Dipeptidyl peptidase-4 inhibitors and cardiovascular safety

Recent cardiovascular outcome trials confirm the safety of the dipeptidyl peptidase-4 inhibitors, but new questions emerge

In 2007, a meta-analysis of clinical trial data implicated the blood glucose-lowering drug rosiglitazone as unexpectedly increasing, rather than reducing, cardiovascular risk in patients with type 2 diabetes.1 This report proved controversial, but its major consequence was that the United States Food and Drug Administration (FDA) required the manufacturer of every new blood glucose-lowering therapy to conduct a postmarketing (Phase IV) cardiovascular safety study when there was no convincing evidence of non-inferiority or superiority over conventional agents, such as metformin and sulfonylureas, based on cardiovascular end point data collected in Phase II–III studies.2 In line with this requirement, the manufacturers of most incretin-based therapies, which comprise the glucagon-like peptide-1 (GLP-1) analogues and dipeptidyl peptidase-4 (DPP-4) inhibitors (gliptins), have initiated Phase IV cardiovascular safety studies. The exceptions are the GLP-1 analogue exenatide, which was already registered and marketed in the US at the time the FDA requirement came into effect, and the DPP-4 inhibitor vildagliptin, whose manufacturer elected not to do such a study and therefore not to market the drug in the US.

The first two postmarketing randomised placebo-controlled cardiovascular safety studies of incretin-based therapies to have reported involved the DPP-4 inhibitors saxagliptin and alogliptin. The Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus–Thrombolysis in Myocardial Infarction (SAVOR-TIMI) 53 trial recruited 16 492 patients with type 2 diabetes who had established, or were at high risk of, cardiovascular disease.3 It found that saxagliptin was non-inferior to placebo on a background of conventional blood glucose-lowering therapy for the commonly used end point of major adverse cardiovascular events (MACE; comprising cardiovascular death, non-fatal myocardial infarction and non-fatal ischaemic stroke) during a median follow-up of 2.1 years. Saxagliptin was well tolerated and safe, apart from a significant 27% increased risk of hospitalisation with heart failure (a prespecified and adjudicated clinical end point) compared with patients receiving placebo (3.5% v 2.8%). The Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) trial involved 5380 patients with type 2 diabetes and a history of an acute coronary syndrome between 15 and 90 days before recruitment.4 It found that alogliptin was non-inferior to placebo when added to usual diabetes care for the same primary end point of MACE over a median of 1.5 years. Alogliptin was similarly well tolerated and safe. Given the smaller sample size and shorter duration of follow-up than in the SAVOR-TIMI 53 trial, there were understandably fewer hospitalisations for heart failure, with a non-significant but nevertheless consistent 19% excess in the alogliptin group.

Given the rosiglitazone controversy, these findings seem reassuring in terms of MACE. However, an analysis of pooled outcome data from 41 959 patients followed for a mean of 44 weeks in Phase II and III studies of the available DPP-4 inhibitors (sitagliptin, vildagliptin, saxagliptin, linagliptin and alogliptin, all of which are registered in Australia) suggested benefit over conventional blood glucose-lowering therapies for MACE.5 This apparent inconsistency could be interpreted as showing the value of Phase IV studies, in which statistical power considerations and end point ascertainment are relatively robust, compared with the hypothesis-generating nature of earlier phase studies. Alternatively, it might reflect differences between Phase II–III and Phase IV studies in patient characteristics, glycaemic separation or comparator drugs. However, in the Phase II–III and Phase IV studies, there was a similar mean difference in glycated haemoglobin (HbA1c) levels between DPP-4 inhibitor and comparator arms (which included metformin, sulfonylureas and glitazones, as well as placebo) of between 3 and 6 mmol/mol.3-5

In relation to patient characteristics, those recruited to the SAVOR-TIMI 53 and EXAMINE trials were, in contrast to those who participated in Phase II–III studies, at a relatively late stage of their diabetes and its complications. Thus, glycaemic intervention alone would be unlikely to alter their vascular outcomes.6 In any case, the effects of improved glycaemic control on MACE typically take much longer than the duration of largely short-term Phase II–III studies and both the SAVOR-TIMI 53 and EXAMINE trials. In addition, the glycaemic separation between DPP-4 inhibitor and comparator groups has been modest. In intensively versus conventionally treated newly diagnosed patients recruited to the United Kingdom Prospective Diabetes Study, in which the updated mean HbA1c level was 10 mmol/mol lower in the intensive group, a reduction in MACE was only evident after 3 years and became maximal beyond 10 years of follow-up.6 Therefore, it is possible that potential cardiovascular benefits of DPP-4 inhibitors might be confined to patients with relatively early-stage diabetes (such as those who typically participate in Phase II and III studies) and, if they are observed within a few years of initiation of treatment, independent of glycaemia.5 However, it is very unlikely that even short-term comparative trials involving such patients will ever be conducted because of cost and patent life considerations.

In light of the non-inferiority of DPP-4 inhibitor treatment for MACE in the SAVOR-TIMI 53 and EXAMINE trials, the increase in hospitalisations for heart failure was unexpected. It may have been a chance finding, especially as there has been no similar signal in Phase II and III studies of DPP-4 inhibitors, and given that it was confined to a relatively small percentage of patients at high cardiovascular risk. However, other published data may provide clues as to its possible pathogenesis. In a study of patients with the metabolic syndrome, sitagliptin lowered blood pressure during placebo or incomplete angiotensin-converting enzyme (ACE) inhibitor treatment, but it increased blood pressure during maximal ACE inhibition.7 DPP-4 inhibition prevents the breakdown of the vasoconstrictor neuropeptide Y and, when ACE is inhibited, substance P. Increased circulating concentrations of these peptides may therefore increase the risk of heart failure in susceptible DPP-4 inhibitor-treated patients taking high doses of ACE inhibitors, through adverse effects on haemodynamics and/or myocardial remodelling.7,8

Consistent with this hypothesis, the Vildagliptin in Ventricular Dysfunction Diabetes trial found that patients with type 2 diabetes and heart failure taking vildagliptin showed unexpected increases in left ventricular end-diastolic and end-systolic volumes at 12 months, relative to placebo-treated participants.9 Although not associated with worsening heart failure or an increase in plasma N-terminal pro-brain natriuretic peptide (NT-proBNP) concentrations, these changes would usually be considered unfavourable drug-related effects on systolic function. In addition, there was a significant excess of angio-oedema in saxagliptin-treated patients in the SAVOR-TIMI 53 trial, albeit in small numbers (eight patients v one in the placebo group).3 This might reflect substantially increased circulating concentrations of vasoactive peptides such as substance P in some patients allocated saxagliptin and treated with an ACE inhibitor.

There are detailed patient-level data in the SAVOR-TIMI 53 and EXAMINE trials that could be used to shed light on possible clinically significant interactions between DPP-4 inhibitor allocation, ACE inhibitor dose and heart failure events. This could be done in association with multivariable analyses of available plasma NT-proBNP data and possible imbalances in baseline therapy, which included a numerical excess of patients treated with glitazones (a drug class known to increase the risk of heart failure) in the saxagliptin arm of the SAVOR-TIMI 53 trial.3 Heart failure will also be an end point of interest when the other ongoing DPP-4 inhibitor cardiovascular safety trials (TECOS and CAROLINA, involving sitagliptin and linagliptin, respectively) report in the next few years. In the meantime, and pending further data and analyses, there may be an argument for avoiding DPP-4 inhibitors in patients with, or at risk of, heart failure and using alternative blood glucose-lowering therapies with a confirmed safety profile for these patients.

New therapies for diabetes, such as those based on the incretin hormones, are inevitably expensive, especially given the cost of large-scale Phase IV safety studies. In light of this, as well as the heart failure data, there is therefore an economic argument for preferring conventional blood glucose-lowering therapies over DPP-4 inhibitors unless there is intolerance or a contraindication to established treatments. Nevertheless, the class has a relatively good pharmacological profile, including bodyweight neutrality, a low risk of hypoglycaemia, and safety in patients with renal impairment when given in an appropriate dose. Further, the SAVOR-TIMI 53 and EXAMINE trials have provided a range of reassuring safety data that extend beyond cardiovascular events to other important potential adverse effects, such as pancreatitis.3,4

It seems ironic that the FDA recently cleared rosiglitazone of an association with MACE after reanalysis of patient-level data from the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) study.10 Although the original rationale for Phase IV cardiovascular safety studies for new diabetes therapies may now be questionable, clinically useful data are nevertheless being generated through studies such as SAVOR-TIMI 53 and EXAMINE, which can only be considered an asset in an era of enhanced pharmacovigilance.