Effects of antidiabetic drugs on the incidence of macrovascular complications and mortality in type 2 diabetes mellitus: a new perspective on sodium–glucose co-transporter 2 inhibitors

Abstract

Elevated hemoglobin A_1c (HbA_1c) values correlate with microvascular and macrovascular complications. Thus, patients with type 2 diabetes mellitus (T2DM) are at an increased risk of developing macrovascular events. Treatment of T2DM should be based on a multifactorial approach because of its evidence regarding reduction of macrovascular complications and mortality in T2DM. It is well known that intensive glucose control reduces the risk of microvascular complications in T2DM, but the effects of antidiabetic drugs on macrovascular complications and mortality in T2DM are less clear. The results of recent trials have demonstrated clear evidence that empagliflozin and liraglutide reduce cardiovascular (CV) and all-cause mortality in T2DM, an effect that is absent in other members of antidiabetic drugs. Empagliflozin is a member of a novel class of antidiabetic drugs, the sodium–glucose co-transporter 2 (SGLT2) inhibitors. Two ongoing randomized clinical trials involving other SGLT2 inhibitors, canagliflozin and dapagliflozin, will provide additional evidence of the beneficial effects of SGLT2 inhibitors in T2DM population. The aim of this paper is to systematically present the latest evidence regarding the usage of antidiabetic drugs, and the reduction of macrovascular complications and mortality. A special emphasis is put on the novel class of antidiabetic drugs, of SGLT2 inhibitors.

Key messages

• Macrovascular complications and mortality are best clinical trial endpoints for evaluating the efficacy of antidiabetic drugs.

• The first antidiabetic drug that demonstrated a reduction in mortality in the treatment of type 2 diabetes mellitus (T2DM) was empagliflozin, a sodium–glucose co-transporter 2 (SGLT2) inhibitor.

• SGLT2 inhibitors are novel class of antidiabetic drugs that play a promising role in the treatment of T2DM.

Keywords:

• Type 2 diabetes mellitus
• macrovascular complications
• mortality
• antidiabetic drugs
• sodium–glucose co-transporter 2 inhibitors

Introduction

Type 2 diabetes mellitus (T2DM), a heterogeneous disorder characterized by insulin resistance, is a result of interactions between environmental, genetic, and behavioral risk factors. In 2015, approximately 415 million adults suffered from diabetes mellitus (DM), and it is estimated that this number will increase to 642 million by 2040. In high-income countries, the majority of patients with DM have T2DM (91%) (1). The pathophysiology of cardiovascular (CV) diseases in patients with T2DM is characterized by the progression of vascular inflammation, endothelial cell dysfunction, and subsequent formation of atherosclerotic plaques. These changes in the vascular endothelia result in the development of atheroma. It is important to note that in patients with T2DM, atheroma has more inflammatory changes, higher lipid levels and greater amounts of thrombi compared to atheroma in people without T2DM (2). The results from the United Kingdom Prospective Diabetes Study (UKPDS) 35 trial confirmed that the risk of microvascular (retinopathy, nephropathy, and neuropathy) and macrovascular (cerebral, coronary, and peripheral vascular disease) complications correlate with elevated levels of hemoglobin A_1c (HbA_1c) (3). Patients with T2DM are at an increased risk of developing CV events, such as stroke, coronary heart disease, peripheral arterial disease, congestive heart failure (HF), or cardiomyopathy. The CV complications are the main cause of morbidity and mortality among the diabetic population, which makes these complications the best clinical trial endpoint for evaluating the efficacy of antidiabetic drugs. Several international guidelines have also highlighted the importance of the prevention and reduction of the CV risk, in patients with T2DM (2,4).

Cardiovascular drugs in T2DM

The treatment of T2DM should be based on multifactorial approach because of the evidence regarding reduction of macrovascular complications and mortality in T2DM (5). Besides regulation of glycaemia, it is of critical importance to focus treatment on the factors that have the greatest impact on improving CV outcomes.

Hypertension

Hypertension is one of the major risk factors for cardiovascular disease (CVD), with very high prevalence rates among the T2DM population (6). Current treatment recommendations strongly encourage prescribers to lower blood pressure (BP) in hypertensive diabetics, although no consensus about the desired BP values has been reached. Most relevant guidelines recommend different BP values for the majority of patients; <140/90 mmHg (7), <140/85 mmHg (2), and <130/80 mmHg (8). Usually multiple drug therapy is necessary to achieve BP targets in T2DM patients; therefore, treatment should be individualized according to characteristics of each patient. Any antihypertensive drug can be used; however, the use of an angiotensin-converting-enzyme inhibitor (ACE-I) or an angiotensin-receptor blocker (ARB) to block renin-angiotensin-aldosterone system (RAAS) is preferred by most of the up-to-date guidelines, particularly in the presence of proteinuria (2,7,8). RAAS inhibitors seem to be of great value in T2DM patients due to their beneficial effects in reducing mortality and macrovascular complications (9–11). While multiple drugs are often needed to control BP, the combination of drugs that work on RAAS should be avoided (2,7). A dual RAAS blockade using ACE-I and ARB demonstrated no beneficial effects regarding macrovascular complications and mortality, but showed an increased incidence of adverse events (12). At least one antihypertensive drug, if possible, should be administered at bedtime, due to a reduction of CV morbidity and mortality compared with morning intake by patients with T2DM (13).

Dyslipidemia

Patients with T2DM have an increased prevalence of lipid abnormalities and consequently high risk of atherosclerotic cardiovascular disease (ASCVD) (7). According to a multivariate analysis from UKPDS 23 trial, increased concentration of low-density lipoprotein cholesterol (LDL-C) and decreased concentrations of high-density lipoprotein cholesterol (HDL-C) are the strongest predictors of coronary artery disease (CAD) in T2DM patients (14). Consequently, management of LDL-C concentrations remains one of the main targets of T2DM treatment. Several placebo-controlled randomized controlled trials (RCTs), sub-analyses of RCTs, and meta-analysis of 14 RCTs have demonstrated that lowering LDL-C with statin therapy has a beneficial effect in patients with DM by reducing CV events and mortality (15–18). Interestingly, the same effect has not been documented in patients with T2DM on hemodialysis (19). The additional reduction of LDL-C can be achieved by combining ezetimibe and statin therapy. The analysis of a subpopulation in the IMPROVE-IT trial showed that this combination has a beneficial effect on the reduction of CV events in a diabetic population with recently diagnosed acute coronary syndrome (ACS) (20). Statins slightly increase the risk for development of DM, but their usage still seems to have positive benefit-risk ratio (21). In T2DM population, this risk is irrelevant due to an already manifested disease of T2DM. Current guidelines advice prescribes to lower LDL-C levels in patients with T2DM and CV risk factors. They recommend use of statins, which have shown dose dependence efficacy (2,7,8,22). Low levels of HDL-C are commonly associated with increased levels of triglycerides. Together with high LDL-C levels are a common pattern of dyslipidemia in patients with T2DM (23). HDL-C levels have been connected in inverse correlation with CV risk, but pharmacological interventions that elevate HDL-C concentrations did not demonstrate evidence of a beneficial effect on CV risk (14,24–26). To improve the understanding of the role of HDL on CV risk, further research is needed (27). Despite unclear evidence, whether modifying HDL-C and triglyceride levels impacts CVD risk, these parameters have been incorporated into up-to-date guidelines (7,8).

Antiplatelet therapy

Patients with DM have an increased thrombotic tendency, and a number of antiplatelet drugs have shown to be of benefit in preventing CV events in certain groups of patients (23). Current evidence suggests that there is no difference in the efficacy of aspirin in secondary prevention in patients with DM, compared to people without DM (28). The use of aspirin is recommended for secondary prevention, whereas clopidogrel is an alternative treatment in cases of aspirin intolerance. Most current guidelines recommend the same dose of aspirin, 75–162 mg per day, for DM patients and individuals without DM (2,7). Dual antiplatelet therapy in the combined administration of a P2Y12 receptor antagonist and aspirin should be used for up to one year in patients following an ACS. Dual antiplatelet therapy should also be used in those patients subjected to percutaneous coronary intervention (PCI) in which the duration of treatment depends on the stent type (2). The usage of aspirin in the primary prevention of CV events in patients with DM is controversial. In addition to its beneficial effect on CV events, the use of aspirin is associated with increase in the risk of hemorrhage. Therefore, its usage needs careful evaluation (29). At this point of time, current evidence does not support routine use of aspirin in the primary prevention in most subgroups of patients with DM (30–33). Although evidence is weak, the usage of aspirin in primary prevention should only be considered in patients that have an ASCVD 10 year risk of >10% without an increased risk for bleeding (7). Fortunately, two large, ongoing RCTs will shed more light on this matter (34,35).

Glycemic control and its effect on macrovascular complications and mortality

The results from two large RCTs confirmed that intensive glucose control in T2DM patients reduces the risk of microvascular complications (36,37). However, there is weak or insufficient evidence to show that treatment with biguanides, sulfonylureas (SU), thiazolidinediones (TZD), alpha-glucosidase inhibitors, glinides, dipeptidyl peptidase-4 (DPP-4) inhibitors, or insulin can reduce macrovascular complications and mortality in patients with T2DM. Several RCTs have documented that the intensive regulation of glycaemic control did not prove to be beneficial in reducing either macrovascular complications or mortality (37–39). Moreover, the ACCORD trial demonstrated a 22% increase in the total mortality in the intensive therapy group (targeting HbA_1c below 6%) compared to the standard therapy group (targeting the HbA_1c from 7% to 7.9%). Mortality was mainly driven by the CV events and interestingly, the rates of hypoglycemia were three-fold higher in the intensive therapy group (39). Although several RCTs have confirmed that insulin, SU, and glinides are particularly associated with hypoglycemia, only the ACCORD trial implicate correlation between hypoglycemia and higher mortality rate (36,39–41).

Biguanidines

Various members of several classes of antidiabetic drugs have been evaluated in RCTs regarding their effects on macrovascular complications and mortality in T2DM. Metformin, a member of the biguanides, is the most often used antidiabetic drug to treat T2DM. However, the results of RCTs have provided inconsistent evidence for the effectiveness of metformin in reducing macrovascular complications and mortality. The UKPDS 34 trial compared effects of metformin on macrovascular complications and mortality in overweight patients (more than 120% of ideal body weight) treated using conventional policy (diet alone) and other antidiabetics (insulin, chlorpropamide, or glibenclamide). The median follow-up of the trial was 10.7 years. Metformin, compared to conventional treatment demonstrated beneficial effects in terms of the risk reduction for myocardial infarction (MI) (39%, p = .01), diabetes-related death (42%, p = .02) and all-cause mortality (36%, p = .01). Compared with other antidiabetics, metformin showed a statistically significant effect on the reduction of all-cause mortality (p = .021) and stroke (p = .032). Unexpectedly, metformin did not significantly reduce microvascular complications. More surprisingly, the combination of metformin and SU caused a significant increase in the risk of diabetes-related death (96%, p = .04) and all-cause mortality (60%, p = .04) compared to SU treatment alone. Metformin was added to obese and normal weight patients who were initially assigned to SU and had inadequate responses to SU treatment (fasting plasma glucose 6.1–15.0 mmol/L) (42). Metformin was also compared to glipizide in a double-blind RCT that included non-overweight patients with T2DM and CAD. The results demonstrated that metformin significantly reduced the risk of developing macrovascular events [hazard ratio (HR) 0.54, 95% CI 0.30–0.90] but had no significant effect on all-cause mortality (43). However, the clinical significance of this high-quality RCT is limited due to the almost universally accepted recommendations of the guidelines to use metformin as the initial drug for T2DM therapy (44). Interestingly, as metformin demonstrated an increased risk of mortality when used in combination with SU, compared to SU alone, the similar effects of metformin and insulin could not be excluded. Fortunately, the HOME trial demonstrated that compared to insulin alone, metformin in combination with insulin has beneficial effects in reducing macrovascular complications (45). In contrast to the above-mentioned RCTs, a meta-analysis of the RCTs that had durations of ≥52 weeks showed that overall metformin did not have any effect on macrovascular events and all-cause mortality in T2DM. A significant reduction of macrovascular complications with metformin in dysglycemia was only evident in trials comparing metformin versus placebo or no therapy, but not in the active-comparator trials. Additionally, the authors highlighted a potentially harmful effect of the combination of metformin and SU, as shown by the UKPDS study; thus, further investigations on this matter were recommended (46).

Sulfonylureas and insulins

SU and insulin are classes of antidiabetic drugs with long histories in the treatment of T2DM. Several RCTs have evaluated their effects on macrovascular complications in T2DM (36,41–43). The UKPDS 33 trial compared intensive glucose control with SU (chlorpropamide, glibenclamide, or glipizide) or insulin treatment to conventional therapy (dietary restriction). The results indicating any reduction between the groups in macrovascular complications and mortality were not statistically significant (36). However, the post-interventional 10-year follow-up of the UKPDS survivor cohort demonstrated a positive correlation between the intensive glucose control with SU or insulin compared to the conventional therapy in terms of the reduction of MI and mortality (47). Although, the results of that study cannot be considered to be sufficient evidence due to the observational study design. A comparison of insulin to either a conventional therapy (diet alone) or SU (chlorpropamide or glibenclamide) did not demonstrate any reduction in macrovascular complications during T2DM treatment (36). Similar results were documented in the ORIGIN trial, which compared basal insulin with standard methods of care (any antidiabetic drug on the basis of the investigator’s best judgment and local guidelines) (41). In those trials, the insulin usage also failed to show a beneficial effect in terms of all-cause mortality in T2DM and dysglycemia. The UKPDS 33 trial enrolled patients with newly diagnosed T2DM (911 in the insulin group, 619 in the chlorpropamide group, 615 in the glibenclamide group, and 896 in the conventional therapy group) and had a median follow-up of 11.1 years. The ORIGIN trial enrolled patients with dysglycemia (6264 in the insulin group and 6273 in the standard care group), of which the majority (∼82%) were T2DM patients. The ORIGIN trial had a shorter follow up (6.2 years) compared to the UKPDS 33 trial but both trials included patients who were very well controlled, with baseline median HbA_1c values between 6.1–6.4% in both the insulin and control groups (36,41). Because these values were below the level at which insulin therapy usually starts, these trials describe what could be considered an effect of early use of insulin-based treatment in T2DM. As a conclusion from the available RCTs, early use of insulin-based treatment in T2DM, though safe, does not show any beneficial effects in terms of macrovascular complications and all-cause mortality.

Thiazolidinediones

TZDs are high-efficacy antidiabetic drugs (mean HbA_1c>1–2%) that increase insulin sensitivity (44). The PROactive trial, a placebo-based RCT, that assessed pioglitazone treatment in patients with T2DM and extensive macrovascular disease, demonstrated statistically significant results in terms of risk reduction for reaching the composite secondary endpoint (all-cause mortality, nonfatal MI, and stroke) (48). However, the results from the PROactive trial are compromised in several ways. The primary endpoint consisting of the composite of macrovascular complications failed to reach statistical significance. Therefore, the results of the trial regarding the secondary endpoint can only be considered exploratory and not confirmed because the primary endpoint failed to reach statistical significance (49). Furthermore, there is a difference between the predefined secondary endpoints in the published design of the study and the secondary endpoints published in the results after completion of the study (48,50). The incidence of HF, which was interestingly not defined as an endpoint but rather as an adverse event, was more frequent in the pioglitazone group than in the placebo group (HR: 1.41, p = .007) (48). Finally, the post-interventional 10-year follow-up of the trial did not document significant differences regarding macrovascular complications, with the exception of a leg amputation (51). However, pioglitazone compared to placebo demonstrated a reduction in an incidence of stroke and MI but only in patients with insulin resistance and a recent ischemic stroke or transient ischemic attack (52). At this point, the role of pioglitazone in reducing the CV risk in T2DM is still unclear. Pioglitazone is also compromised with a slightly increased risk of bladder cancer (HR 1.23, 95% CI: 1.09–1.39), according to a meta-analysis of the controlled studies (53). This risk was confirmed in a large retrospective cohort study, with an even higher incidence (HR 1.63, 95% CI: 1.22–2.19) when pioglitazone use was compared to the absence of TZD treatment. Interestingly, another TZD drug, rosiglitazone, did not demonstrate this risk (54). The RECORD trial, which compared rosiglitazone to an active comparator, did not demonstrate a difference regarding the incidence of CV hospitalization, CV death, MI, and stroke. Interestingly, rosiglitazone was associated with a higher incidence of HF, causing hospital admission or death (HR 2.10, 95% CI: 1.35–3.27) (55).

Incretin-based antidiabetics

DPP-4 inhibitors and glucagon-like peptide-1 (GLP-1) receptor agonists are incretin-based antidiabetic drugs with intermediate-efficacy (mean HbA_1c >0.5–1%) and without clinically significant risk of hypoglycemia (44). Although it has been assumed that DPP-4 inhibitors and GLP-1 receptor agonists could have a beneficial effect on the reduction of macrovascular complications and mortality, this hypothesis was unsupported by several large, multicenter, double-blinded, placebo-controlled RCTs but was supported only by recent LEADER trial with liraglutide (56–60). Unfortunately, the SAVOR-TIMI 53 trial with saxagliptin, the EXAMINE trial with alogliptin, and the ELIXA trial with lixisenatide had short median follow-ups of 2.1, 1.5, and 2.1 years, respectively. Thus, the possible longer-term benefits on macrovascular complications could not be detected (56–58). In the SAVOR-TIMI 53 trial, saxagliptin was compromised by a higher incidence of hospitalization for HF (HR 1.27, 95% CI: 1.07–1.51) and with a tendency toward an increased risk of death from any cause (HR 1.11, 95% CI: 0.96–1.27) (56). In the EXAMINE trial, the DPP-4 inhibitor alogliptin was associated with a higher incidence of hospitalization for HF (HR 1.76, 95% CI: 1.07–2.90) but, interestingly, only in a subpopulation without a history of HF at baseline (61). Recently, US Food and Drug Administration (FDA) announced new warnings regarding the increased risk of HF associated with alogliptin and saxagliptin, particularly in patients who already have CV or kidney diseases (62). Higher incidences of hospitalization for HF due to saxagliptin and alogliptin treatment have been documented in placebo-controlled RCTs, suggesting that this could be a class effect (56,61). However, lixenatide, sitagliptine, and liraglutide demonstrated a neutral effect on the incidence of hospitalization for HF (58–60). The only exception in the reduction of macrovascular complications and mortality in this class of antidiabetic drugs were documented with liraglutide, as shown by the results of the recent LEADER trial. Liraglutide demonstrated beneficial effects regarding the reduction of CV mortality and all-cause mortality in a double-blind placebo-controlled RCT with a median duration of treatment of 3.5 years and a follow-up of 3.8 years. The LEADER trial enrolled 9340 patients with T2DM and a high CV risk (Table 1). However, liraglutide demonstrated an increase in acute gallstone disease, and it seems that liraglutide has an influence on carcinomas. Notably, liraglutide reduced the incidence of prostate cancer but tended to increase the incidence of pancreatic cancer (60). These findings seem worrying, but unfortunately, the LEADER trial was not designed and did not have sufficient power to detect the influence of liraglutide on cancer risk. Thus, additional research on this matter is needed.

Table 1. Cardiovascular outcomes in LEADER trial.

Outcome	Liraglutide (N = 4668)	Placebo (N = 4672)	Hazard ratio (95% CI)	p Value
Primary composite endpoint^a	13%	14.9%	0.87 (0.78–0.97)	.01
Expanded composite endpoint^b	20.3%	22.7%	0.88 (0.81–0.96)	.005
Death from CV causes	4.7%	6.0%	0.78 (0.66–0.93)	.007
Death from any cause	8.2%	9.6%	0.85 (0.74–0.97)	.02
MI	6.3%	7.3%	0.86 (0.73–1.00)	.046
Coronary revascularization	8.7%	9.4%	0.91 (0.80–1.04)	.18
Hospitalization for HF	4.7%	5.3%	0.87 (0.73–1.05)	.14
Stroke	3.7%	4.3%	0.86 (0.71–1.06)	.16
Hospitalization for UA	2.6%	2.7%	0.98 (0.76–1.26)	.87

^a Death from CV causes, nonfatal MI, or nonfatal stroke.

^b Death from CV causes, nonfatal MI, nonfatal stroke, coronary revascularization, or hospitalization for unstable angina (UA) or HF.

Alpha-glucosidase inhibitors

At the moment, there are no trials that can prove the effect of alpha-glucosidase inhibitors on the reduction of macrovascular complications and mortality in T2DM. Acarbose has been evaluated in RCT only in patients with impaired glucose tolerance (IGT), in whom this drug demonstrated beneficial effects (63). Currently, that evidence is searching for additional confirmation through a large-scale trial (64).

Glinides

Similar to alpha-glucosidase inhibitors, glinides have been evaluated in RCTs only in patients with IGT. In the NAVIGATOR trial, nateglinide did not reduce macrovascular complications and mortality in patients with IGT and either CV risk factors or CV disease (40).

Sodium–glucose co-transporter 2 (SGLT2) inhibitors

Renal gluconeogenesis, renal glucose utilization, and the reabsorption of glucose from the glomerular filtrate into the circulation are conducted via kidneys. Thus, the kidneys play a role in the mechanisms that regulate glucose homeostasis in the euglycemic state and DM. Glucose reabsorption is conducted through active transport via sodium–glucose co-transporters (SGLTs) (65). Of the six members of the SGLTs family, SGLT2 (a high-capacity, low-affinity glucose transporter) is the most important, as this transporter is responsible for the reabsorption of ∼90% of the glucose in the kidneys (66,67). SGLT2 inhibitors are a novel insulin-independent class of antidiabetic drugs that reduce renal glucose reabsorption and, consequently, increase the urinary glucose excretion (UGE) (68). Currently, three SGLT2 inhibitors (dapagliflozin, canagliflozin, and empagliflozin) have been approved by the European Medicines Agency (EMA) and the FDA from November 2012, and several others are currently in the pipeline. In addition to their primary effect on the regulation of glycaemia, SGLT2 inhibitors have some additional properties, such as improvements in glucose perturbations and insulin sensitivity, the attenuation of intraglomerular hypertension, and the reduction of BP, body weight, body fat mass, albuminuria, and uric acid levels (69). Glycosuria is probably responsible for weight loss, the enhancement of insulin sensitivity, and the reduction of uric acid levels (70–72). The proposed mechanisms for the antihypertensive effect are osmotic diuresis, mild natriuresis, weight loss, and possible indirect effects on the nitric oxide release (73). Due to their insulin-independent mechanisms of action, SGLT2 inhibitors have a low potential to induce hypoglycemia except when used in combination with insulin or insulin secretagogues (74). The disadvantages of the SGLT2 inhibitors include polyuria, electrolyte imbalance, urinary and genital tract infections, impaired renal function, and a small increase in LDL-C (70). Nevertheless, it is clear that SGLT2 inhibitors are safe, effective, novel antidiabetic drugs, with promising roles in the treatment of T2DM.

The effects of sodium–glucose co-transporter 2 inhibitors on the macrovascular complications and mortality in type 2 diabetes mellitus

The results from the EMPA-REG OUTCOME study (a double-blind placebo-controlled RCT) presented a revolution in the treatment of T2DM. The study enrolled 7020 patients with T2DM and established CV diseases who were randomly assigned in a 1:1:1 ratio to receive either 10 mg or 25 mg of empagliflozin or placebo once a day. The median duration of treatment was 2.6 years, and the median observation time was 3.1 years. Compared to the placebo, empagliflozin demonstrated risk reductions for all-cause mortality, death from CV-related causes and hospitalization for HF. However, there were no significant differences between the groups in the incidences of MI, stroke, or hospitalization for unstable angina (UA) (Table 2). Interestingly, the majority of patients in the EMPA-REG OUTCOME trial received metformin therapy (∼74%), but in a subpopulation analysis, empagliflozin tended to have a higher efficacy considering the primary composite outcome (death from CV causes, nonfatal MI, or nonfatal stroke) and the single outcome (death from CV causes) in patients who were not taking metformin (Table 3). The most pronounced effect of empagliflozin on the reduction of the primary composite outcome was in the subpopulation with HbA_1c values <8.5%, and BMI <30 kg/m² (75). Another beneficial feature of empagliflozin is its effect on the kidneys. The recently published results showing the prespecified secondary objectives of the EMPA-REG OUTCOME trial revealed that compared to the placebo, empagliflozin was associated with a slower progression of kidney disease and lower rates of clinically relevant renal events when added to standard care (76). These effects of empagliflozin on the kidneys seem to be of great value considering that the FDA published warnings indicating that the risk of acute kidney injury is associated with the usage of two other members of the SGLT2 class of drugs (canagliflozin and dapagliflozin) (77). There was an increased rate of genital infection among patients receiving empagliflozin, but no increase in other adverse events. It is important to note that the beneficial effects were documented in a population with established CV diseases that were well treated with the use of RAAS inhibitors, statins, and aspirin. Furthermore, there were no significant differences in the baseline characteristics between the pooled empagliflozin groups and the placebo group, including the usage of antidiabetic and CV drugs. All these facts make the results of this trial very valuable for clinical practices.

Table 2. Cardiovascular outcomes in EMPA-REG OUTCOME trial.

Outcome	Empagliflozin (N = 4687)	Placebo (N = 2333)	Hazard ratio (95% CI)	p Value
Primary composite endpoint^a	10.5%	12.1%	0.86 (0.74–0.99)	<.001^c  .04^d
Secondary composite endpoint^b	12.8%	14.3%	0.89 (0.78–1.01)	<.001^c .08^d
Death from CV causes	3.7%	5.9%	0.62 (0.49–0.77)	<.001
Death from any cause	5.7%	8.3%	0.68 (0.57–0.82)	<.001
Hospitalization for HF	2.7%	4.1%	0.65 (0.50–0.85)	.002
MI	4.8%	5.4%	0.87 (0.70–1.09)	.23
Stroke	3.5%	3.0%	1.18 (0.89–1.56)	.26
Hospitalization for UA	2.8%	2.8%	0.99 (0.74–1.34)	.97

^a Death from CV causes, nonfatal MI, or nonfatal stroke.

^b Death from CV causes, nonfatal MI, nonfatal stroke, or hospitalization for UA.

^c Hypothesis was non-inferiority.

^d Hypothesis was superiority.

Table 3. Hazard ratios for the primary outcome considering metformin use in the EMPA-REG OUTCOME trial.

	Patients with event/patients analyzed
Metformin	Empagliflozin	Placebo	Hazard ratio (95% CI)	p Value for interaction
Primary outcome^a
No	146/1228	93/599	0.72 (0.56, 0.94)	.14
Yes	344/3459	189/1734	0.92 (0.77, 1.10)
CV death
No	54/1228	53/599	0.46 (0.32, 0.68)	.07
Yes	118/3459	84/1734	0.71 (0.54, 0.94)

^a Death from CV causes, nonfatal MI, or nonfatal stroke.

The pooled analysis which included nine phase 2 and phase 3 trials involving canagliflozin and the interim analysis from the ongoing CANVAS trial, documented no statistically significant risk reduction in the 4-point MACE (composite of death from CV causes, nonfatal MI, nonfatal stroke, or hospitalization for UA). Interestingly, canagliflozin showed a tendency toward higher stroke incidence (HR 1.46, 95% CI: 0.83–2.58), while the interim analysis of the CANVAS trial documented increased risk of CV events in the early treatment period (during the first 30 days of treatment). However, it was a small number of events (78).

Pooled analysis of the 14 phase 2b/3 trials with dapagliflozin showed that this drug only tended to reduce the 4-point MACE. Unfortunately, the data pertaining to the early treatment period are unavailable (79).

Discussion

The choice of antidiabetic drugs in the treatment of T2DM should be individualized and based on the patient’s characteristics and the pathohistological mechanism underlying the disease (beta cell dysfunction or insulin resistance). The aim of T2DM treatment should be the reduction of macrovascular complications and mortality. The reduction of microvascular complications should specifically not be overlooked and undervalued, because it can have a significant effect on the quality of a patient’s life. Therefore, T2DM treatment should be based on a multifactorial approach, as treating other factors such as hypertension, dyslipidemia, and aggregation that influence CV outcomes share equal importance with regulating glycaemia. Out of many antidiabetic drugs available for the treatment of T2DM, none have been proven by large, double-blinded placebo-controlled RCTs to have the clear potential to reduce macrovascular complications and mortality until recently, when empagliflozin and liraglutide were shown to have these effects. Although evidence of the effects of metformin on reducing macrovascular complications and mortality in T2DM are still limited, metformin was the only drug that showed some evidence of a beneficial effect on the clinical endpoints mentioned above. Metformin seems to have the most pronounced beneficial effect as a first-line treatment in overweight patients with T2DM. In contrast to metformin, empagliflozin demonstrated impressive results in non-overweight patients with T2DM and established CV diseases. Empagliflozin also tended to be more effective in patients who were not taking metformin. Thus, empagliflozin should be considered as a treatment of choice in non-overweight patients with T2DM and CV diseases as well as those patients who are not receiving metformin in their therapy.

The use of insulin in T2DM treatment, though safe, is unfortunately without any beneficial effects on macrovascular complications and mortality. On the other hand, SU in combination with metformin has documented negative effects on mortality. Nevertheless, more evidence is needed to confirm the negative effects of this combination. Until then, the combination of those two antidiabetic drugs should be used with caution.

Analyzing effect of antidiabetic drugs on HF, several antidiabetic drugs, including the TZDs pioglitazone and rosiglitazone and the DPP-4 inhibitors saxagliptin and alogliptin, have been associated with increased rates of hospitalization for HF (48,55,56,61). Several additional members of incretin-based antidiabetic drugs demonstrated only neutral effects on hospitalization for HF (58–60). On the other hand, empagliflozin demonstrated impressive results in reducing hospitalization for HF in patients with T2DM and CV diseases.

Unfortunately, there are no data regarding the combined treatment of liraglutide and empagliflozin, which could reveal their possible additive effects on macrovascular complications and mortality. We hope that future investigations will provide additional evidence on this matter and improve the efficacy of T2DM treatment. Although the comparison of the placebo-controlled RCTs is not the best way to compare these two drugs, it seems that treatment with empagliflozin has better efficacy compared to treatment with liraglutide. Our opinion is based on the fact that empagliflozin required shorter time intervals and fewer patients compared to liraglutide to demonstrate its beneficial effects (7020 patients and follow-up 3.1 years versus 9340 patients and follow-up 3.8 years) in similar populations. Moreover, the HR values and the p values regarding a reduction of CV and all-cause mortality were lower in the EMPA-REG OUTCOME trial (HR 0.62, p < .001 and HR 0.68, p < .001) compared to the LEADER trial (HR 0.78, p = .007 and HR 0.85, p = .02) (60,75). Furthermore, empagliflozin demonstrated a better effect in populations with CV diseases in terms of a decreased incidence of hospitalization for HF and a better safety profile. To the detriment of liraglutide, the LEADER trial demonstrated an increase in acute gallstone disease and the insufficiently revealed effect of liraglutide on cancer risk.

The EMPA-REG OUTCOME trial presented impressive results on the reduction of CV mortality, all-cause mortality, and hospitalization for HF in patients with T2DM and established CV disease. However, the potential mechanism responsible for the beneficial effects is difficult to determine. In the trial, the empagliflozin group showed better glycaemic control compared to the placebo group, but the difference in the HbA_1c levels at week 206 was very small (7.81% versus 8.16%), and was unlikely to have an influence on the reduction of macrovascular risk and mortality (75). Empagliflozin also lowers BP, reduces body weight, reduces uric acid levels, and has a small effect on lipids (80). However, it is unlikely that these small changes in the additional properties of empagliflozin could explain the impressive results of its beneficial effects. Interestingly, empagliflozin demonstrated more clinically significant results on the reduction of mortality and hospitalization for HF in a much shorter period (2.6 years), compared to a trial, such as the 5-year long HOPE trial, which involved ramipril, a drug that acts on the CV system (75,81). In a retrospective cohort study, empagliflozin was also shown to have favorable effects on the markers of arterial stiffness and vascular resistance in patients with T2DM, which could have an impact on the CV risk reduction (82). Nevertheless, the potential mechanism of the beneficial effects of empagliflozin in terms of reducing the CV risk and mortality remains unclear; thus, further research is needed.

In subpopulation analyses, empagliflozin demonstrated heterogeneous results. Empagliflozin did not prove to be effective in reducing the primary composite endpoint (HR 1.14, 95% CI: 0.86–1.5), tending instead to reduce death from CV causes (HR 0.78, 95% CI: 0.46–1.03) in subpopulations with HbA_1c values ≥8.5%. These results also assume a higher incidence of the remaining components of the primary composite endpoint (nonfatal MI and nonfatal stroke) in this subpopulation. A similar effect, but less obvious, was observed in patients with a BMI ≥30 kg/m² (75). BMI and the HbA_1c level reduction are the primary goals of T2DM treatment, and we can expect poorer regulation of these factors in patients with a lower adherence (83). Lower adherence corresponds to skipping therapy, which can possibly simulate a brief initiation of the therapy. The hypothesis is that empagliflozin may cause acute and potentially hazardous effects in patients with a lower adherence, as a direct consequence of skipping therapy. Since there are no data regarding early treatment periods with empagliflozin, this hypothesis cannot be excluded. These hazardous effects have already been observed during the early stages of canagliflozin treatment (78).

Currently, two ongoing multicenter, double-blind, placebo-controlled RCTs are evaluating the effects of SGLT2 inhibitors on mortality and macrovascular complications in patients with T2DM and increased CV risk (84,85). The CANVAS trial with canagliflozin involves approximately 4330 participants, and it is planned to continue until the occurrence of 420 events of the 3-point MACE (CV death, nonfatal MI, nonfatal stroke) (84). The DECLARE-TIMI 58 is the largest ongoing trial involving SGLT2 inhibitors; in that trial, 17,150 participants were recruited and 1390 3-point MACE events are targeted, which are the same as in the CANVAS trial (85). These trials will provide more evidence regarding the hypothesis that SGLT2 inhibitors have a class-effect in T2DM as well as the potential mechanisms by which SGLT2 inhibitors exert their beneficial effects. Though the evidence from RCTs showing that antidiabetic drugs reduce macrovascular complications and mortality in T2DM treatment is lacking, fortunately, this fact seems to be largely historical from today’s perspective. The recent results with empagliflozin and liraglutide are going to bring changes not only in T2DM treatment but also in future standards for evaluation of antidiabetic drugs.

Disclosure statement

Dario Rahelić has served as principal investigator or co-investigator in clinical trials of AstraZeneca, Eli Lilly, MSD, Novo Nordisk, Sanofi Aventis, Solvay and Trophos. He has received honoraria for speaking or advisory board engagements and consulting fees from Abbott, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Lifescan - Johnson & Johnson, Novartis, Novo Nordisk, MSD, Merck Sharp & Dohme, Pfizer, Pliva, Roche, Salvus, Sanofi Aventis and Takeda.