The pleiotropic cardiovascular effects of dipeptidyl peptidase‐4 inhibitors

Abstract

Patients with Type 2 diabetes have an excess risk for cardiovascular disease. One of the several approaches, included in the guidelines for the management of Type 2 diabetes, is based on dipeptidyl peptidase 4 (DPP‐4; also termed CD26) inhibitors, also called gliptins. Gliptins inhibit the degradation of glucagon‐like peptide‐1 (GLP‐1): this effect is associated with increased circulating insulin‐to‐glucagon ratio, and a consequent reduction of HbA1c. In addition to incretin hormones, there are several proteins that may be affected by DPP‐4 and its inhibition: among these some are relevant for the cardiovascular system homeostasis such as SDF‐1α and its receptor CXCR4, brain natriuretic peptides, neuropeptide Y and peptide YY. In this review, we will discuss the pathophysiological relevance of gliptin pleiotropism and its translational potential.

Introduction

Patients with Type 2 diabetes (DMT2) have an excess risk for cardiovascular disease (CVD) 1: as a rule, the normalization of glucose, blood pressure, lipid profile and body weight is considered a priority in the treatment of patients affected by DMT2. One of the several approaches, included in the National Institute for Health and Care Excellence Guidelines for the management of DMT2, is based on dipeptidyl peptidase 4 (DPP‐4; also termed CD26) inhibitors (DPP‐4‐I), also called gliptins 2. Gliptins and the glucagon‐like peptide‐1 receptor agonists (GLP‐1RA) are broadly defined as incretin‐based therapy. In the seminal definition of Creutzfeldt, incretins (intestin secretion insulin) are gut‐derived hormones that increase glucose‐stimulated insulin secretion 3. This description implicates three important concepts: (i) gut hormones cosecrete with insulin in response to a meal; (ii) gut hormones potentiate insulin release in response to a meal; (iii) the contribution of gut hormones to insulin secretion is exerted only when glucose levels are increased. The two most important incretin hormones are glucose‐dependent insulinotropic polypeptide (GIP) and GLP‐1. More than 70% of insulin release in response to oral glucose is determined by Incretin hormones 4. In patients with DMT2 there is almost a completely loss of the insulinotropic response to GIP. Although GLP‐1 maintains significant insulinotropic activity, GLP‐1‐induced insulin secretion is reduced in DMT2 patients compared with healthy individuals 5.

The role of dipeptidyl peptidase‐4

Fifty percent of GLP‐1 is degraded in ~1 min and 50% of GIP in ~7 min by the DPP‐4 or CD26, a 110 kDa peptidase that belongs to a unique class of membrane associated peptidases: the DPP family, which also includes DPP6, DPP8, DPP9, fibroblast activation protein and prolyl endopeptidase (PEP) 6. The DPP‐4 gene encodes a 766 amino acids protein, which has a high cleavage selectivity for peptides with a proline or alanine at the second position and cleaves dipeptides at the N‐terminus of such peptides. The soluble form of DPP‐4, which is the main form targeted by the DPP‐4‐I, lacks the intracellular tail and transmembrane regions; it is found at high levels in seminal fluid, and moderate and low amounts are present in plasma and cerebrospinal fluid, respectively 7. DPP‐4 is strongly expressed on epithelial cells (kidney proximal tubulus, intestine, bile duct) and distributed throughout the body, with particularly high expression on the apical surface of endothelial and differentiated epithelial cells. Mice deleted for Dpp4 are fertile and appear healthy and partially resistant to the development of obesity and hyperinsulinaemia and demonstrate improved postprandial glucose control 8.

DPP‐4‐I and GLP‐1 degradation

Native GLP‐1 (_7‐36), is degraded within minutes by DPP‐4: the DPP‐4‐cleaved form of GLP‐1 (_9‐36), is the predominant circulating form of the hormone, but it has no significant insulin stimulatory role 9. Rather, GLP‐1 (_9‐36) exerts positive effect on myocardial function, such as improving left ventricular performance 10, protecting against ischaemia–reperfusion injury 11, and vasodilation 12. Another GLP‐1 metabolite (_28‐36), has remarkably positive effects on intrahepatic lipid metabolism by limiting free fatty acid oxidation and gluconeogenesis 13. DPP4 inhibition should theoretically prevent the positive effects of GLP‐1 cleavage products. Data in support of this hypothesis are scarce. For instance, sitagliptin does not interfere with the positive myocardial effect of GLP‐1 (9‐36) in isolated cardiomyocites 14.

The administration of DPP‐4‐I avoids the rapid degradation of native GLP‐1 by DPP‐4, and, accordingly, increases its circulating concentration in states of reduced incretinergic tone, such as in patients with DMT2. This is important also because we have found that plasma DPP‐4 activity of patients with DMT2 is significantly higher (~33%) than in controls and is not lowered by intensified glucose control 15. The inhibition of circulating DPP‐4 leads to a ~3‐fold increase in the concentration of active GLP‐1 and decreases HbA1c of ~0.5–0.7% 16. DPP‐4‐I are neutral or minimally active on body weight, blood pressure, and lipid profile. The net effect of DPP‐4 inhibition on the potential beneficial actions of GLP‐1 cleaved forms on the cardiovascular (CV) system are largely unknown.

The commercially available DPP‐4‐‐I are sitagliptin (Merck), alogliptin (Takeda), linagliptin (Boehringher Ingelheim), saxagliptin (BMS/Astra Zeneca), and vildagliptin (Novartis). Other DPP‐4‐I are omarigliptin (Merck), tenagliptin (Steris Healthcare), anagliptin (generic available in Japan), gemigliptin (LG Life Science). Sitagliptin, alogliptin and linagliptin form noncovalent interactions with residues in the catalytic site so that they dissociate rapidly from the enzyme, while saxagliptin and vildagliptin form a reversible covalent enzyme–inhibitor complex in which there is a slow rate of inhibitor binding and a slow rate of dissociation 17. The different characteristics of DPP‐4‐I are reported in Table 1.

Table 1.

Pharmacodynamic properties of dipeptidyl peptidase (DPP‐4) inhibitors

Drug	Sitagliptin	Vildagliptin	Saxagliptin	Alogliptin	Linagliptin
Therapeutic dose	100 mg od	50 mg bid	5 mg od	25 mg od	5 mg od
Fold selectivity vs DPP‐8 and DPP‐9	>2600	<100	<100	>14 000	>10 000
Effect on active GLP‐1 levels	~2‐fold increase at 100 mg od	~3‐fold increase at 50 mg bid	1.5‐ to 3‐fold increase at ‡2.5 mg od	2‐ to 3‐fold increase at 25 mg (single oral dose)	4‐fold increase at 25 mg (single oral dose)
Kidney excretion	87%	85%	75%	76%	5%
Liver excretion	13%	4.5%	22%	13%	85%
Binding on enzyme	Noncovalent	Covalent	Covalent	Noncovalent	Noncovalent
Dissociation	Rapid	Slow biphasic	Slow biphasic	Rapid	Rapid

od, once daily; bid, twice daily

The pleiotropism of DPP‐4‐I

Beside incretin hormones, there are several proteins that have a penultimate alanine, proline or serine in the N‐terminus start site: therefore, they may be affected by DPP‐4 and its inhibition. These substrates are broadly divided into physiological substrates and pharmacological substrates: the former are defined as peptides whose endogenous circulating levels of intact vs. NH2‐terminally cleaved forms are altered following reduction or elimination of DPP‐4 activity in vivo 18. The latter are substrates NH2‐terminal cleaved by DPP‐4 in vitro, the in vivo relevance of which is unknown. In this review, we will highlight the effect of DPP‐4 inhibition on those substrates, which may be relevant for the CV system.

DPP‐4 inhibition and stromal cell‐derived factor‐1α

Stromal cell‐derived factor‐1α (SDF‐1α) and its receptor CXCR4 play a prominent role in the trafficking and homing of haematopoietic stem cells: SDF‐1α is encoded by the CXCL12 gene located on chromosome 10, and it is ubiquitously expressed in many tissues and cell phenotypes 19. The SDF‐1α receptor CXCR4 is highly expressed on different population of primitive cell populations such as human CD34+ haematopoietic stem cells, which migrate towards SDF‐1α gradients 20. SDF‐1α/CXCR4 signalling pathway plays an important role in vascular and cell protection and in vascular development in multiple organs.

Full‐length SDF‐1α _1–68 undergoes processing first at the COOH terminus to produce SDF‐1α _1–67, and then it is cleaved by DPP‐4 at the NH2 terminus to produce SDF‐1α 3–67 21. The decreased oxygen tension, present in the bone marrow environment, enhances expression of SDF‐1α by stromal cells, thus inducing a retentive effect on stem cells 22. SDF‐1 levels increase in both plasma and ischaemic tissue shortly after ischaemic injury, in response to hypoxia, which upregulates hypoxia inducible factor (HIF)‐1α 23. HIF‐1α upregulates SDF‐1 α, by binding to the promoter of SDF‐1, and initiating its transcription. Administration of a CXCR4 antagonist, such as AMD3100 (plerixafor), by disrupting the SDF‐1α/CXCR4 pathway, leads to a sustained mobilization of CD34+ haematopoietic stem cells to the peripheral blood 24. Also, the administration of granulocyte colony‐stimulating factor (G‐CSF) induces a 10‐fold to 100‐fold increase in the level of circulating CD34+ haematopoietic stem cells in humans by decreasing SDF‐1α concentrations in the bone marrow, thereby generating a gradient of SDF‐1α toward the circulation 25, 26. G‐CSF also increases the enzymatic N‐terminal cleavage of SDF‐1α by DPP‐4 in the marrow, resulting in the release of CXCR4+ stem cells from the bone marrow into the blood 27. After an episode of acute ischaemia, such as an acute coronary syndrome, SDF‐1α is released from damaged cells resulting in mobilization and recruitment of progenitor cells through its gradients. In this context, it has been shown that DPP‐4 contributes to SDF‐1α degradation after myocardial infarction: DPP‐4 inhibition blocks the degradation of SDF‐1α, which, in turn, allows a much more efficient recruitment of progenitor cell in the site of ischaemia 28. While there is considerable positive experimental evidence that DPP‐4 inhibition of SDF‐1α cleavage is protective in the ischaemia–reperfusion injury, there is a very limited evidence in humans 29.

The hypothesis that DPP‐4 inhibition prevents the degradation of SDF‐1α, thus favouring the homing of progenitor cells with a consequent amelioration of ulcer healing, has been tested in humans in the context of peripheral artery disease and its complication. We have shown that diabetes delayed wound healing, with reduced granulation tissue thickness and vascularity, and increased apoptosis as a consequence of an increased apoptosis, and decreased proliferation of bone marrow‐derived progenitor cell 30. Marfella and associates reported that in patients with DMT2 and at least one full‐thickness wound below the ankle, the treatment with the DPP‐4‐I vildagliptin leads to a more rapid wound closure rate at week 12 than in controls, and doubling in complete healing of the index ulcer 31. In a retrospective analysis conducted in 82 169 propensity score‐matched pairs of DPP‐4 inhibitor users and nonusers with DMT2, DPP‐4 inhibitor users were associated with a lower risk of both peripheral arterial disease and risk of lower‐extremity amputation than nonusers 32. More recently, Long and colleagues have tested the hypothesis that DPP‐4‐I can improve diabetic wound healing, independent of their beneficial effects on glycaemic control. They showed that these drugs promote the migration and epithelial–mesenchymal transition of keratinocytes both directly, and indirectly by inducing SDF‐1α production of fibroblasts in vitro and in diabetic mice 33.

A combined strategy using both G‐CSF and DPP‐4 inhibition has been proposed to improve cardiac homing of mobilized stem/progenitor cells, after acute myocardial infarction: the inhibition of DPP‐4 improves homing of mobilized stem/progenitor cells, while G‐CSF will enhance release of stem cells from the bone marrow 28, 34, 35. These data suggest that the SDF‐1–CXCR4 axis may be particularly important especially in patients with DMT2 and CVD. Diabetes is known to impair stem cell mobilization from the bone marrow, associated with altered regulation of SDF‐1α and inability to upregulate muscle HIF‐1α 36, 37. Along with the reduced mobilization of stem cells, diabetes features a reduction of bone marrow‐derived progenitor cells, which show increased apoptosis and decreased proliferation compared to nondiabetic wound tissues 38, 39. The disturbed kinetic of CD34+ cells between bone marrow and blood was associated with aging and diabetes, and strongly related to high activity of DPP‐4 40. Therefore, diabetes affected DPP‐4 activity both in blood and in the bone marrow, with a consequent impairment of stem/progenitor cell mobilization after ischaemia or G‐CSF administration 41. For these reasons, in a small, controlled, nonrandomized clinical trial, we tested the hypothesis that a DPP‐4 inhibitor, sitagliptin, would improve the progenitor cell mobilization 42. After 4 weeks, we showed that, as compared with control subjects, patients receiving sitagliptin showed a significant increase in endothelial progenitor cells and SDF‐1α. This finding has been replicated in crossover trial using the DPP‐4‐I linagliptin administered for 4 days in patients with DMT2 43.

Thus, DPP‐4 inhibition in patients with DMT2 seems to restore the ability of the bone marrow to release progenitor cells. The importance of this observation is two‐fold. First, diabetes is characterized by a diabetic stem cell mobilopathy determined, at least in part, by the presence of a bone marrow microangiopathy 44, 45 and by the maladaptive regulation of CXCR4/SDF‐1. Second, both age and risk factors, such as hyperglycaemia, independently associate with the blood levels of progenitor cells, with lower levels in older subjects and those with higher risk factor burden or CVD 46. One may thus postulate that the chronic administration of DPP‐4 inhibitor would allow a sustained capacity of the bone marrow to release progenitor cells in response to bouts of hypoxia also in elderly individuals with CVD.

The hypothesis that the DPP‐4 inhibition may have a positive effect on vascular disease by avoiding DPP‐4‐mediated cleavage of SDF‐1α is not unanimously accepted. Activation of CXCR4 receptor may enhance thrombopoiesis in humans 47, and may contribute to plaque progression by perturbing neutrophil adhesive capacity, thus confirming the complexity of this chemokine in the context of the pathophysiology of vascular disease 48.

SDF‐1α plays also an important role in kidney homeostasis: data on this issue are conflicting, and particularly scarce on the role of DPP‐4 inhibition. In general, upregulation of SDF‐1‐CXCR4 axis exerts a protective effect on kidney function by decreasing oxidative stress, profibrotic effect and ischaemia 49, 50, 51. There is ample evidence that the administration of DPP‐4‐I in humans exerts positive effect on markers of kidney function, especially albuminuria, beyond their ability to decrease plasma glucose 52, 53, 54.

DPP‐4 inhibition may also affect natriuresis in both animals and humans 55, 56. In humans, one month of sitagliptin treatment increases circulating levels of SDF‐1α _1–67, and induces natriuresis by blocking distal tubular sodium reabsorption, distal to the macula densa, without affecting renal haemodynamic or blood pressure 57. Potential distal tubular ion transport channels linking DPP‐4 inhibition to distal natriuresis may involve the Na⁺/Cl⁻ thiazide‐sensitive channel, and the epithelial sodium channel 4. All things considered, the effect of DPP‐4 inhibition on SDF‐1/CXCR4 axis affects kidney function in several ways; nonetheless there is no information regarding the interactions between this inhibition and those with other DPP‐4 sensitive substrates in the overall regulation of renal function.

Brain natriuretic peptides and DPP‐4‐I

In patients with heart failure (HF), cardiac brain natriuretic peptidase (BNP) secretion increases in response to congestion, such as fluid overload, and increased afterload following neurohumoral stimulation by angiotensin II and endothelin. The role of BNP is particularly relevant in patients with diabetes, which represents one of the major risk factors for HF, being associated with a >50% increase in risk and playing a relevant prognostic effect 58. Human BNP is synthesized as a 134‐amino acid precursor protein, which is subsequently processed during secretion to form the 108‐aa peptide, proBNP 59. On the extracellular surface of cardiomiocytes, proBNP interacts with corin to form the active BNP _1–32 and NT‐proBNP _1–76; there is also a circulating proBNP _1–108. All these forms of BNP are substrates for DPP‐4 since they all have a proline in the second N‐terminal position. DPP‐4 cleaves BNP _1–108 to BNP _3–108, BNP _1–32 to BNP _3–32, and NT‐proBNP _1–76 to NT‐proBNP _3–76. BNP _3–32 exerted a reduced natriuretic than BNP _1–32. BNP concentrations are reduced in people with obesity, insulin resistance, and diabetes, and this deficiency may contribute to their CV risk 60. In theory DPP‐4 inhibition could exert beneficial effects on cardiac function, by increasing the proportion of circulating BNP _1–32, and NT‐proBNP _1–76. In a randomized, single‐blinded, placebo‐controlled, cross‐over study, we tested whether the DPP‐4 inhibitor linagliptin could affect the circulating BNP and pro‐BNP concentrations 61. We found no significant changes in BNP and NT‐proBNP levels after treatment with linagliptin or placebo in patients with or without CKD. Our results, however, do not rule out the possibility that DPP‐4 inhibition may exert a positive effect on BNP response in patients with a history or ongoing HF. Similar results have been obtained by Devin and associates showing that, in healthy subjects, sitagliptin increased venous GLP‐1 concentrations without affecting BNP concentrations 62. It has indeed been shown that, in these patients, an inverse correlation was observed between serum DPP‐4 activity and left ventricular ejection fraction, thus suggesting that DPP‐4 inhibition may alleviate the pathophysiological mechanisms leading to HF independently of its ability to modify BNP circulating concentrations 63.

Neuropeptide Y and Peptide YY

Neuropeptide Y (NPY) and peptide YY (PYY) belong to pancreatic polypeptide (PP) family involved in the neuroendocrine control of feeding associated processes 64. These polypeptides bind to receptors identified as NPY receptors. There are six identified NPY receptors: Y1‐Y6 belong to the G protein‐coupled receptor family except for Y3. Vessels wall as well as platelets release NPY and norepinephrine together in the presence of sympathetic nerves over‐activity. The endothelial cells express NPY–Y1, Y2 and Y5 receptors (Figure 1); under pathological conditions, NPY binds to Y1 receptor to induce vasoconstriction, and favouring the norepinephrine vasoconstriction with a consequent increase in blood pressure 65. While the effects of DPP‐4 inhibition on NPY has been extensively studied in experimental models, there are fewdata in humans, and most relate to the interactions between DPP‐4 inhibition and the renin‐angiotensin system. This hypothesis has been tested by Jackson and colleagues in humans where they observed a significant attenuation of ACE inhibitor‐mediated blood pressure lowering effects by DPP‐4‐I 66, 67. Thus, both an increased sympathetic tone (induced by simultaneous ACE and DPP‐4 blockade in the central nervous system) and a reduced metabolism of NPY_1–36 to NPY_3–36 could counterbalance the antihypertensive effect of ACE inhibition. Whether or not this has significant clinical implications needs further evaluation. Based on these findings, in humans, the relationships between NPY and DPP‐4 inhibition appear, at best, complex, and determined by several factors, among which, most importantly, is the concomitant antihypertensive therapy 68.

Figure 1.

Neuropeptide Y (NPY) receptors and their effects on the cardiovascular system

PYY is a 36‐amino acid polypeptide colocalized with GLP‐1 in the endocrine L‐type cells of the gastrointestinal mucosa 69. PYY has structural homology with NPY and pancreatic polypeptide, and together form the NPY family of peptides. PYY release is stimulated by intraluminal nutrients: after its release, DPP‐4 cleaves the N‐terminal tyrosine‐proline residues, forming PYY_3–36. PYY_1–36 represents about 60% and PYY_3–36 40% of circulating PYY. However, a significant degradation from the COOH terminus has also been described 70; in humans, PYY acts through Y‐receptor subtypes: Y1, Y2, Y4 and Y5. PYY increases postprandial natriuresis and elevates blood pressure 71. Both forms of PYY cross the blood–brain barrier and modulate weight control regulation, with PYY_1–36 acting through Y1‐ and Y5‐receptors increasing weight gain and PYY_3–36 acting through Y2‐receptors decreasing body weight 72. High serum fasting PYY _3–36 concentration are associated with high body mass index, waist circumference, HbA1c, fasting blood glucose, leptin, triglyceride and serum insulin along with a low high‐density lipoprotein cholesterol concentration 73. Further studies aimed at understanding the role of PYY in CVD and the specific role of DPP‐4 inhibition are warranted.

CV outcome trials and real‐world evidence on DPP‐4‐I CV safety

From preclinical and clinical studies, one can speculate that the administration of DPP‐4‐I should, supposedly, positively affect the CV system beyond their ability to decrease plasma glucose: these purported effects should be expected based on their ability to interfere with the cleavage of SDF‐1α, BNP and neurohormones. A meta‐analysis on published randomized clinical trials has shown that DPP‐4 inhibition is associated with neutral results in term of blood pressure lowering effect 74. In this context, several meta‐analysis have been performed to assess the CV safety of these compounds as reported in the Table 2.

Table 2.

Meta‐analysis performed to identify the cardiovascular (CV) safety of dipeptidyl peptidase inhibition

First Authors	Number of patients included in analysis	Risk for CV mortality (OR 95% CI)	All‐cause mortality (OR 95% CI)	Stroke (OR 95% CI)
Elgendy et al. 81	66 730	1.02 (0.92–1.14)	1.03 (0.94–1.12)	0.99 (0.85–1.15)
Xu et al. 82	29 600	1.01 (0.91–1.12)	1.03 (0.95–1.11)	1.02 (0.88–1.18)
Savarese et al. 83	107 100	0.975 (0.887–1.073)	1.010 (0.935–1.091)	0.933 (0.820–1.062)
Abbas et al. 84	36 543	1.01 (0.91–1.12)	NA	1.00 (0.86–1.16)
Monami et al. 85	20 312	0.689 (0.528–0.899)*	0.668 (0.396–1.124)	ND

^* MACE: cardiovascular death, nonfatal myocardial infarction and stroke, and hospitalizations

due to acute coronary syndromes and/or heart failure.

OR, odds ratio; CI; confidence interval; NA, not available; ND, not determined

As can be appreciated, strong evidence for neutrality emerges regarding all the major CV end‐point as well as for all cause death.

Similarly to the previously reported meta‐analysis, in all CV outcome trials (CVOTs), the DPP‐4‐I saxagliptin, alogliptin, and sitagliptin met the primary endpoint of noninferiority vs. placebo with respect to MACE (CV mortality, nonfatal myocardial infarction, and nonfatal stroke) 75, 76, 77. Interestingly, in all three CVOTs, no significant differences were reported in those parameters such as heart rate and blood pressure, potentially modifiable by the DPP‐4 inhibition, and linked to its pleiotropic activity.

The Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus (SAVOR) randomized 16 492 patients with DMT2, with a history of or at high risk for CVD, to saxagliptin or placebo in addition to usual care. At the end of follow‐up period (median of 2.1 years), the rate of primary end‐point (a composite of CV death, nonfatal myocardial infarction or ischaemic stroke) was similar in the two groups. Interestingly, in those randomized to saxagliptin, the risk for hospitalization for HF was significantly higher 78. Several speculations have been proposed for this increased risk of HF, which was observed only for this drug of the class but not with the others. Notably, in both SAVOR and in the Examination of CV Outcomes with Alogliptin vs. Standard of Care (EXAMINE) trials, HF was not included among the prespecified end points. Recently, a randomized controlled trial specifically assessing the effect of vildagliptin on heart function in patients with DMT2 and HF, has been completed 79, and showed that 52 weeks treatment with vildagliptin 50 mg twice daily was neutral vs. placebo in left ventricular ejection fraction and geometry 79. From a recent posthoc analysis of the SAVOR trial, the risk of HF hospitalization appears tightly linked to the worsening of renal function 80. Following the results of the SAVOR trials, numerous observational studies have conducted either to confirm or rule‐out the association between DPP‐4‐I treatment and HF. As can be seen in Table 3, there have been conflicting results, which tend to exclude a class effect but to indicate saxagliptin as the only drug involved in this increased risk.

Table 3.

Results from observational studies linking dipeptidyl peptidase 4 (DPP‐4) inhibitors treatment with the risk of heart failure (HF)

First authors	Findings
Weir et al. 86	Sitagliptin was not associated with an increased risk of all‐cause hospitalizations or death, but was associated with an increased risk of HF‐related hospitalizations among patients with type 2 diabetes with pre‐existing HF.
Monami et al. 87	Increased risk of heart failure, without any clear evidence of differences among drugs of the class
Giorda et al. 88	Not associated with an increased risk of HF
Fadini et al. 89	The use of DPP‐4 inhibitors was associated with a reduced risk of HF hospitalization when compared with sulfonylureas.
Li et al. 90	Risk may be increased
McGuire et al. 91	Sitagliptin use does not affect the risk for hypertensive HF in Type 2 Diabetes
Kongwatcharapong et al. 92	Saxagliptin significantly increases the risk of HF by 21% especially among patients with high cardiovascular risk while no signals were detected with other agents
Toh et al. 93	The risk for hypertensive HF was not higher with DPP‐4 inhibitors than with the other study drugs
Savarese et al. 94	No risk
Verma et al. 95	No risk
Ou et al. 96	No signals
Yamada et al. 97	Reduced risk
Fadini et al. 98	No intraclass difference in the incidence rate of first and total hypertensive HF events between sitagliptin, saxagliptin, and vildagliptin.

There is an ongoing CVOT testing long‐term impact on CV morbidity and mortality of treatment with linagliptin against glimepiride: the CARdiovascular Outcome Trial of LINAgliptin vs. Glimepiride in DMT2 (CAROLINA). It will be interesting to assess the risk of a DPP‐4 inhibitor vs. a sulfonylurea (glimepiride) on HF.

Conclusions

DPP‐4‐I (gliptins) play an important role in the management of patients with DMT2; beyond their ability to improve glucose control, they also exert numerous pleiotropic actions, which have been proven in several experimental conditions but their evidence in humans are flimsy. Large, prospective trials involving a total of >40 000 high‐risk patients with tDMT2 either have been published or are ongoing: the first three confirmed the noninferiority of gliptins compared with placebo with no significant effect on those parameters potentially interrelated to their pleiotropic activities. For these reasons, there may be the possibility that their putative, favourable, CV actions may be offset by other, less known, negative off‐target effects, mediated by disparate DPP‐4 substrates.

Competing Interests

A.A. received research grants, lecture or advisory board fees from Merck Sharp & Dome, AstraZeneca, Novartis, Bayer, Boeringher‐Ingelheim, Sanofi, Mediolanum, Janssen, NovoNordisk, Eli Lilly, Servier, Vifor Pharma, Jannsen, and Takeda. G.P.F. received grant support, lecture or advisory board fees from AstraZeneca, Boehringer‐Ingelheim, Eli Lilly, NovoNordisk, Sanofi, Genzyme, Abbott, Novartis, Merck Sharp & Dohme.

Contributors

A.A. and G.P.F. collected the data, performed data analysis, and wrote the manuscript. Both authors approved the final version of the manuscript.

Avogaro, A. , and Fadini, G. P. (2018) The pleiotropic cardiovascular effects of dipeptidyl peptidase‐4 inhibitors. Br J Clin Pharmacol, 84: 1686–1695. 10.1111/bcp.13611.