1. Introduction

About 30% of the patients with type 2 diabetes mellitus (T2MD) will develop heart failure ^[1]. T2DM patients have an increased cardiovascular mortality, heart failure related hospitalization and longer hospital stays when compared to patients without DM ^[2]. Moreover, the treatment of both conditions presents several therapeutic challenges. It is well known that previous trials have shown an increased risk of heart failure hospitalizations in patient treated with specific oral antidiabetic drugs ^[3].

The sodium-glucose cotransporter 2 inhibitors (SGLT2i) are a new type of oral antidiabetic drugs. They inhibit renal glucose reabsorption and increase renal glucose excretion, thus lowering plasma glucose levels. This unique mechanism of SGLT2i action is insulin independent, thus improving glycemic control without promoting hypoglycemia in the absence of exogenously administered insulin ^[4]. Recently, these drugs have shown important improvement in cardiovascular outcomes in patient with T2DM ^[5,6].

The objective of this review is to analyze the impact of SGLT2i on cardiovascular outcomes and chronic heart failure in randomized clinical trials, proposed physiological mechanism, and future perspectives.

---

2. Sodium-glucose cotransporter 2 inhibitors

---

2.1. Mechanism of action and efficacy

SGLT2 are a member of a sodium-glucose cotransporter family being the major cotransporter involved in glucose reabsorption in the kidney, in the epithelial cells of luminal membranes of S1 and S2 segment in the convoluted tubule. This co-transporter is responsible of the 97% renal glucose reabsorption (Figure 1) ^[7]. The SGLT2 receptor is not expressed in the myocardial tissue ^[8].

Other members of this family are the SGLT1, responsible for the 10% glucose renal reabsorption, being its main role related to intestinal glucose absorption ^[9].

The SGLT2 inhibition causes a higher rate of glycosuria (40 to 80 g per day) and a decrease of the HbA1c of 0.7% to 0.9%. Osmotic diuresis contributes to a reduction in the extracellular volume decreasing blood pressure by 3 to 5 mmHg. Moreover, the impaired energy balance due to glycosuria results in weight loss (2–3 kg) (Figure 2) ^[4,5,6,9].

---

2.2. Adverse effects and security warnings

The most common side effects in patients treated with SGLT2i were related to glycosuria, such as genital and urinary infections. A complete list of adverse events is shown in Table 1.

Regarding dapaglifozin, a high rate of bladder cancer was registered, being considered as a contraindication in patients with such affection ^[10]. Meanwhile, with canagliflozin, in randomized clinical trials, a higher than expected rate of low trauma fractures and lower limb amputations (predominantly toe and midfoot) were found ^[6]. Patients with history of prior amputation, peripheral vascular disease, and neuropathy were at highest risk for amputation. Canagliflozin is not recommended in patients with these risk factors. It remains uncertain whether these adverse effects are specific of each drug or could be presented as a class effect.

An increased risk of ketoacidosis has been observed in patients in treatment with SGLT2 inhibitors. Even if glycaemia is in the normal range, if nausea, vomiting or metabolic acidosis appear, patients treated with this type of drugs must be evaluated for ketone bodies in urine and serum. We also have to remember that SGLT2i are not recommended for the treatment of Type 1 diabetes ^[11].

---

3. SGLT2i and cardiovascular outcomes

Currently, there are two main randomized clinical trials that compare efficacy and impact on cardiovascular outcomes: EMPA-REG OUTCOME (Empagliflozin) and CANVAS Program (Canagliflozin) (Table 2).

---

3.1. EMPA-REG outcome

In the EMPA-REG trial, patients in the empagliflozin group presented a lower rate of major adverse cardiovascular events (MACE) when compared to placebo. This decrease was similar when comparing two different treatment doses (10mg vs 25mg) and independent of the glycemic control. A reduction in cardiovascular mortality appeared to be. Of note, the treatment with empagliflozin was not associated with a reduction in ischemic events, such as myocardial infarction (MI) and stroke ^[5]. Thus, the lower cardiovascular mortality seems to be driven by a reduction in non-ischemic events.

---

3.2. CANVAS Program

The CANVAS Program integrates data from the CANVAS and CANVAS-R trials ^[6]. The patients treated with canagliflozin presented a lower rate of MACE when compared to placebo. Eventually, regarding the individual endpoints, there was no difference in the all-cause mortality or in cardiovascular mortality between canagliflozin and placebo. Furthermore, there was no difference in the ischemic events (MI and stroke).

---

3.3. EMPA-REG vs CANVAS

Both trials were performed according to the FDA regulation of antidiabetics drugs and cardiovascular safety ^[12]. The objective and design are very similar, although there are some differences that should be mentioned. Regarding the population of the trials, the EMPA-REG trial included patients with history of ischemic events such as: MI, coronary artery disease (CAD), stroke and periphery arterial disease (PAD) ^[5]. Meanwhile, the CANVAS, performed a different inclusion criteria definition: Patients aged between 30 and 49 years should have a documented symptomatic CAD and patients older than 50 years should have two or more prespecified cardiovascular risk factors with or without documented CAD or previous ischemic events (MI or stroke) ^[6]. This difference in the selection criteria was related to differences in the baseline characteristics of the patients in both trials. In the EMPA-REG trial a 75.6% suffered CAD (46.7% with a previous MI). However, in the CANVAS Program, patients with CAD only represented a 54.4%.

These differences should be considered when the outcomes of the trials are compared. Despite both trials met their primary composite endpoint, in the EMPA-REG, a reduction in the cardiovascular mortality existed, whereas not in the CANVAS. Moreover, neither studies had impact on the ischemic events (MI or stroke) (Table 2). It remains uncertain whether these issues could explain the differences in the results or if a real difference exists in the efficacy of the drugs.

---

3.4. Findings in perspectives

The positive results of these studies have created a great expectation in the cardiology and endocrinology scientific communities. However, the findings need to be kept in perspective. The important cardiovascular benefit of empagliflozin and canagliflozin, while statistically significant and clinically relevant, was observed in high-risk population, with established cardiovascular disease (CVD) at baseline. Specifically, in patients treated with canagliflozin, the increased risk of amputations should be considered. The low impact of the SGLT2i in the glycemic control suggests that extra-glycemic effects were responsible for the CVD outcome. However, it remains unknown if these drugs could cause the same effects on T2DM patients without CVD. Furthermore, the precise mechanism by which reduction in MACE was produced is still not well understood.

---

4. SGLT2i and heart failure

Considering the incremented risk of heart failure with counted oral hypoglycemic agents, the regulatory agencies focus special attention on heart failure outcomes ^{[3,4,5,6,9,10,11,12]}. In both EMPA-REG and CANVAS, the impact of SGLT2i on heart failure, defined as hospitalization due to these event, was analyzed (Table 2).

Despite the evidence between glycemic control and microvascular events, no trials have demonstrated the relation between glucose levels in blood stream and heart failure.

---

4.1. EMPA-REG outcome

In the EMPA-REG trial an impressive reduction in the hospitalization due to heart failure was found (NNT: 74.1). Moreover, in the empagliflozin group there was a reduction of the investigator-reported incidence of heart failure (NNT: 58.8) and introduction of loop diuretic treatment (NNT: 21.7). This decrease was observed in both, patients with and without heart failure, and was consistent in both empagliflozin doses. These data suggest a reduction in the incidence and progression of heart failure ^{[5,6,9,10,11,12,13]}. Only 10% of the included patients suffered a previous cardiac failure event, based on the narrow standard MedDRA query 'cardiac failure' ^[14]. It should be highlighted that the decrease in cardiovascular deaths was not associated to ischemic events but with a reduction in heart failure events.

---

4.2. CANVAS Program

In the CANVAS, the canagliflozin group showed an important reduction in hospitalization due to heart failure (NNT: 31.3). There was trend towards a reduction in both, patients with and without heart failure. Patients with heart failure history represent an 11.9% of the sample ^[6].

---

4.3. EMPA-REG vs CANVAS

Both trials showed consistent and clinical important reduction on heart failure events. At present, there is more detailed information of the EMPA-REG trial regarding heart failure events, meanwhile post-hoc analysis of the CANVAS is still missing ^[13,14,15].

---

4.4. Findings in perspective

Given the previous unsatisfactory heart failure outcomes with thiazolidinediones and DPP-IV inhibitors, the safety profile of oral hypoglycemic agents have been under scrutiny and several trials have been published regarding the cardiovascular safety of oral hypoglycemic agents ^[16,17]. Incretin mimetics have also been evaluated ^[18,19,20], one of the most recent published studies, the exenatide study of cardiovascular event lowering trial (EXSCEL) ^[21], evaluated the CV safety of exenatide, added to usual care in patient with type 2 diabetes, demonstrating that it was noninferior to placebo with regarding safety.

Thus, the impressive results of the EMPA-REG and CANVAS are more than welcome. Eventually, there are several issues in relation to the design of these trials. First, none of these trials were sample powered for heart failure endpoints. Second, the selection criteria in both trials were not as strict as in a conventional heart failure trial. In the EMPAG-REG were based on a questionnaire and in the CANVAS on medical history. Third, baseline or follow-up left-ventricular ejection fraction (LVEF) assessment was not performed regularly. Fourth, in both trials approximately only a 10% of the enrolled patients had previous history of heart failure. Fifth, in the subgroup analysis, no benefit in patients taking aldosterone receptor blockers appeared, not knowing if an overlap in the mechanism of these two treatments exists ^[5,6]. Finally, the proportion of patients with preserved or reduced heart failure remains unknown. Considering all these reasons, despite of an important clinical benefit, further research is needed to completely understand the mechanism and to define which population could benefit more from these drugs.

---

5. Physiological mechanism of SGLT2i

The beneficial outcomes of SLGT2i related to decrease in progression of heart failure have not been related to glycemic control ^[22,23,24]. Several of these mechanisms are still not completely understood and it remains unknown if they can be defined as class effects or specific for each SLGT2i drugs ^[25].

Some of the proposed mechanisms related to beneficial (✓) and adverse effects (✕) are the following (Figure 2):

---

5.1. Glucosuria

✓    Reduction in the HbA1C

✓    Uricosuria and reduction in uric acid

✓    Decrease of insulin and increased in glucagon production. Energy substrate change by use of ketones by the heart

✓    Reduction in body weight

✕    Possible risk of diabetic ketoacidosis

✕    Hypoglycemia, specially when combined with insulin or sulphonylureas

✕    Genitourinary infections

---

5.2. Natriuresis

✓    Increase of tubuloglomerular feedback: Maintenance of euvolemic and reduction of proteinuria

✓    Decrease of renal cortical ischemia, increase in the erythropoietin and hematocrit and therefore an increase in myocardial oxygen delivery

✓    Decrease in arterial stiffness

✓    Decrease in plasma volume: in preload y afterload

✕    Hypotension: Acute kidney injury, falls, fracture, tissue hypoperfusion

✕    Hyperviscosity: Tissues ischemia and risk of lower limbs amputation

---

5.3. Inhibition of sodium hydrogen exchangers (NHE)

✓    Reduction in LV hypertrophy

✓    Reduction in LV fibrosis and remodeling

✓    Reduction in high-sensitivity troponin I and N-terminal pro-B-type natriuretic peptide

---

5.4. Direct cardiac effects

Recently, it has been proposed that the direct cardiac effects of the SGLT2i can be related to their interaction with sodium-hydrogen exchangers (NHE). It has been demonstrated that NHE exhibits an increased function in patients with heart failure ^[26,27]. A higher intracellular calcium concentration, secondary to an increased intracellular sodium, has been related to myocyte injury and hypertrophy, leading to a higher rate of fibrosis, cardiac hypertrophy and heart failure ^[24].

---

6. Future perspective

The positive results of the EMPA-REG and CANVAS, specially the reduction in heart failure events, have led to an extensive basic and clinical research. Currently, several randomized clinical trials with the objective to evaluate the impact of SGLT2i in heart failure in diabetic patients, are ongoing (Table 3). These trials were specifically designed with this aim, considering sample size, primary endpoint, cardiac imaging assessment, heart failure specific treatment such as cardiac resynchronization therapy, etc. Eventually, EMPEROR trials will assess both types of heart failure, preserved and reduced. The DECLARE-TIMI58 primary endpoint should be presented in 2019 adding an extensive and high-quality evidence of the cardiovascular safety of SGLT2i. Finally, small trials are evaluating the hypothesis of the possible application of SGLT2i to non-diabetic patients, but probably there is still a long way to go before this could be applied to clinical practice ^[28,29].

---

7. Conclusions

Due to regulatory agencies' requirements, cardiovascular safety of oral hypoglycemic agents has been extensively studied, but inconsistent results have been previously found. The EMPA-REG outcome and CANVAS Program have not only showed a reduction in major cardiovascular outcomes, but also an improvement in heart failure events. Furthermore, the paradigm of the glycemic control begins to shift to a more clinical outcome related focus. Despite of the important clinical effects of SGLT2i, these results should be taken in perspective, without forgetting the selected and high-risk population included and methodological limitations regarding heart failure outcomes. Additionally, serious adverse events such as increased risk of bone fracture and lower limb amputations, should be considered when prescribing SGLT2i. Further research is needed to determine the mechanism related to benefits and adverse effects of these drugs, and determine their target population in daily clinical practice.

---