OUP user menu

Potential cardiovascular effects of dipeptidyl peptidase-4 inhibitors in patients with type 2 diabetes: current evidence and ongoing trials

Ofri Mosenzon, Itamar Raz
DOI: http://dx.doi.org/10.1093/eurheartj/sus003 B22-B29 First published online: 5 November 2012


Cardiovascular disease (CVD) is a major cause of morbidity and mortality in patients with type 2 diabetes mellitus (T2DM). Cardiologists, who often treat patients with CVD and T2DM, are faced with the unmet need for an agent that provides glycaemic control yet does not pose CV risk. No antidiabetic therapy is currently indicated to improve macrovascular outcomes. Results of studies assessing the association between intensive antidiabetic therapy and a reduction in the risk of major CV events in patients with T2DM have been inconsistent, and independent reports have linked certain T2DM therapies (e.g. rosiglitazone, sulphonylureas) with negative CV outcomes. These findings prompted the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) to develop guidelines for assessing CV risk in investigational antidiabetic therapies. The FDA guidelines specifically call for meta-analyses of completed phase 2 and 3 trials; long-term, prospective, CV safety studies; or both. Results from meta-analyses involving dipeptidyl peptidase-4 (DPP-4) inhibitors, including those approved before the issuance of the FDA guidelines, suggest that these agents are not associated with an increase in CV risk and may potentially provide CV benefits. Prospective, large-scale, long-term trials designed in accordance with the FDA guidelines examining the CV risks and potential benefits of DPP-4 inhibitors are under way. This review discusses the current evidence and ongoing trials that may support the potential CV benefit of DPP-4 inhibitors in T2DM.

  • Cardiovascular disease
  • Dipeptidyl peptidase-4 (DPP-4) inhibitors
  • Guidelines
  • Type 2 diabetes mellitus


Cardiovascular disease (CVD) is a major cause of morbidity and mortality in patients with diabetes.1 Patients with diabetes are two to four times more likely to develop CVD than those without diabetes.1 Furthermore, CVD accounts for ∼50–60% of deaths in patients with type 2 diabetes mellitus (T2DM).2,3 Because of the potential for such important consequences, the National Cholesterol Education Program identifies diabetes as a coronary heart disease risk equivalent.4 Accumulating evidence suggests that the negative impact of T2DM on CV status may be attributed to a constellation of pathogenic processes, which include accelerated atherosclerosis, as well as abnormalities in inflammatory pathways and in endothelial, myocardial, and platelet function.58

Many CV therapies (e.g. antiplatelet, antilipidaemic, antihypertensive agents) have been shown to impart CV benefits in patients with T2DM. However, an antidiabetic agent that both effectively reduces blood glucose levels and improves CV outcomes in this population has yet to be definitively identified. Current antidiabetic therapies have been approved and introduced into practice without definitive long-term CV safety and efficacy data.9 Additionally, results of studies assessing the association between tight glycaemic control and a reduction in the risk of major CV events in patients with T2DM have been inconsistent.1013 Results from long-term, randomized clinical studies suggest that intensive therapy with some antidiabetic agents may not only fail to confer CV benefit but also be associated with increased mortality compared with standard therapy.14 No antidiabetic therapy currently is indicated to improve macrovascular outcomes. However, results from a 10-year follow-up from the UK Prospective Diabetes Study (UKPDS) in newly diagnosed patients15 showed a significant reduction in myocardial infarction (MI) in the intensive therapy group.10 Additionally, a meta-analyses of data from randomized clinical studies revealed a reduced risk of major coronary events, but not of CV mortality, with intensive therapy in patients with T2DM.1113 In the metformin treatment group, which consisted of patients who were overweight, substantial risk reductions for MI (39%, P = 0.01), and death from any cause (36%, P = 0.01) were observed during the original trial;15 however, these results were based on a small number of patients (n = 342).

In addition to these inconsistent findings, independent reports have linked certain commonly prescribed therapies for T2DM (rosiglitazone and sulphonylureas) with increased CV events and mortality.16,17 Taken together, these findings suggest the need for additional long-term studies assessing CV risks of T2DM treatments and have prompted the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) to develop guidelines for assessing CV risk with investigational antidiabetic therapies.18,19 In addition to providing important CV safety information, these studies may demonstrate CV benefits associated with these agents and address other long-term safety issues.

Emerging evidence suggests that some newer antidiabetic therapies, such as the incretin-based glucagon-like peptide-1 (GLP-1) analogues and dipeptidyl peptidase-4 (DPP-4) inhibitors, may help fill that need.20 The objectives of this article are to summarize the current EMA and FDA guidelines for establishing CV safety of antidiabetic agents, review findings from recent meta-analyses assessing the CV safety and efficacy of DPP-4 inhibitors, and introduce ongoing trials with these therapies.

FDA guidelines for the clinical investigation of antidiabetic agents

In 2008, the FDA released guidelines for the evaluation of CV risk in new antidiabetic agents.18 According to the FDA guidelines, before submission of a New Drug Application or Biologics License Application for antidiabetic agents, the incidence of important CV events must be assessed in a completed phase 2/3 clinical study programme through an integrated meta-analysis.18 These FDA guidelines impose statistical hurdles that affect the regulatory consequences for new antidiabetes agents (Figure 1).21 FDA guidance cites an upper bound of the two-sided 95% confidence interval (CI) of <1.8 for the estimated risk ratio (RR) of such events in the investigational group vs. the control group. In the case of a CI ≥1.8, additional studies would be required before approval of the drug to satisfy the requirement for risk assessment.18 If the upper bound of CI of pooled estimated RRs is between 1.3 and 1.8, one or more post-marketing safety trials should be performed to definitely demonstrate an overall upper bound CI of <1.3 (results may be from one adequately powered post-marketing trial or from a trial pooled with relevant pre-marketing clinical data).18 If the upper bound of the CI is <1.3 and the overall risk–benefit analysis supports approval, a post-marketing cardiovascular trial generally may not be necessary (Figure 1).18,21

Figure 1

Recent FDA guidelines impose statistical hurdles for approval of antidiabetes agents. The figure illustrates five hypothetical examples of possible hazard ratios (HRs) and the upper limit of the 95% confidence interval of a development plan. The regulatory consequences of each outcome also are indicated. Reproduced with permission of the American Diabetes Association and Boaz Hirshberg. Copyright 2011.21

The FDA guidelines further state that newly initiated clinical studies should include the following: (i) prospective, blinded adjudication of CV events; (ii) patients at higher risk of CV events; and (iii) a long study duration (e.g. ≥2 years). The FDA also recommends the use of meta-analyses to explore the incidence of important CV events across phase 2/3 studies, as well as the assessment of potential differences in CV RRs by subgroups (e.g. age, patient sex, race).18 In accordance with these regulatory requirements, long-term prospective studies have been initiated with various antidiabetic agents, including agents that were approved by the FDA before the guidance was issued.

EMA guidelines for the clinical investigation of antidiabetic agents

In 2010, the EMA released guidelines for the clinical investigation of antidiabetic agents similar to those issued by the US FDA.19 According to the EMA guidelines, antidiabetic agents should not be associated with an increased risk for CV events.19 Importantly, these guidelines stress that exploration of potential CV effects should occur throughout the course of the drug development programme. Clinical studies should include the following: (i) assessment of effects on atherothrombosis, cardiac functionality, and repolarization and conduction abnormalities in preclinical studies and (ii) the inclusion of study populations with comorbidities and concomitant drug regimens representative of patients treated with antidiabetic agents in clinical practice.19 Furthermore, the EMA recommends that clinical studies involving antidiabetic agents enrol a study population large enough to adequately detect safety signals, include patients at high risk for CV events, have a long-term duration of treatment (i.e. 18–24 months), and utilize prospective definitions of CV outcomes to assess CV risk accurately.19

These recent guidelines show the wider change in the evolution of drug approval, resulting in Adaptive Licensing (AL) approaches.22 AL approaches are based on stepwise learning of the safety and efficacy of new drugs, under acknowledged uncertainty, with repeated phases of data gathering and regulatory evaluation. Drug approval is no longer binary but a continuous process that relies on a combination of data from randomized control trials (RCTs), as well as observational data and real-world use.

Incretin hormones and DPP-4 inhibitors

The incretin hormones GLP-1 and glucose-dependent insulinotropic polypeptide (GIP) play important roles in regulating glucose homeostasis.23 Both are released from the gut in response to food intake and augment insulin secretion by pancreatic β cells in a glucose-dependent manner.23 In addition, GLP-1 lowers glucagon secretion by pancreatic α cells in a glucose-dependent manner.23 However, as part of the natural physiological process, incretin hormones are rapidly inactivated by the enzyme DPP-4.23 One pharmacological approach for potentiating the actions of incretin hormones is the oral DPP-4 inhibitors (alogliptin, linagliptin, saxagliptin, sitagliptin, and vildagliptin).20 Compared with older medications for T2DM, DPP-4 inhibitors are associated with a lower risk of hypoglycaemia,24 are weight neutral,24 and do not have negative effects on blood pressure.2527

Data from both pre-clinical2835 and early clinical studies3638 support a cardioprotective effect of GLP-1. It is suggested that this cardioprotective effect is mediated through GLP-1R–dependent39,40 and GLP-1R–independent mechanisms via GLP-1 metabolites.28 The evidence supporting a cardioprotective benefit for GLP-1 suggests that DPP-4 inhibition also might be associated with CV benefit based on the associated increase in the availability of GLP-1 due to this inhibition. Additionally, preliminary evidence suggests that DPP-4 inhibitors may have CV benefits mediated through several other DPP-4 substrates (e.g. stromal cell-derived factor-1, brain natriuretic peptide).20

Current evidence for cardiovascular safety and efficacy with DPP-4 inhibitors

Because phase 2 and 3 registration trials for these therapies were designed or completed before the 2008 FDA guidance on CV risk assessment, individual, retrospective meta-analyses of trial data were conducted to assess the effects of DPP inhibitors on CV risk (Table 1).4145 In all studies, the primary endpoint was a composite of major adverse cardiovascular events (MACE), which included CV death, stroke, and MI at a minimum. Some trials also included CV ischaemic events43 or hospitalization for unstable angina pectoris42 in their MACE definitions. The MACE endpoint was based on investigator-reported adverse events (AEs) that were systematically categorized according to preferred terms in the Medical Dictionary for Regulatory Activities.4244 Identified CV events were then retrospectively43 or prospectively42,45 adjudicated by a blinded independent committee of experts in all but the sitagliptin trial.44 Exposure-adjusted incidence rates were calculated per 1000 patient-years, compared between the DPP-4 inhibitor therapy group and comparator group, and expressed as RRs.

View this table:
Table 1

Study designs of individual meta-analysis assessing cardiovascular risk of dipeptidyl peptidase-4 inhibitors

DPP-4 inhibitorsaNo. of trialsbStudy duration (weeks)Total no. of patientsNo. of patients exposed to DPP-4 inhibitorPatient-years of exposure to DPP-4 inhibitorbComparator treatments
Alogliptin41812−2647023489Not availablePlacebo
Linagliptin42818−5252393319 (3159/160)c2060Placebo, glimepiride, or voglibose
Saxagliptin43816–116460733563758Placebo, metformin, and up-titrated glyburide
Sitagliptin441912–10610 24654294709Placebo, glipizide, glimepiride, insulin, metformin, pioglitazone, rosiglitazone, or combination
Vildagliptin452512 to ≥10413 5707509 (1393/6116)d686/7034dPlacebo, metformin, gliclazide, acarbose, rosiglitazone, pioglitazone, glimepiride, sulphonylurea, or combination
  • DPP-4, dipeptidyl peptidase-4.

  • aThe DPP-4 inhibitor could have been taken as monotherapy, initial combination therapy, or as add-on combination therapy with other glucose-lowering agents.

  • bPhase 2 or 3.

  • cLinagliptin 5 mg/linagliptin 10 mg.

  • dVildagliptin 50 mg/vildagliptin 100 mg.

Results from these meta-analyses show RR point estimates <1.0, with the upper bounds of the 95% CIs below the 1.3 FDA limit for most compounds (Table 2).4145 The upper bound of the 95% CI was >1.8 for alogliptin41 and vildagliptin 50 mg/day.45 Taken together, these findings suggest that treatment with a DPP-4 inhibitor is not associated with an increased risk of CV events. Moreover, risk reductions were statistically significant, with the upper bounds of the 95% CIs <1.0 in the saxagliptin and linagliptin meta-analyses, supporting a potential reduction in CV events.42,43

View this table:
Table 2

Results of individual meta-analysis assessing cardiovascular risk of dipeptidyl peptidase-4 inhibitors

DPP-4 inhibitor (no. of patients treated/comparator)Primary MACE endpointNumber (%) of patients experiencing an eventRisk ratio (95% CI)
DPP-4 inhibitorAll comparators
Alogliptin41 (3489/1213)Adjudicated CV death, non-fatal MI, or non-fatal stroke9 (0.26)5 (0.41)0.63 (0.21–1.91)
Linagliptin42 (3319/1920)Adjudicated CV death, non-fatal MI, non-fatal stroke, or UAP with hospitalization11 (0.3)23 (1.2)0.34 (0.16−0.70)a
Saxagliptin43 (3356/1251)Adjudicated CV death, MI, stroke22 (0.7)18 (1.4)0.43 (0.23−0.80)a
Sitagliptin44 (5429/4817)Reported CV ischaemic AEs0.6 per 100 patient-years0.9 per 100 patient-years0.68 (0.41−1.12)b
Vildagliptin45Adjudicated CCV death, ACS, TIA, stroke
 50 mg (1393/1555)10 (0.72)14 (0.90)0.88 (0.37−2.11)c
 100 mg (6116/4872)81 (1.32)50 (1.64)0.84 (0.62−1.14)c
  • ACS, acute coronary syndrome; AEs, adverse events; CCV, cardiovascular and cerebrovascular; CV, cardiovascular death (including fatal stroke and MI); DPP-4, dipeptidyl peptidase-4; MACE, major adverse cardiovascular event; MI, myocardial infarction; TIA, transient ischaemic attack; UAP, unstable angina pectoris.

  • aCox hazards ratio.42,43

  • bPoisson risk ratio.44

  • cMantel–Haenszel risk ratio.45

A recent meta-analysis combined the results from 53 RCTs that included 20 312 patients treated with different DPP-4 inhibitors and 13 569 controls treated with placebo or active comparators for 24 weeks or longer. The meta-analysis showed that the odds ratio for MACEs in the DPP-4 inhibitor treated group compared with all other treatment groups was reduced by 31% (odds ratio, 0.69; 95% CI, 0.53–0.90; P = 0.006).46

These recent meta-analyses provide important systematically collected information regarding CV events with DPP-4 inhibitors, but have inherent limitations. Most meta-analyses were retrospective in nature, and thus limited by the lack of pre-specified CV definitions or case report forms.4245 The low incidence of CV events, short duration of disease (mean, 3–8 years), and inclusion of monotherapy trials suggest that many patients had less advanced T2DM, and therefore a generally lower risk of CVD.42,43,45 Because data were based on drug registration trials, patients at increased CVD risk also may have been underrepresented. Moreover, despite large total patient exposures (10 24644 4607,43 13 570,45 523942), individual patient exposure was <2 years. Consequently, long-term data are still required to test the hypothesis that these therapies decrease CV risk.

Ongoing cardiovascular outcome trials with DPP-4 inhibitors

Epidemiological surveillance study

Although clinical trials provide valuable information on the efficacy, safety, and CV outcomes associated with T2DM therapies, it often takes years to obtain such results, which must be interpreted in the context of a controlled trial. In contrast, epidemiological databases provide additional real-time information on the safety of T2DM medications in real-world patients that can be evaluated repeatedly as evidence accumulates.

The US FDA Sentinel Initiative is an active, sustainable system that is being developed to leverage existing electronic, de-identified healthcare data from ∼100 million patients to monitor the safety profile and AEs associated with marketed medications, including the incidence of MIs in patients taking oral T2DM treatments.47,48 To assist in the development of the Sentinel System, a Mini-Sentinel pilot project using electronic health information obtained from claims data, in-patient and out-patient medical records, and patient registries is ongoing.48,49 Additionally, a prospective cohort study with data obtained from the Mini-Sentinel database will compare the incidence of acute MIs in patients receiving saxagliptin with those using other approved antidiabetic therapies [sitagliptin, long-acting insulin, pioglitazone, and second-generation sulphonylureas (glimepiride, glipizide, and glyburide/glibenclamide)].50 In the future, findings from this study may be compared with results from the prospective, controlled large-scale outcomes study of saxagliptin, Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus-Thrombolysis in Myocardial Infarction (SAVOR-TIMI 53).

Randomized clinical trials

There are several ongoing prospective, randomized, double-blind clinical trials assessing the CV safety of DPP-4 inhibitors (Table 3).9,5155 These trials address some of the limitations of the previously conducted meta-analyses. Specifically, these trials incorporate prospective blinded adjudication of CV events, inclusion of patients at increased risk for CV events (e.g. advanced age, pre-existing CVD, specific CV risk factors, renal disease), and long treatment periods.9 Similar to the meta-analyses, the primary MACE endpoint includes CV death, non-fatal MI, and non-fatal stroke; the linagliptin and sitagliptin studies also include hospitalization for unstable angina as part of the primary MACE endpoint.9,5155 The estimated completion dates for these trials range from 201455 to 2018,52 with planned sample sizes of ∼5400–16 500 patients. However, these trials are event-driven, and will end when a pre-specified number of adjudicated CV events have occurred.

View this table:
Table 3

Study design and inclusion criteria of ongoing cardiovascular outcome studies of dipeptidyl peptidase-4 inhibitors

DPP-4 inhibitor (clinical study)Study designPrimary endpointPlanned sample sizeInclusion criteria
HbA1c (%)MedicationsAge and CV history
Alogliptin (EXAMINE)51R, DB, PBO-controlled, phase 3; non-inferioritySafety54006.5–11.0OAD monotherapy or combination therapya≥18 years + ACS (past 15–90 days)
Linagliptin (CAROLINA)52R, DB, active-controlled (glimepiride) parallel-group, phase 3/4; non-inferioritySafety/efficacy60006.5–8.5Treatment-naïve or MET ± AGIb40–85 years
6.5–7.5SU/glinide (±MET or AGI)CVD, diabetes-related end-organ damage, ≥70 years, or ≥2 CV risk factors
Saxagliptin (SAVOR-TIMI 53)9R, DB, PBO-controlled, phase 4; superiorityEfficacy/safety16 5006.5–12.0Treatment-naïve, or antidiabetic treatment/insulina≥40 years with CVD or ≥55 years (men) or ≥60 years (women) with ≥1 CVD risk factor
Sitagliptin (TECOS)53,54R, DB, PBO-controlled, parallel-group, phase 3; non-inferioritySafety14 0006.5–8.0Stable dose(s) of antihyperglycaemic agent(s), including insulin≥50 years
Pre-existing CVD
  • ACS, acute coronary syndrome; AGI, α-glucosidase inhibitor; CAROLINA, Cardiovascular Outcome Study of Linagliptin versus Glimepiride in Patients with Type 2 Diabetes; CV, cardiovascular; CVD, cardiovascular disease; DB, double-blind; DPP-4, dipeptidyl peptidase-4; EXAMINE, EXamination of cArdiovascular outcoMes with alogliptIN versus standard of carE in patients with type 2 diabetes mellitus and acute coronary syndrome; HbA1C, glycated haemoglobin; MET, metformin; OAD, oral antidiabetic agent; PBO, placebo; R, randomized; SAVOR-TIMI, Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus−Thrombolysis in Myocardial Infarction; TECOS, Trial Evaluating Cardiovascular Outcomes with Sitagliptin; SU, sulphonylurea.

  • aExcluding other incretin-based therapy.

  • bExcluding treatment with other antidiabetic drugs.

In most trials, the primary MACE endpoint is considered a CV safety outcome to rule out an excess risk of events. However, some trials include evaluation of efficacy in their statistical plans. For example, the SAVOR-TIMI 53 trial is a superiority trial powered to assess the effect of saxagliptin on the reduction in CV events.9 In this study, patients with documented type 2 diabetes, glycated haemoglobin (HbA1C,) ≥6.5% and ≤12.0%, and either a history of established CV disease (secondary prevention) or multiple risk factors for vascular disease but without established CV disease (primary prevention) are being randomized. By enrolling patients with diabetes who are at high risk for CV complications, SAVOR-TIMI 53 is designed and powered to test for the superiority of saxagliptin vs. placebo and to exclude definitively any excess risk. Moreover, with few study limitations on concomitant use of other diabetic therapy, SAVOR-TIMI 53 will evaluate the efficacy and safety of saxagliptin across a broad spectrum of patients with T2DM. In the Examination of Cardiovascular Outcomes: Alogliptin versus Standard of Care in Patients with Type 2 Diabetes Mellitus and Acute Coronary Syndrome (EXAMINE) trial, the superiority of alogliptin to placebo for the primary MACE composite will be evaluated if non-inferiority is first demonstrated for the safety endpoint.51 Similar to the SAVOR-TIMI 53 study, the higher-risk CV study population (acute coronary syndrome in the past 15–90 days) included in EXAMINE will test for the superiority of alogliptin and exclude any excess risk. Thus, findings from these ongoing clinical studies will not only provide practitioners with information regarding the CV safety of DPP-4 inhibitors but also give important insights as to whether some of these agents are effective in reducing CV morbidity and mortality in patients with T2DM.


Cardiovascular disease greatly contributes to morbidity and mortality in patients with T2DM. There currently is an unmet need for a safe and effective antidiabetic therapy that provides both glycaemic control and CV benefits in patients with T2DM. No antidiabetic therapy currently is indicated to improve macrovascular outcomes. Results of studies assessing the association between intensive antidiabetic therapy and a reduction in the risk of major CV events in patients with T2DM have been inconsistent, and independent reports have linked certain T2DM therapies (e.g. rosiglitazone, sulphonylureas) with adverse CV outcomes. In the light of these issues, the FDA released guidelines in 2008 for evaluating the CV safety of T2DM medications. Preliminary evidence from meta-analyses suggests that DPP-4 inhibitors may reduce CV events in patients with T2DM. However, as of yet, no antidiabetic agent has been definitively proven to provide CV benefits in this patient population. Ongoing randomized trials designed in accordance with FDA guidelines and ongoing epidemiological surveillance studies are examining the CV safety and efficacy of DPP-4 inhibitors. These trials do not allow direct comparison of the agents, and even limited comparison will be difficult, given the differences in inclusion/exclusion criteria and in primary endpoints. Nonetheless, these trials may provide unique opportunities to gather valuable information about these agents. Additional data regarding the overall safety of DPP-4 inhibitors in the treatment of diabetes will certainly be elucidated. Most importantly, data from these trials will provide definitive conclusions on the potential CV benefits of this class of agents used to treat patients with T2DM.


The authors wish to acknowledge Scientific Connexions (Newtown, PA, USA), funded by Bristol-Myers Squibb (Princeton, NJ, USA) and AstraZeneca (Wilmington, DE, USA), for providing writing and editorial support.

Conflict of interest: O.M.: Consultant for AstraZeneca, Speakers Bureau: AstraZeneca, Merck Sharp & Dohme, Sanofi, Lilly, Novo Nordisk, and Novartis. I.R.: Advisory Board: AstraZeneca, Bristol-Myers Squibb, Eli Lilly, Merck Sharp & Dohme, and Novo Nordisk; Consultant: Andromeda, AstraZeneca/Bristol-Myers Squibb, Eli Lilly, Johnson & Johnson, HealOr, Insuline, Teva, and TransPharma; Speakers Bureau: AstraZeneca, Bristol-Myers Squibb, Eli Lilly, Johnson & Johnson, Novo Nordisk, and Roche; Honoraria European Society of Cardiology 2012 presentation: AstraZeneca/Bristol-Myers Squibb.


The authors take full responsibility for the content of this publication and confirm that it reflects their viewpoint and medical expertise.


View Abstract