Diabetes mellitus (DM) is a metabolic disease characterized by chronic hyperglycemia due to impaired insulin secretion or activity. There are four etiologic categories of diabetes mellitus, as follows: diabetes mellitus type 1 (DMT1); diabetes mellitus type 2 (DMT2); gestational diabetes; and other specific DM types. DMT2 accounts for the great majority of DM patients (95%). Diagnosis of DM is made by measuring fasting blood glucose, 2-h postprandial blood glucose and glycated hemoglobin A1c (HbA1c) determination.

Along with arterial hypertension and hyperlipidemia, DMT2 is one of the leading cardiovascular risk factors. A large meta-analysis found the presence of DMT2, independently of other risk factors, to double the risk of coronary heart disease, myocardial infarction, ischemic stroke and cardiovascular death caused by various vascular events (1). While efficient control of risk factors such as hypertension and hyperlipidemia, as well as antiaggregation therapy in DMT2 reduces the risk of macrovascular complications, and good control of the ‘glucotriad’ (target HbA1c <6.5%, fasting glycemia <6.6 mmol/L and postprandial glycemia <7.8 mmol/L) has favorable effect on microvascular complications (2), the effects of tight glycemic control on macrovascular complications reported from different studies are controversial (UKPDS, VADT, ACCORD and ADVANCE) (3-6). In some studies, tight glycemic control had favorable influence on macrovascular complications, whereas in the ACCORD study it increased mortality.

Prompted by the adverse effects of rosiglitazone in the RECORD study and subsequent meta-analyses (7-9), the US Food and Drug Administration (FDA) since 2008 and European Medicines Agency (EMA) since 2012 require, besides beneficial hypoglycemic action, appropriate clinical cardiovascular outcome trials (CVOT) and safety evidence for all antidiabetic drugs (Figure 1) (10). Assessment of the effect on cardiovascular risk is based on the results of phase 2 and 3 clinical trials for all developing antidiabetic drugs, as well as for those already on the market. On assessing the effects of antidiabetic drugs on cardiovascular risk, the two-sided confidence interval upper borderline value of 95% (95% CI) is highly relevant for the estimated risk ratio. At the upper borderline value <1.3, the drug is considered safe and can be registered without additional safety testing. If the upper borderline value of risk ratio is >1.8, the drug cannot be registered and requires additional large safety testing (CVOT). Antidiabetic drugs with the risk ratio upper borderline limit between 1.3 and 1.8 can be registered, but appropriately designed and statistically powered postmarketing trial must be conducted to demonstrate the risk ratio upper borderline limit <1.3 (10).

Cardiovascular safety of oral antidiabetic drugs is of particular importance in patients with heart failure. Croatia belongs to a group of European countries with a high cardiovascular risk and increasing prevalence of DMT2. The prevalence of heart failure in the general population of Europe and Croatia has been estimated to 2%, while 0.4% of the general population suffer from both heart failure and DMT2 (11, 12). According to data of the National Diabetes Registry (CroDiab registry), a total of 254,296 individuals aged >18 suffering from diabetes were registered in 2014, yielding a prevalence of 7.9% (13). However, previous studies have shown that even 40% of those suffering from DM have not been diagnosed with the disease, thus it is estimated that as many as 400,000 individuals or every tenth adult in Croatia have DM. The prevalence of DM is significantly higher in symptomatic patients with heart failure (12%-30%), in hospitalized patients increasing up to 40% (14, 15). There are no recent epidemiological studies or reliable data on the prevalence and incidence of heart failure in Croatia; in particular, data on heart failure in DMT2 are lacking (16). The relevant European epidemiological study reports mentioned above can quite certainly be extrapolated to Croatia. With the 4.3 million population and presuming a 2% prevalence of heart failure, about 86,000 individuals would suffer heart failure in Croatia. With the prevalence of both heart failure and DMT2 of 0.4%, both conditions would be present in about 17,200 individuals. The likely prevalence of heart failure in 250,000 DMT2 patients in Croatia is 8% or fourfold that recorded in the general population.

Heart failure is responsible for impaired quality of life and disability, is associated with high morbidity and mortality, and can be induced by any disease causing damage to the heart structure and function (11, 12, 16, 17). Heart failure and DMT2 are frequently found as comorbidities and exert unfavorable mutual effect on the natural course of both conditions. Coronary heart disease and arterial hypertension as potent risk factors for heart failure have a high prevalence in diabetic patients. Hyperglycemia per se has adverse effect on myocardium, increasing the risk of myocardial dysfunction. Diabetic cardiomyopathy, a term denoting a specific clinical entity, includes numerous pathophysiological mechanisms of myocardial damage in DMT2, e.g., accumulation of advanced glycation end products, oxidative stress, inflammatory reaction, impaired intracellular calcium metabolism, altered microRNA expression, atherosclerotic lesion promotion, and development of coronary heart disease. And vice versa, the very presence of heart failure increases the risk of diabetes development due to hypersympathetic tone, pancreas and liver hypoperfusion and congestion, insulin resistance, and reduced physical activity (2, 18).

Metformin therapy of overweight DMT2 patients for 10 years after the well-known United Kingdom Prospective Diabetes Study (UKPDS) significantly reduced all diabetes dependent adverse outcomes by 21% (p=0.01), myocardial infarction by 33% (p=0.005) and overall mortality by 27% (p=0.002) (3, 19, 20). Such a notable reduction of cardiovascular events made metformin the first drug of choice in overweight DMT2 patients. A large meta-analysis confirmed the favorable effect of metformin on cardiovascular events and mortality, in particular in long-term therapy and in younger patients (21). However, caution was warranted due to the possible adverse effects of a combination of metformin and sulfonylurea agents.

Earlier, metformin was considered to be contraindicated in heart failure for fear from lactic acidosis, but later it showed reduction in total mortality, number of hospitalizations and adverse events (22, 23). In a comparator study with other oral hypoglycemic drugs and insulin, metformin as monotherapy reduced mortality by 35% and in combined therapy by 28%, whereas other drugs without metformin had neutral effects (24). Besides decreasing hyperglycemia, metformin acts favorably on dyslipidemia and reduces platelet aggregation, plasminogen activator inhibitor-1 (PAI-1) activity, endothelial dysfunction and chronic vascular inflammation (25). In their specific study, Masoudi et al. demonstrated lower prevalence of lactic acidosis in patients on metformin as compared with control group (2.3% vs. 2.6%) (26). In a systematic Cochrane analysis of 347 prospective comparator and observational cohort studies, the prevalence of lactic acidosis in metformin treated patients was 4.3/100,000 patients versus 5.4/100,000 patients in the non-metformin group (27). The risk of lactic acidosis is increased in patients with impaired renal function and estimated glomerular filtration rate (eGFR) <50, and metformin should be avoided at eGFR <30. In diabetic patients with eGFR 30-50, metformin should be used with caution, and decision on therapy depends on other characteristics of each individual patient (28).

Sulfonylureas and sulfonylurea analogues (meglitinides) are the oldest group of oral antihyperglycemics. These agents stimulate insulin secretion by blocking the adenosine triphosphate (ATP) sensitive potassium channels on the islands of Langerhans β cells and therefore are called insulin secretagogues. They represent the second line of DMT2 treatment when monotherapy with metformin, glitazones, dipeptidyl peptidase-4 (DPP-4) inhibitors or glucagon-like peptide-1 (GLP-1) analogues fails to produce satisfactory glucotriad control. These agents are administered as monotherapy in case of contraindications to other antidiabetic drugs.

Cardiovascular safety of these agents has been a subject of debate for years now, i.e. since the publication of the results of the known University Group Diabetes Programme (UGDP) study (29), in which a higher cardiovascular risk was recorded in diabetic patients treated with appropriate diet and tolbutamide as compared with those treated with diet alone. Later generally retrospective cohort studies have reported contradictory results. In the large UKPDS trial (3), no increase was recorded in cardiovascular risk, whereas an opposite conclusion has been reported from some other studies. It has been concluded that besides the effect of the group of drugs, there are differences in the action of particular drugs. The cardiovascular risk increase may be consequential to the anti-vasodilatory action due to blocking the ATP sensitive potassium channels on coronary arteries and the proarrhythmic action on the myocardium. Studies performed on animal myocardium models found glibenclamide to increase the risk of extrasystole, tachyarrhythmia and fibrillation in ischemic conditions, whereas gliclazide had protective action in baseline and ischemic conditions (30-32). A retrospective cohort study including 5631 diabetic patients found an annual incidence of heart failure of 4.4/100 diabetic patients treated with sulfonylureas versus 3.3/100 diabetic patients on metformin; the risk was dose dependent [hazard ratio (HR) 1.38; 95% confidence interval (CI) 1.20-1.60] (33). According to the observational study in diabetic patients on combined therapy with insulin secretagogues and metformin, mortality was higher in the glibenclamide group as compared with the groups on repaglinide, gliclazide or glimepiride (34). Similar results have also been reported from a study in 107,806 diabetic patients, where total mortality was higher in the groups treated with glimepiride, glipizide, glibenclamide or tolbutamide as compared with those on metformin, but there was no statistically significant difference for gliclazide and repaglinide groups in comparison with metformin group (35). As some of these studies suffered from certain drawbacks in design, it is not clear whether the increased cardiovascular risk was a consequence of sulfonylurea action or there was a protective effect of metformin (36, 37).

Thiazolidinediones (TZD, PPARγ agonists) increase insulin sensitivity of skeletal muscle and reduce hepatic glucose production through activation of the peroxisome proliferator-activated receptor γ (PPARγ), thus acting favorably on glycemic regulation, in obese DMT2 patients in particular (38). These agents do not increase the risk of hypoglycemia and have longer action than metformin and sulfonylureas (39). An unfavorable property of TZD is their ‘aldosterone’ effect in distal and collecting tubules of the kidney, where they cause sodium and water reabsorption, thus increasing the risk of edema and manifest heart failure in diabetic patients with asymptomatic left ventricular dysfunction (2, 24). Meta-analyses of clinical trials with rosiglitazone found an increased risk of all macrovascular events, myocardial infarction in particular, in comparison with various comparators (8, 9). In the RECORD study, rosiglitazone increased the risk of nonfatal and fatal heart failure significantly, i.e. in the intent-to-treat analysis by 110% (p<0.001) and in the per-protocol + 30 days analysis by 91% (p=0.013) (7). Therefore, EMA suspended rosiglitazone from the European market, whereas on the US market the use of rosiglitazone is approved under a special prescription program.

In contrast to rosiglitazone, pioglitazone showed beneficial effects on preventing cardiovascular events. In the PROactive study, pioglitazone in comparison with placebo reduced the combined secondary outcome (total mortality, myocardial infarction and stroke) in high-risk diabetic patients with macrovascular disease by 16% (p=0.027) (40). The risk of several isolated secondary outcomes was also reduced significantly, as follows: stroke by 47% (p=0.008); recurrent acute coronary syndrome by 37% (p=0.035); and myocardial reinfarction by 28% (p<0.045) (41, 42). In a meta-analysis of the cardiovascular risk of TZD, unlike rosiglitazone, pioglitazone reduced the risk of all macrovascular events and myocardial infarction in comparison with comparators (43). The FDA meta-analysis of clinical trials of pioglitazone also demonstrated the beneficial effect of this drug on cardiovascular event reduction (44). Due to the increased risk of edema and heart failure, pioglitazone is not allowed for use in patients with NYHA grade I-IV heart failure in Europe and in those with NYHA grade III-IV in the USA. In favor of pioglitazone, it should be noted that those pioglitazone studies did not involve fatal heart failure, i.e. there was no mortality increase, while the occurrence of edema could be well controlled by diuretic therapy. Besides the PROactive study, favorable therapeutic effects of pioglitazone have also been reported from a number of small clinical trials with surrogate endpoints. In the QUARTET study, pioglitazone monotherapy proved superior to gliclazide in regulating fasting glycemia, less frequent hypoglycemia and better effect on triglycerides, high density lipoprotein (HDL) cholesterol and total cholesterol/HDL cholesterol ratio (45). In the CHICAGO study, pioglitazone, but not the comparator glimepiride, prevented progression of carotid atherosclerosis as measured by the intima/media thickness (IMT) (46). A similar anti atherosclerotic action of pioglitazone in comparison with glimepiride in the prevention of coronary atherosclerosis was found in the PERISCOPE study, in which the progression of coronary plaques was analyzed by intravascular ultrasound (IVUS) (47). Although these were small trials with surrogate endpoints, the overall results of all clinical trials of pioglitazone have confirmed the importance of this agent in the management of DMT2 patients, those overweight in particular, where peripheral insulin resistance has a major pathophysiological role (48). A fixed combination of metformin and pioglitazone is an excellent therapeutic choice in these patients, with the exception of those with seriously damaged renal function and NYHA grade I-IV heart failure.

These are a newer group of antidiabetic drugs that decrease the level of blood glucose by inactivating the dipeptidyl peptidase-4 (DPP-4) enzyme. Inhibition of this enzyme reduces incretin breakdown, thus increasing insulin secretion, decreasing glucagon secretion and slowing down gastric emptying. DPP-4 inhibitors act favorably on appetite, have neutral effect on body weight, and do not cause hypoglycemia. The first drug from the group of DPP-4 inhibitors is sitagliptin, and the others are vildagliptin, linagliptin, omarigliptin and alogliptin, some of these still in the phase of development or research (49). Cardiovascular effects of DPP-4 inhibitors in high-risk DMT2 patients have been investigated in three large randomized studies (alogliptin in EXAMINE, saxagliptin in SAVOR-TIMI 53, and sitagliptin in TECOS), while a number of similar studies are just being under way (linagliptin in CAROLINA and CARMELINA, and omarigliptin in MK-3102-015 AMI and MK-3102-018) (Figure 2) (49-52). In the large, well designed, placebo-controlled EXAMINE study, alogliptin did not increase the risk of major adverse cardiovascular events (MACE) in DMT2 patients with recent acute coronary syndrome (<90 days of randomization). A decreasing trend of cardiovascular mortality is described, however, without reaching statistical significance (49-51). Results of the EXAMINE, SAVOR-TIMI 53 and TECOS studies confirm cardiovascular safety of DPP-4 inhibitors, which neither reduced nor increased the prevalence of MACE in these three studies (49).

However, in the SAVOR-TIMI 53 study, an unexpected and significant 27% (p=0.007) increase in the rate of hospitalizations for heart failure was recorded in the group of patients on saxagliptin (52). A recent meta-analysis found no differences in MACE and total mortality between DPP-4 inhibitors and placebo but the prevalence of heart failure was by 16% greater (p=0.04) in the group of patients on DPP-4 inhibitors (53). Another meta-analysis of cardiovascular safety of DPP-4 inhibitors has also pointed to a comparable increase in the rate of heart failure (54). In the EXAMINE study, the 19% increase in hospitalizations for heart failure recorded in the group of patients on alogliptin was not statistically significant (p=NS) and subsequent post-hoc analysis revealed no difference between alogliptin and placebo in the combined endpoint of hospitalization for heart failure and cardiovascular mortality (55). In the TECOS study, sitagliptin did not increase the rate of heart failure; this variation from other studies with DPP-4 inhibitors in the prevalence of heart failure could be explained by differences in the characteristics of study populations, other therapies, definition and recording of heart failure, intrinsic pharmacological differences among particular DPP-4 inhibitors, or just as mere coincidence (56).

Anyway, the absolute risk of heart failure with DPP-4 is low, associated with other frequently taken drugs (sulfonylurea derivatives and thiazolidinediones) and still a controversial issue. Results of the large randomized studies that are under way (CAROLINA, CARMELINA, MK-3102-015 AMI and MK-3102-018) and future clinical trials will certainly contribute to better understanding of the cardiovascular effects and safety of DPP-4 inhibitors^49.

Subtype 2 sodium-glucose transport (SGLT-2) inhibitors are a relatively novel group of oral antihyperglycemics, so as yet there are little data on their cardiovascular safety. SGLT-2 is a transmembrane protein performing sodium dependent glucose reabsorption and is responsible for about 90% of overall glucose reabsorption in proximal renal tubule. This new group of drugs stimulate renal excretion of glucose by inhibiting this protein activity, thus reducing hyperglycemia, increasing desirable total calorie deficit, stimulating osmotic diuresis and lowering arterial pressure, thus eventually reducing the cardiovascular risk. The efficacy of SGLT-2 inhibitors has been investigated in a number of studies. In the CANTATA-SU study, which included 1452 diabetic patients that failed to achieve satisfactory glycemia control on metformin (mean HbA1c 7.8%), the efficacy of add-on canagliflozin in a dose of 100 mg or 300 mg was compared with glimepiride (mean dose 5.6 mg). In both canagliflozin groups, HbA1c reduction was similar to that recorded with glimepiride, and was somewhat better in the group on a higher dose of canagliflozin (0.81% and 0.82%, respectively, vs. 0.93%). Canagliflozin had a beneficial effect on weight loss (-4.2 to -4.4 kg) as compared with glimepiride (+0.8 kg), but with a higher prevalence of genital fungal infections (57). According to a meta-analysis that included data on 2313 diabetic patients, 1332 of them on antihypertensive therapy, canagliflozin in doses of 100 mg and 300 mg decreased systolic blood pressure by a mean of 4.3 mm Hg and 5.0 mm Hg, respectively, in comparison with placebo. The respective diastolic blood pressure decrease was 2.5 mm Hg and 2.4 mm Hg versus 0.6 mm Hg on placebo. Greater arterial pressure decrease was recorded in the group of hypertensive than in normotensive diabetic patients (58).

The EMPA-REG OUTCOME study assessed cardiovascular safety of empagliflozin during a mean 3.1-year treatment. The primary composite endpoint consisted of cardiovascular mortality, nonfatal myocardial infarction and stroke. Empagliflozin was superior to placebo in primary outcome reduction [10.5% vs. 12.1%; hazard ratio (HR) 0.86; 95% confidence interval (CI) 0.74-0.99; p=0.0382]. Analysis of particular outcomes revealed empagliflozin to have significantly decreased cardiovascular mortality by 38% (p<0.0001), rate of hospitalization for heart failure by 35% (p=0.0017) and total mortality by 32% (p<0.0001) (59).

A meta-analysis of 21 phase 2b and 3 studies investigated the effect of dapagliflozin on MACE. Upon patient stratification for additional risk factors for MACE, it was concluded that dapagliflozin did not increase the risk of MACE either in diabetic patients without [hazard ratio (HR) 0.77; 95% confidence interval (CI) 0.54-1.10] or with additionally increased cardiovascular risk [hazard ratio (HR) 0.80; 95% confidence interval (CI) 0.52-1.22] (60).

A number of studies on cardiovascular safety and efficacy of SGLT2 inhibitors are just being under way (REFORM, CANVAS, CREDENCE, and DECLARE-TIMI 58) (Figure 2) (61-64).

The possible side effects of gliflozin include hypoglycemia (more frequently in the population of diabetic patients treated with a combination of gliflozin and sulfonylureas or insulin), renal function worsening, orthostatic hypotension and urogenital infections. Several cases of urosepsis and euglycemic ketoacidosis have also been described. Therefore, a specific warning was issued by FDA in May 2015 and by EMA in July 2015. Namely, SGLT2 was expressed on α cells of the islands of Langerhans and its inhibition stimulates glucagon secretion; the more so, ketone transporters in the kidney may also be inhibited, which would increase the level of ketone bodies in the blood (65). The risk of ketoacidosis is greater in patients treated with a combination of gliflozin and metformin. SGLT2 inhibitors are not recommended at eGFR <60 mL/min/1.73 m². If eGFR falls below 60 upon drug introduction, the dose should be decreased and at eGFR <45 the drug should be discontinued. In patients with heart failure, the use of gliflozin is limited by the use of diuretics and mineralocorticoid receptor antagonists due to the higher risk of orthostatic hypotension, renal function impairment and hyperkalemia (65).