controlled RCT (Phase 3)	Canagliflozin 100 mg daily vs. placebo	4 years	3,627 patients with Stage 2 or 3 CKD and macroalbuminuria and on ACEi/ARB (>30 years)	All T2D	ESKD, S-creatinine doubling, renal/CV death	CV death, non-fatal MI, non-fatal stroke, hospitalized UAP, hospitalized CHF, composite renal endpoint (ESKD, doubling of serum Cr. Renal death)	NCT02065791	Anticipated 2019
Multicenter Trial to Evaluate the Effect of Dapagliflozin on the Incidence of Cardiovascular Events (DECLARE-TIMI 58)	Double-blind, placebo controlled RCT (Phase 3)	Dapagliflozin 10 mg vs. placebo	Up to 6 years	17,276 patients with high risk for CV events (≥40 years)	All T2D	CV death, non-fatal MI, non-fatal ischemic stroke; CV death, hospitalization due to HF	Renal composite endpoint (≥40% decrease in eGFR to <60 and/or ESRD and/or renal or CV death, all-cause mortality	NCT01730534	Anticipated 2019
Cardiovascular Outcomes Following Ertugliflozin Treatment in Type 2 Diabetes Mellitus Participants With Vascular Disease (VERTIS)	Double-blind, placebo controlled RCT (Phase 3)	Ertugliflozin vs. placebo	Up to 6.1 years	8,000 patients with established vascular complications (≥40 years)	All T2D	CV death, non-fatal MI, non-fatal stroke	CV death, non-fatal MI, non-fatal stroke and hospitalized UAP	NCT01986881	Anticipated 2019
Large Randomized Controlled Clinical Trials in Patients with Heart Failure
Dapagliflozin Effect on Symptoms and Biomarkers in Diabetes Patients With Heart Failure (DEFINE-HF)	Double blind, placebo controlled RCT (Phase 4)	Dapagliflozin 10 mg daily vs. placebo	12 weeks	250 patients with HF (≥19 years)	All T2D	Change in NTproBNP	Change in SBP, weight, HbA1c, BNP and QoL score by questionnaire	NCT02653482	TBD
Dapagliflozin in Type 2 Diabetes or Pre-diabetes, and Preserved Ejection Fraction Heart Failure (PRESERVED HF)	Double blind, placebo controlled RCT (Phase 4)	Dapagliflozin 10 mg daily vs. placebo	12 weeks	320 patients with HF (≥19 years)	T2D or pre-diabetes	Change in NTproBNP	Change in SBP, weight, HbA1c, BNP and QoL score by questionnaire	NCT03030235	TBD
Study to Evaluate the Effect of Dapagliflozin on Incidence of Worsening Heart failure or Cardiovascular Death in Patients with CHF (DAPA-HF)	Double blind, placebo controlled RCT (Phase 3)	Dapagliflozin 10 mg daily vs. placebo	36 months	4,500 patients with HFrEF (≥18 years)	Non-diabetic and T2D (T1D excluded)	Time to CV death or hospitalization for HF or an urgent HF visit.	Time to ≥50% sustained decline in eGFR or ESRD. QoL score by questionnaire. Time to death by any cause.	NCT03036124	TBD
EMPagliflozin outcomE tRial in Patients With chrOnic heaRt Failure With Preserved Ejection Fraction (EMPEROR-Preserved)	Double blind, placebo controlled RCT (Phase 3)	Empagliflozin 10 mg daily vs. placebo	38 months	4,126 patients with HFpEF (≥18 years)	Non-diabetic, T1D and T2D eligible	Time to first event of adjudicated CV death or adjudicated hospitalization for HF.	Change in eGFR. Time to sustained reduction in eGFR. Time to all-cause mortality. Time to DM.	NCT03057951	TBD
EMPagliflozin outcomE tRial in Patients With chrOnic heaRt Failure With Reduced Ejection Fraction (EMPEROR-Reduced)	Double blind, placebo controlled RCT (Phase 3)	Empagliflozin 10 mg daily vs. placebo	38 months	2,850 patients with HFrEF (≥18 years)	Non-diabetic, T1D and T2D eligible	Time to first event of adjudicated CV death or adjudicated hospitalization for HF.	Change in eGFR. Time to sustained reduction in eGFR. Time to all-cause mortality. Time to DM.	NCT03057977	TBD
Mechanistic Clinical Trials in Patients with Heart Failure
Empagliflozin Impact on Hemodynamics in Patients With Diabetes and Heart Failure (EMBRACE-F)	Double blind, placebo controlled RCT (Phase 4)	Empagliflozin 10mg daily vs. placebo	12 weeks	60 patients with HF (≥19 years)	All T2D	Change in pulmonary artery systolic and diastolic pressure.	Change in mean pulmonary artery pressure. Change in QoL score by questionnaire.	NCT03030222	TBD
Study and Effectiveness of SGLT-2 Inhibitors in Patients with Heart Failure and Diabetes (REFORM)	Double blind, placebo controlled RCT (Phase 4)	Dapagliflozin 10 mg daily vs. placebo	12 months	56 patients with HF (18-75 years)	All T2D	Change in LV ESV or EDV (primary) by cardiac MRI	Change in LVEF, LVMi, fluid status by BIA and QoL score by questionnaire	NCT02397421	TBD
Empagliflozin in Heart Failure: Diuretic and Cardio-Renal Effects (EMPA)	Double blind, placebo controlled RCT (Phase 2)	Empagliflozin 10mg daily vs. placebo	14 days	50 patients with stable HF (≥18 years)	All T2D	Changes in urine sodium concentrations via ion selective electrodes.	Changes in blood volume by radiolabeled albumin.	NCT03027960	TBD
SGLT2 Inhibition in Diabetes and Heart Failure	Prospective cohort study	SGLT2i	4 weeks	31 patients with HF (≥18 years)	All T2D	Changes in peak oxygen consumption (VO2)(mL/kg/min) and minute ventilation (VE)/carbon dioxide production (VCO2) slop	Not specified	NCT02862067	TBD

HF = heart failure; SGLT2i sodium-glucose co-transporter inhibitor; RCT = randomized controlled trial, T2D = type 2 diabetes, QoL = quality of life, LV = left ventricular, ESV = end-systolic volume, EDV = end-diastolic volume, LVMi = left ventricular mass indexed, EF = ejection fraction, CV = cardiovascular, UAP = unstable angina pectoris, CHF = congestive heart failure, ESKD = end-stage kidney disease; TBD = to be determined

The results of the EMPA-REG OUTCOME trial have translated into recognition of empagliflozin as a CV protective therapy for patients with T2D by the Food and Drug Administration (FDA), the American College of Cardiology/American Heart Association, Diabetes Canada, and the European Society of Cardiology; similar perspectives may soon emerge for canagliflozin based on the CANVAS Program. Consistent with the observations in EMPA-REG OUTCOME and CANVAS, the broad impact of SGLT2 inhibition on CV outcomes has been observed in several “real-world” studies such as CVD-REAL, which reported significant benefits on mortality and hospitalization for HF risk with several SGLT2 inhibitor agents in an analysis of administrative data from 6 countries³³. Separate analyses examining dapagliflozin vs. other glucose lowering therapies revealed similar beneficial CV effects^{34, 35}, and an additional analysis demonstrated a 56% decreased risk of all-cause mortality and CVD with SGLT2 inhibition in a general practice cohort in the Sweden³⁶.

Mechanisms of Action of SGLT2 Inhibitors and Relationships with CV Outcomes (Figure 1)

High capacity, low-affinity SGLT2 are located in the renal proximal tubular epithelium and reabsorb filtered glucose^{37, 38}. SGLT2 inhibition-related glycosuria results in an insulin-independent HbA1c reduction of approximately 0.7-1.0% in patients with T2D, and also a body weight loss (~2 to 3 kg) due to induction of a negative caloric balance in patents with normal renal function³⁹ (Figure 3). SGLT2 inhibition-induced glycosuria also mediates a uricosuric effect^40–42 via the GLUT9 transporter⁴⁰, which may be cardioprotective, as increased plasma uric acid levels are associated with CV complications⁴³ and congestive HF⁴⁴. Despite being safely used in T2D patients with CKD in clinical trials, canagliflozin dosing reductions are required for patients with eGFR of <60 and >45 ml/min/1.73m². Furthermore, empagliflozin and canagliflozin are contraindicated in patients with eGFR of <45 ml/min/1.73m² and dapagliflozin is contraindicated in patients with eGFR of <60 ml/min/1.73m² due to reduced glycemic efficacy. Therefore, based on these glucose-centric GFR cut-offs, SGLT2 inhibitor use in HF patients with co-existing CKD may be “off-label”. Based on available evidence, it is unlikely that SGLT2 inhibitor-mediated reductions in death, HF or nephropathy events are mediated by HbA1c lowering, since HbA1c was only marginally decreased by <0.4% according to the trial design. A number of trials with other antihyperglycemic agents, perhaps with the exception of LEADER, have failed to show a positive impact of improved glycemic control on cardiorenal endpoints of magnitude seen in EMPA-REG OUTCOME or CANVAS Program¹². Similarly, glucosuria-associated weight loss was also unlikely to have accounted for the CV benefits with SGLT2 inhibition. While <3.5% decreases in total body weight was observed in EMPA-REG OUTCOME and CANVAS Program, trials such as LOOK AHEAD had no significant effects on lowering CV events over 10 years follow-up in overweight adults with T2D despite an average 6% loss of body weight⁴⁵.

Aside from effects on glycemic control and weight loss, SGLT2 are responsible for ~5% of sodium reabsorption at the proximal tubule under normal conditions. In the setting of chronic hyperglycemia, the capacity of SGLT2 in the kidney is increased leading to more pronounced effects on sodium homeostasis which decreases plasma volume and blood pressure with SGLT2 inhibition, as reviewed in more detail elsewhere^{39, 46–48}. Importantly for HF, SGLT2 inhibition-mediated effects on effective circulating volume contraction could reduce preload, thereby lowering ventricular filling pressure. Overall, the effects of SGLT2 inhibition on natriuresis and associated plasma volume contraction would be expected to reduce cardiac preload, while afterload reductions may occur through blood pressure and arterial stiffness lowering, thereby improving subendocardial blood flow.

In the kidney, SGLT2 inhibition-mediated proximal tubular natriuresis increases sodium delivery to the macula densa, which activates tubuloglomerular feedback, afferent vasoconstriction and reduces glomerular pressure, as reviewed elsewhere^{41, 49, 50}. These intrarenal hemodynamic effects may account for the 30-40% reduction in albuminuria observed with SGLT2 inhibitor agents including empagliflozin, dapagliflozin and canagliflozin²⁵. The hemodynamic effects of SGLT2 inhibition occur even in the presence of renal dysfunction, despite attenuated HbA1c lowering effects in this setting⁵¹. Preservation of GFR might be of particular importance in HF patients in order to avoid volume overload and complications arising from diuretic resistance (Figure 3). Therefore, natriuresis mediated by SGLT2 inhibition is likely a major factor leading to cardiovascular and renoprotective effects observed with empagliflozin and canagliflozin, which appears to extend across CKD stages⁵². Whether other SGLT2 inhibitors will also demonstrate cardiac protection remains to be determined in large prospective CV safety studies (Table 1)⁵³.

Despite the clinical relevance of maintaining sodium and water homeostasis in HF patients, reliable measurements of changes in total body sodium in humans remains difficult due to the potential for multiple sodium reservoirs throughout the body. Recent studies suggest that tissues such as skin and muscle store sodium, and that sodium accumulation is associated with CV disease (CVD)⁵⁴. In fact, the concentration of sodium in skin measured by ²³sodium-magnetic resonance imaging (²³Na-MRI) correlates strongly with left ventricular mass and blood pressure in CKD, suggesting that it is a surrogate marker for volume expanded states⁵⁵. SGLT2 inhibition with dapagliflozin for 6 weeks significantly reduces skin sodium concentration measured by ²³Na-MRI in 59 patients with T2D⁵⁶ suggesting that SGLT2 inhibition decreases both plasma volume and total body sodium content, which may protect against volume expansion and the risk of being hospitalized for HF.

Treatment of T2D patients with SGLT2 inhibitors typically reduces SBP by 4-6 mmHg and DBP by 1-2 mmHg^{46, 47}. In response to SGLT2 inhibition, similar magnitudes in SBP reduction (~2.7 mmHg) have also been observed in healthy individuals without T2D^57–61. Apart from natriuresis and osmotic diuresis, other mechanisms that might contribute to blood pressure lowering in response to SGLT2 inhibition include a decrease in arterial stiffness⁶², and effects on endothelial function or vascular architecture⁶³. Improvements in arterial stiffness with SGLT2 inhibition has been partly attributed to effective circulating volume contraction via mild diuretic effects and vascular smooth muscle relaxation⁶⁴. Despite these antihypertensive effects, the risks of myocardial infarction, hospitalization for unstable angina, or stroke were not reduced in the EMPA-REG OUTCOME or CANVAS Program trials. This lack of effect on atherosclerosis-related endpoints argues against a major role for blood pressure lowering alone leading to CV benefits. Moreover, previous meta-analyses using other agents have suggested that modest blood pressure lowering effects in EMPA-REG OUTCOME and the CANVAS Program were probably insufficient to account for the large impact on CV or renal endpoints⁶⁵. Therefore, based on available evidence from these clinical trials, it seems highly unlikely that the reduction in CV risk with empagliflozin or canagliflozin can be attributed to glycemic control, blood pressure, or body weight reduction suggesting the involvement of other SGLT2 inhibition-mediated mechanisms.

It is unclear whether SGLT2 inhibitors have direct effects on cardiomyocytes, since SGLT2 is not expressed in cardiac tissue in humans⁶⁶. Recent preclinical studies in rat ventricular myocytes treated with empagliflozin however, demonstrated direct SGLT2 inhibition-mediated cardiac effects by lowering myocardial intracellular sodium concentrations via inhibition of the myocardial Na⁺/H⁺ exchanger flux with a secondary decrease in intracellular calcium and an increase in mitochondrial calcium⁶⁷. Regulation of mitochondrial calcium concentrations by SGLT2 inhibition is of interest, because mitochondrial calcium is an activator of ATP synthesis and of antioxidant enzymatic pathways^{68, 69}. In fact, increasing mitochondrial calcium concentrations prevents sudden death and overt HF in porcine models⁶⁸. Additionally, increases in myocardial intracellular sodium and calcium concentrations are early hallmarks and drivers of cardiovascular death and HF^{70, 71} (Figure 3). Further cardioprotective effects associated with SGLT2 inhibition may be derived through attenuating the pro-thrombotic milieu associated with hyperglycemia, which may be achieved through dapagliflozin-mediated effects on neutrophil-derived S100 calcium-binding proteins A8/A9 (S100A8/A9), which interact with the receptor for advanced glycation end products (RAGE)⁷². In light of the rapid clinical benefits observed in EMPA-REG OUTCOME and the CANVAS Program in particular, and animal models demonstrating protective effects on left ventricular mass and left ventricular end diastolic diameter⁷³, SGLT2 inhibition may have protective effects on cardiomyocytes or on inflamamtory pathways in the CV system – an intriguing possibility that merits further investigation.

Beyond effects on traditional CV risk factors or mitochondrial cardiomyocyte pathways, is the possibility that the reduction in HF risk in EMPA-REG OUTCOME and in the CANVAS Program were related to metabolic substrate shifts from conventional fatty acids to ketone bodies. SGLT2 inhibition is associated with increased glucagon (possibly via direct pancreatic α-cell effects) and reduced insulin concentrations, thereby promoting the production of ketones such as β-hydroxybutyrate⁷⁴. Ketones are freely taken up by the heart and can act as an efficient substrate for myocardial energy generation, as discussed elsewhere^75–77. This metabolic substrate shift leads to improved energy utilization, potentially contributing to improved cardiac efficiency, contractility and CV protection. In conjunction with increases in hematocrit leading to improved oxygen delivery to the heart⁷⁸, SGLT2 inhibition may both increase oxygen supply to the heart, and reduce cardiac oxygen demand, thereby improving cardiac function.

Given the acute nature of the HF benefit in EMPA-REG OUTCOME²⁵, discussion around CV benefits have generally focused on either volume/hemodynamic pathways or on changes in energy utilization in the heart. However, it is important to recognize and highlight the possibility that both of these pathways may have protective and injurious potential. For example, natriuresis may be beneficial for reasons discussed above, but may also increase the risk of volume depletion and acute kidney injury, and promote neurohormonal activation. In EMPA-REG OUTCOME, empagliflozin was well tolerated hemodynamically and decreased the risk of AKI, possibly due to reduced loop diuretic use, which could help preserve renal function by the avoidance of volume depletion⁷⁹. In the CANVAS Program, canagliflozin also did not increase the risk of AKI. Fortunately, SGLT2 inhibition-mediated effects on the RAAS appear to be modest^80–82 and effects on symapthetic activity are neutral in patients without established HF⁸³.

Despite empirical evidence showing HF benefits in the EMPA-REG OUTCOME and CANVAS Program trials, volume-related effects of SGLT2 inhibition in patients with HF are not yet fully understood. For example, in a pilot study involving 20 Japaneses patients with T2D and HF, the addition of ipragliflozin to conventional loop or thiazide diuretics for 4 days decreased plasma natriuretic peptide levels without influencing plasma angiotensin II, aldosterone or noradrenaline levels⁸⁴. It is not known whether participants in EMPA-REG OUTCOME had HFrEF or HFpEF, since echocardiography was not performed at baseline. In a subsequent report, the EMPA-REG OUTCOME investigators explored HF outcomes from this trial and concluded that a decrease in HF hospitalization and CV death was consistent among patients with and without HF at baseline. Empagliflozin may therefore prevent HF decompensation as well as incident HF in predisposed individuals⁷⁹. The time course for the reduction in HF hospitalization risk with canaglifozin has not yet been reported in the CANVAS Program.

Impact of SGLT2 Inhibition on Cardiac Structure and Function

Until ongoing small mechanistic studies and clinical trials are completed (Table 1), the relative role(s) of SGLT2 inhibition on HF pathophysiology described above remains speculative. There are limited data evaluating whether SGLT2 inhibitors modify cardiac structure and function. In animals, SGLT2 inhibition reduces cardiac fibrosis, inflammation, and oxidative stress⁸⁵. In several rodent models of T2D, SGLT2 inhibition improves diastolic function and left ventricular hypertrophy while simultaneously reducing myocardial fibrosis and expression of profibrotic/prohypertrophic proteins^86–88. Verma et al. have also demonstrated cardioprotective effects in a zebrafish model of HF, suggesting conservation of cardiovascular protective effects across species⁸⁹.

In humans, an analysis of 10 patients with T2D (baseline HbA_1c 7.3%) and established CVD who underwent transthoracic echocardiograms before and after three months of treatment with empagliflozin was recently reported⁹⁰. The authors observed a significant reduction in left ventricular mass index, an established surrogate marker of CV risk and improved diastolic function as determined by early lateral annular tissue Doppler velocity. Considering SGLT2 are not expressed in the human myocardium⁹¹, it is difficult to reconcile whether and how SGLT2 inhibition could have direct effects on ventricular remodelling. It is possible these SGLT2 inhibition-mediated effects on ventricular remodelling occur secondary to indirect effects on sodium and calcium entry into cardiomyocyte cellular compartments or via unknown effects in the heart. Further mechanistic studies are needed to better interpret the exsiting clinical data and to better define the effects of SGLT2 inhibition in cardiovascular tissue.

SGLT2 Inhibition and Potential Side Effects

While SGLT2 inhibition has emerged as a promising new cardio-renal protective therapy, it is important to recognize potential adverse effects related to drugs in this class, including increased rates of genitourinary infections, postural hypotension, polyuria, diabetic ketoacidosis, acute kidney injury and possible increased rates of bone fractures (Table 2)^{31, 39, 92–95}. The most commonly reported side effect due to glycosuria is genitourinary infections, mostly genital mycotic infections but in some studies bacterial urinary tract infections as well^{39, 92–94}. Less common but perhaps more important in the HF population in the context of concomitant diuretic use, is the potential for volume depletion. Whether this hemodynamic effect will prove a serious risk for ill and elderly patients with comorbid medical conditions is unclear. Although the EMPA-REG OUTCOME trial did not report a specific approach to adjusting baseline diuretic agents, it reasonable to be cautious when combining SGLT2 inhibitors with other diuretics ⁹⁶, as suggested elsewhere³⁹, especially in patients with severe systolic dysfunction. While AKI has been reported with SGLT2 inhibition, rates of AKI were less common in empagliflozin versus placebo-treated patients in the EMPA-REG OUTCOME trial, and similar trends were observed in the CANVAS Program²⁶. This is reassuring, since 80% of patients in EMPA-REG OUTCOME were taking concommitant RAAS inhibitors, which also decrease intraglomerular pressure²⁵. From a safety point of view, SGLT2 inhibitors should be part of “sick day” T2D mangement strategies, and may need to be held in the context of NSAID use or radiocontrast administration due to the potential for volume depletion and hemodynamic side effects ⁹⁶.

Table 2.

Summary of Adverse Effects of Sodium Glucose Co-Transport-2 Inhibitors

Type	Proposed Mechanism of Action	Drug vs. Class	References
Genitourinary Infections	Increased infection risk in the setting of glycosuria	SGLT2i	39, 92–94
Polyuria/polakyuria	Glycosuria-induced osmotic diuresis, initial increase in natriuresis	SGLT2i	39, 92–94
Postural hypotension	Volume depletion, blood pressure lowering effects through volume independent mechanisms (i.e., glucose lowering, weight loss, reduced arterial stiffness) possibly in the setting of concomitant diuretic, anti-hypertensive use	SGLT2i	39, 92–94
Diabetic ketoacidosis	Ketoacidosis from reduced carbohydrate availability due to increased glycosuria and reduced insulin concentrations. SGLT2 inhibition is also associated with increase levels of counter-regulatory hormones such as glucagon, thereby increasing ketogenesis.	SGLT2i	105
Decrease in eGFR	Decreased intraglomerular pressure due to afferent vasoconstriction, leading to an expected rise in creatinine (acute eGFR “dip” of 4-6 ml/min/1.73m2). Acute kidney injury risk is not increased with SGLT2 inhibition in clinical trials (EMPA-REG OUTCOME, CANVAS Program) but has been reported in post-marketing data. Further work is required to determine the long-term significance of the eGFR dip, and to determine whether or not it is protective (i.e. associated with reduced glomerular hypertension and proteinuria).	SGLT2i	26
Fracture risk	Increased parathyroid hormone and fibroblast growth factor 23 secretion due to increased serum phosphate concentration. SGLT2i may predispose to hypotension and hence falls and bone fracture	Only reported in the CANVAS Program	108
Amputation risk	AMP kinase activation and inhibition of Complex I of the respiratory chain, changes in oxidative metabolism leading to ischemia Hyperviscosity due to haemoconcentration Hypotension leading to reduce limb perfusion	Only reported in the CANVAS Program	31

In the CANVAS Program, patients treated with canagliflozin had a significantly higher risk of lower extremity amputations compared to the placebo group³¹. The increased risk of lower extremity amputations observed in CANVAS with canagliflozin has not been reported with dapagliflozin, empagliflozin, or other SGLT2 inhibitors⁹⁷. While the potential mechanisms responsible for differences in risk rates between members of this drug class for lower extremity amputations are not known, it may be relevant to recognize that canagliflozin has less SGLT2 selectivity compared to the other currently available SGLT2 inhibitors, and this may as a consequence induce more glucosuria⁹⁸. While it is not yet known if this translates to greater volume depletion with canagliflozin, a more pronoucned rise in hematocrit and hemoconcentration could conceivably increase blood viscosity and hence tissue ischemia in the periphery. Interestingly, other diuretics that cause hemoconcetration may similarly increase amputation risk, although this has not been carefully studied⁹⁹. In addition, canagliflozin, but not empagliflozin, impacts mitochondrial AMP-kinase activity, which is involved in mitochondrial energy pathways¹⁰⁰. In animals, canagliflozin associated AMP kinase activation was associated with inhibition of Complex I of the respiratory chain, leading to increases in cellular AMP and ADP. While it is not known if these pharmacological effects observed preclinically impact lower extremity perfusion, oxygenation or energy delivery, these findings highlight the concept that SGLT2 inhibitors differ pharmacodynamically and thereby lead to differences in both benefit and adverse effect profiles. As an alternative explanation, data collection or reporting of amputations may have differed between the trials, leading to under-reporting with other agents. Whereas further work is required to determine whether or not increased amputation risk occurs across the SGLT2 inhibitor drug class, a recent administrative data analysis using FAERS reported a consistent increased risk of amputation with canagliflozin only, and no increased risk with either empagliflozin or dapagliflozin¹⁰¹. Whether the risk for amputation will be present in HF is unclear, but should be further explored to determine if hypotension and peripheral hypoperfusion, both characteristic of HF, impact amputation risk further.

Notwithstanding the theoretical benefits around energy utilization described above, increased ketonemia in patients with diabetes taking SGLT2 inhibitiors should raise concern for evolving diabetic ketoacidosis. Although not reported more frequently as an adverse effect in EMPA-REG OUTCOME or the CANVAS Program, the US FDA and EMA have issued advisories about the potential risk of diabetic ketoacidosis with SGLT2 inhibitors ^{102, 103}. Fadini et al. recently published an analysis of the FDA Adverse Event Reporting System (FAERS), which demonstrated a proportional reporting ratio (PRR) of 7.9 (95% confidence interval, 7.5-8.4) for diabetic ketoacidosis in patients using SGLT2 inhibitors compared to those using other agents, with a higher risk in patients with type 1 diabetes¹⁰⁴. An increase in the risk of diabetic ketoacidosis was also observed by Fralick et al. in > 76,000 commercially insured T2D patients in the United States taking SGLT2 inhibitors compared to those using DPP4 inhibitors (hazard ratio for diabetic ketoacidosis, 2.2 95% CI 1.4-3.6)⁹⁵. In contrast, a separate meta-analysis of completed clinical trials in >10,000 T2D patients reported a reduction in risk of diabetic ketoacidosis, while a third administrative database analysis in >150,000 T2D patients reported an overall neutral risk of diabetic ketoacidosis, highlighting that more data are needed to better understand this rare but potentially serious risk^105–107.

As a final comment, the CANVAS Program also reported an increased risk for bone fracture – an observation not yet reported with other agents in this drug class. Mechanistically, SGLT2 inhibition may exert adverse effects on bone physiology through increased concentrations of parathyroid hormone and fibroblast growth factor-23¹⁰⁸. The impact on bone fracture risk does not appear to exist across this drug class, since a recent meta-analysis by Tang et al failed to broadly demonstrate bone fracture risk with SGLT2 inhibitor use¹⁰⁹. Whether bone fracture risk is of specific concern with SGLT2 inhibitors in HF is not known. Nevertheless, these agents should be used with caution in frail individuals who are at risk of falling, including those with hypotension.

Ongoing SGLT2 Inhibitors Cardiovascular Trials

In light of the emerging evidence for CV protetction with SGLT2 inhibition, specifically empagliflozin and canagliflozin, clinical trials are currently underway to evaluate the impact of other SGLT2 inhibitors on primary CV outcomes in patients with T2D⁵³. The “Multicenter Trial to Evaluate the Effect of Dapagliflozin on the Incidence of Cardiovascular Events” (DECLARE –TIMI58) is a Phase IIIb trial that is evaluting the effect of dapagliflozin (10 mg daily) versus placebo over 4.5 years on the composite 3-point MACE primary outcome in patients with T2D (Clinicaltrials.gov NCT01730534)⁵³. This trial is enrolling 17,150 patients with T2D and either known CVD (secondary prevention cohort) or at least two risk factors for CVD (primary prevention cohort). The impact of ertugliflozin on CV and renal endpoints is similarly being examined in the VERTIS study, although results are not anticipated until after 2020⁵³.

In addition to these trials, CV endpoints have been included as secondary endpoints in primary renal outcome trials. In the Evaluation of the Effects of Canagliflozin on Renal and Cardiovascular Outcomes in Participants with Diabetic Nephropathy Study (CREDENCE), the objectives are to determine whether canagliflozin versus placebo for 5 years will reduce the primary composite endpoint of the study including end-stage kidney disease (ESKD), doubling of serum creatinine, renal or CV death in patients with T2D with stage 2 or 3 CKD and macroalbuminuria (Clinicaltrials.gov NCT02065791)⁵³. The effects on progression of CKD or CV/renal death of dapagliflozin versus placebo as an addition to standard of care in 4,000 patients with CKD are being examined in a 4 year follow-up Study to Evaluate the Effect of Dapagliflozin on Renal Outcomes and Cardiovascular Mortality in Patients With Chronic Kidney Disease (DAPA-CKD) (Clinicaltrials.gov NCT03036150). Importantly DAPA-CKD is recruiting patients with and without T2D to better define the role of SGLT2 inhibition as a renal protective therapy in the absence of hyperglycemia – perhaps through the natriuresis-related pathways discussed above. A second large CKD prevention trial with empagliflozin was announced on June 11, 2017 that will include patients with and without diabetic kidney disease, although no details are currently available. To date, no human studies have reported renal or cardiovascular effects in the setting of non-diabetic kidney disease. Accordingly, planned and ongoing nephropathy trials with empagliflozin and dapagliflozin in non-diabetic patients are based on the premise that volume-related and other non-glycemic mechanisms will extend to patients without diabetes, since glucose-lowering effects are unlikely to be responsible for end-organ protection¹¹⁰.

In addition to studies such as DAPA-CKD in patients without diabetes that will evaluate primary renal endpoints, several large clinical trials have been announced that will investigate the impact of SGLT2 inhibition in HF patients with and without T2D. These trials include DAPA-HF, which is enrolling HFrEF patients, and the EMPEROR trials, which are including patients with HFrEF and HFpEF (Table 1). In addition, there are a number of ongoing smaller mechanistic studies in patients with T2D in this area that will better define SGLT2 inhibition-mediated mechanisms resulting in clinical HF benefits – including natriuresis, plasma volume contraction and neurohomonal activation, as reviewed in Table 1. As in the case of renal trials including individuals without T2D, the rationale for undertaking HF trials with SGLT2 inhibition in patients who do not have T2D is to take advantage of glucose-independent cardiorenal protective pathways in response to SGLT2 inhibition. Once complete, these ongoing studies will help elucidate the complex pharmacodynamic effects of SGLT2 inhibition on CVD in patients with and without T2D.

Conclusions

SGLT2 inhibition may emerge as an effective and safe adjunctive therapy for HF; that promotes hemodynamic stability and helps correct volume overload, while avoiding the risks of volume depletion, independent of effects on hyperglycemia. Nevertheless, the mechanisms responsible for the acute cardioprotective effects of SGLT2 inhibition, such as sodium and water homeostasis and plasma volume regulation must be examined in diverse populations, including in patients with and without T2D, and in those with HFpEF and HFrEF. Furthermore, the therapeutic landscape will continue to evolve with the use of new agents that impact circulating volume, such endothelin-1A receptor antagonists (sodium retention) and ARNI (natriuresis). The diabetes research community will also need to focus on understanding how these various agents interact in combination – both in terms of blood pressure, albuminuria and cardiorenal benefits, and around the potential for adverse effects¹¹¹.

In light of EMPA-REG OUTCOME, the CANVAS Program and studies with GLP-1RA agents, the management of patients with T2D and existing CVD is evolving. Since clinicians now have access to effective therapies to prevent cardiorenal events, future trials need to consider whether the demonstration of non-inferiority is sufficient, or whether superiority should be targeted, especially for HF events. In addition, while existing trial evidence has demonstrated unexpectedly robust protection in patients with existing complications, it remains to be determined if these agents will be similarly effective as primary prevention strategies.