Miles Fisher
Address for correspondence: Professor Miles Fisher
Department of Diabetes, Endocrinology & Clinical Pharmacology, Glasgow Royal Infirmary, 84 Castle Street, Glasgow G4 0SF, UK
E-mail: miles.fisher@ggc.scot.nhs.uk
https://doi.org/10.15277/bjd.2022.336
Dipeptidyl peptidase-4 (DPP-4) inhibitors were the first class of new antidiabetic drugs to be studied using modern cardiovascular safety trials, as mandated by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA). Trials with saxagliptin, alogliptin and sitagliptin satisfied the safety criteria with no increase in major cardiovascular adverse events (MACE), a composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke. No reduction in MACE was demonstrated in these trials, and an unexpected increase in the secondary outcome of hospitalisation for heart failure was observed in the SAVOR-TIMI 53 trial with saxagliptin.
Unusually, linagliptin was studied in two separate safety trials: CARMELINA compared linagliptin with placebo and CAROLINA compared linagliptin with glimepiride. In both trials MACE events were similar in the linagliptin and comparator groups, and no significant differences were observed in rates of hospitalisation for heart failure. These trials provide evidence of cardiovascular safety for linagliptin but show no clear cardiovascular benefits, and indirectly provide evidence of cardiovascular safety for the sulfonylurea glimepiride.
Br J Diabetes 2022;22:42-46
Key words: diabetes, cardiovascular outcome trial, linagliptin, glimepiride
Licencing requirements for new antidiabetic drugs changed in the US and EU following the rosiglitazone controversy, with a greater requirement to demonstrate cardiovascular safety. Between 2013 and 2015 three dedicated cardiovascular outcome trials (CVOTs) with dipeptidyl peptidase-4 (DPP-4) inhibitors in patients with type 2 diabetes were completed.1-3 SAVOR-TIMI 53 (Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus-Thrombolysis in Myocardial Infarction 53) was the first of these, comparing saxagliptin and placebo. It demonstrated no cardiovascular benefit with no difference in major adverse cardiovascular events (MACE), a composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke, and an unexpected increase in the secondary outcome of hospitalisation for heart failure.1
EXAMINE (Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care) compared alogliptin with placebo in patients following an acute coronary syndrome: it showed no significant difference in MACE or hospitalisation for heart failure.2 TECOS (Trial Evaluating Cardiovascular Outcomes with Sitagliptin) comparing sitagliptin with placebo showed no significant difference in the primary endpoint which was the composite of MACE plus hospitalisation for unstable angina, or in the secondary outcome of hospitalisation for heart failure.3 These three trials have been reviewed earlier in this series.4-6
This review describes results from the two cardiovascular safety trials performed with linagliptin in patients with type 2 diabetes: CARMELINA (Cardiovascular and Renal Microvascular Outcome Study with Linagliptin) comparing linagliptin with placebo,7 and CAROLINA (Cardiovascular Outcome Study of Linagliptin Versus Glimepiride in Type 2 Diabetes) comparing linagliptin with the sulfonylurea glimepiride.8 The review describes the primary endpoint and important secondary outcomes from the principal publications and directs attention to important subsequent publications of data from subgroups and/or post hoc analyses.
Linagliptin was licenced by the FDA in 2011 for use in the US and by the EMA in 2012 for use in Europe. A meta-analysis of cardiovascular events in 5,239 subjects from eight trials in the linagliptin phase 3 development programme was published in 2012.9 The primary endpoint was a composite of death, stroke, myocardial infarction (MI) and hospitalisation for unstable angina. In this cohort study there were significantly fewer cardiovascular events in the linagliptin group than in the comparator group who were treated with placebo, glimepiride or voglibose. A further prespecified meta-analysis of cardiovascular events in 9,459 subjects from 19 phase 2 and phase 3 trials was published in 2016 and showed no significant difference between the linagliptin and comparator groups.10 The cardiovascular safety of linagliptin was formally examined in two cardiovascular outcome trials (CVOTs). CAROLINA, comparing linagliptin with glimepiride, was started in 2010 and completed in August 2018, and CARMELINA was started in 2013 but completed earlier in January 2018.
For both CVOTs the primary endpoint was initially MACE, comprising cardiovascular death, non-fatal MI and non-fatal stroke, plus hospitalisation for unstable angina, sometimes termed ‘MACE plus’ or‘4-point MACE’. This was later changed to MACE alone or ‘3-point MACE’ as it was felt that the inclusion of hospitalisation for unstable angina introduced a degree of clinical subjectivity in assessing that outcome, and that 3-point MACE was diagnostically more precise.11
A paper describing the rationale, design and baseline characteristics of subjects in CARMELINA was published in 2018.12 The principal results were presented later that year at the meeting of the European Association for the Study of Diabetes (EASD) and published simultaneously in the Journal of the American Medical Association (JAMA). The key features of the trial and baseline characteristics of subjects are described in Box 1. Linagliptin is excreted in bile, whereas sitagliptin, saxagliptin and alogliptin are excreted in urine, so there was a focus in CARMELINA on renal outcomes and the study deliberately recruited subjects with diabetic kidney disease and a high risk of renal outcomes.
In CARMELINA there was no significant difference in MACE, so non-inferiority was established but not superiority (Figure 1, Box 2). There was no significant difference in the secondary renal outcome which was a composite of death due to renal failure, end-stage renal disease, or a sustained 40% or greater decrease in eGFR from baseline. A similar composite renal outcome was used in SGLT2 inhibitor trials e.g. DAPA-CKD. There was also no significant difference in rates of unstable angina or hospitalisation for heart failure when comparing linagliptin with placebo.
Other results from CARMELINA
The number of further publications from CARMELINA is small compared to the multiple post hoc publications from most cardiovascular outcome trials. The effect on cardiovascular and renal outcomes was examined according to baseline eGFR status (>60, 45 to <60, 30 to <45, and <30 ml/min/1.73m2).13 Across the different eGFR categories there was no beneficial effect of linagliptin on MACE or on the secondary renal outcome, but albuminuria progression was reduced with linagliptin regardless of eGFR status (Box 2). A further analysis looked at subjects with nephrotic range proteinuria, defined as a urinary albumin:creatinine ratio of > 2200mg/g at baseline.14 In this subgroup the cardiovascular event rate was double that of subjects without baseline nephrotic range proteinuria, and kidney events were greatly increased, but again there was no significant difference comparing linagliptin and placebo, and improvement in albuminuria was seen in the linagliptin group.
A paper describing the rationale for the active comparator trial design of CAROLINA was published in 2013,15 followed by a paper detailing the baseline characteristics of subjects in 2015.16 The principal results were presented in 2019 at the meeting of the EASD and published simultaneously in JAMA. Key features of the trial and baseline characteristics of subjects are described in Box 3. There was no significant difference in MACE in CAROLINA, so non- inferiority was established but not superiority (Figure 2, Box 4). There was also no significant difference in rates of unstable angina or hospitalisation for heart failure when comparing linagliptin with glimepiride.
Other results from CAROLINA
Again, the number of further publications from CAROLINA is small. An important substudy in CAROLINA was the CAROLINA-COGNITION study comparing the effects of linagliptin and glimepiride on accelerated cognitive decline (ACD), as diabetes is associated with an increased risk of cognitive impairment, particularly in patients with concomitant cardiovascular disease. This included 3,163 of the 6,033 subjects in CAROLINA. A paper describing the rationale and design of this substudy was published in 2018,17 with results published in 2021.18 No difference was observed in ACD, measured by the Mini-Mental State Examination (MMSE) and a more sensitive compositive measure of attention and executive functioning. Worryingly, ACD occurred in 27.8% of the linagliptin group and 27.6% of the glimepiride group, and the authors concluded that preventing cognitive impairment remains an unmet need in people with type 2 diabetes.
Cardiovascular outcome trials with the DPP-4 inhibitors saxagliptin, alogliptin, sitagliptin and linagliptin are now completed, and it is unlikely that any other CVOTs will be performed with this class of drugs. Vildagliptin was approved for use in Europe before the EMA introduced detailed cardiovascular safety requirements. Vildagliptin has not been approved for use in the US, and no CVOT has been performed. A meta-analysis of events in 17,446 patients in 40 double-blind phase 3 and phase 4 studies showed no difference in the rate of adjudicated MACE or hospitalisation for heart failure between vildagliptin and comparators.19
Several other DPP-4 inhibitors are approved for use in Asian and South American countries, but these have not been studied in large CVOTs.20 The once weekly DPP-4 inhibitor omarigliptin is approved in Japan and was under development in Europe and the US, including a CVOT. The trial was halted early for commercial reasons when the sponsor decided that approval would not be pursued in the US or Europe. This was an event-driven study and only 228 of the 632 required events were recorded when the trial was halted, so no conclusions can be drawn from the results.21
The results of the completed trials of DPP-4 inhibitors demonstrate safety for atherosclerotic cardiovascular events, with no increase compared to comparators, but no reduction in atherosclerotic events. For people with type 2 diabetes and existing atherosclerotic cardiovascular disease GLP-1 receptor agonists and SGLT2 inhibitors are a better treatment choice to further reduce events, as recommended in the consensus report by the American Diabetes Association and the EASD, with the additional benefit for patients that they are associated with weight loss.
Uncertainty remains about the effects of DPP-4 inhibitors on heart failure. No reduction in hospitalisation for heart failure was observed, and in SAVOR-TIMI 53 an increase in heart failure events was observed with saxagliptin. For the patient with heart failure, or with an increased risk of developing cardiovascular disease, SGLT2 inhibitors are a better choice to reduce further heart failure events.
The updated NICE guideline on the management of type 2 diabetes in adults retains DPP-4 inhibitors as first in a list of possible treatment options to be considered after metformin for patients who do not have chronic heart failure, established atherosclerotic cardiovascular disease or a high risk of cardiovascular disease.22 If a DPP-4 inhibitor is being used, no dose adjustment is required for linagliptin for patients with renal impairment. Pioglitazone is second in the list, sulfonylureas third, and then SGLT2 inhibitors which can be considered “for some people”. If the prescriber truly takes account of patient preferences, as recommended by NICE, then most patients will choose SGLT2 inhibitors for their associated weight loss rather than DPP-4 inhibitors which are weight-neutral or pioglitazone/sulfonylureas which clearly increase body weight.
Based on the results of CAROLINA if a sulfonylurea is to be prescribed then glimepiride has the greatest amount of cardiovascular safety data and should be the sulfonylurea of choice.
Conflict of interest The author has received personal fees from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Lexicon, MSD, NAPP, Novo Nordisk and Sanofi outside the submitted work.
Funding None.